ss
BUOLEERIN OF THE BRITISH MUSEUM (NATURAL MISTORY)
ZOOLOGY Wolk 2 1971-1973
BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1977
DATES OF PUBLICATION OF THE PARTS
INO: i 5 October 1971 Nov 2%. 31 December 1971 INGE 3} 0 31 December 1971 No. 4 . 3 March 1972 INOS in 3 March 1972 No.6. 29 March 1972 No7): 30 March 1972 INO! 8) = 14 June 1972 No.9 . 16 March 1973
ISSN 0007-1498
Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4,5NU
CONTENTS
ZOOLOGY VOLUME 22
Hyoid and ventral gill arch musculature in eae ag fishes. By P. H. GREENWooD (Pls 1-2)
The clupeoid fishes described by Francis aa By P. K. ee & P. J. P. WHITEHEAD
Fine structure of Bodo saltans and Bodo eds ere Protozoa) and their affinities with the lees ae oy B. E. BRooKER (Pls 1-6)
The type specimens and identity of the species Bean in ae genus Lithobius by C. L. Koch and L. Koch from ee to a By E. H. Eason : 5 : : :
Contributions to the life-histories and development of Sones minutus Rudolphi, 1819 and C. heterochrous Bees 1802. 2) D. I. GiBson 2
Bats from Thailand a Geese. By Ne 135 Lebate & K. THONGLONGYA. c 6
A redescription of Suen servatus (Fage) Sue nov. (Mysidacea Lepidomysidae) from the material collected on Aldabra Atoll, with a key to the species of Lepidomysidae. By R. W. INGLE Recent records of mammals (other than bats) from Ethiopia. By G. B. CorBet & D. W. YALDEN . : . : 5
The shell structure and mineralogy of the Bivalvia. II. Lucinacea— Clavagellacea conclusions. By J. D. TayLor, W. J. KENNEDY & A. Hatt (Pls 1-15) ‘
Index to Vol. 22
103
I5I
171
197
211
253 295
eS . MUSCULATURE IN : . E _ OSTEOGLOSSOMORPH FISHES
P. H. GREENWOOD
ee
2) "BULLETIN. OF ° | E ne (NATURAL HISTORY)
Vol. 22 No. 1 _ LONDON : 1971
HYOID AND VENTRAL GILL ARCH MUSCULATURE IN OSTEOGLOSSOMORPH FISHES
BY
PETER HUMPHRY GREENWOOD
Pp. 1-55; 21 Text-figures
BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 22 No. 1 LONDON : 1971
THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series, corresponding to the Departments of the Musewm, and an Historical sertes.
Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year.
In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department.
This paper is Vol. 22 No. i of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals.
World List abbreviation Bull. Br. Mus. nat. Hist. (Zool.).
© Trustees of the British Museum (Natural History), 1971
TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY)
Issued 5 October, 197% Price {1-80
HYOID AND VENTRAE GILL ARCH MUSCULATURE IN OSTEOGLOSSOMORPH FISHES
By P. H. GREENWOOD
CONTENTS Page INTRODUCTION 0 : 4 MATERIALS AND METHODS . . : : : é : : : 7 HIoODONTIDAE: Hiodon alosoides ; 9 OSTEOGLOSSIDAE: Osteoglossum bicirr, aR 0 0 9 g 3 12 Scleropages leichardti . F ° ‘5 3 a 13 Heterotis niloticus 0 5 : : 6 : 14 Avapaima gigas. ; ¢ ¢ : : . 14 PANTODONTIDAE: Pantodon buchholzi ¢ : - : : 0 16 NOTOPTERIDAE: Papyrocranus afer . : a é F é 6 18 Xenomystus nigri . 21 Papyrocranus and Xenomystus eocpered an Notopterus 22 MorRMYRIDAE: Mormyrus kannume 23 Mormyrus caschive 27 Mormyrus lacerda 28 Mormyrus hasselquisti 28 Cyphomyrus discorhynchus . 30 Marcusenius cyprinoides . : . 5 : a 31 Marcusenius victoriae ci : F F Z : 32 Gnathonemus longibarbis . 0 , 4 0 : 32 Campylomormyrus elephas . @ 0 5 : 6 33 Petrocephalus bane. 0 A 2 OD 35 Petrocephalus catostoma * - ; A é a 37 Isichthys henryi : é : é é a : 37 Mormyrops anguilloides : 5 : é : 5 38 Hyperopisus bebe : : : : 0 : : 40 GYMNARCHIDAE: Gymmnarchus niloticus : ‘ F F i 9 44 SUMMARY AND DISCUSSION. ° 0 : ° 6 : c 48 Phyletic relations within the Sopreaity F : Z 5 4 . 53 ACKNOWLEDGEMENTS ¢ cs : : : o 5 . 54 REFERENCES . 5 F o c 3 6 9 el ‘i 0 54
ABSTRACT
THE ventral hyoid and gill arch musculature in several representatives of all osteo- glossomorph families is described and compared. The results of this survey show clearly the unusual specializations of the Mormyridae, and add ‘evidence to the
4 P. H. GREENWOOD
suggested relationships of the Mormyridae to the Notopteridae. The Osteo- glossidae and Pantodontidae show a different (and more usual) type of specialization. The Hiodontidae have a basic teleostean pattern in the hyoid muscles, not far removed from the Amza pattern.
Based on these and other characters, a new intragroup classification is suggested for the Osteoglossomorpha (including the reduction of the Gymnarchidae to sub- familial status within the Mormyridae).
INTRODUCTION
Although the ventral gill arch and hyoid musculature has been described (with varying degrees of thoroughness) for some osteoglossomorph fishes (Holmquist, 1911; Munshi, 1960; Bishai, 1967; Nelson, 1969), to date there has been no fully compara- tive account of these muscles; furthermore, only a few species have been investigated.
The present study is an attempt to fill both these gaps, and is part of a continuing investigation of intragroup relationships among the Osteoglossomorpha. That the musculature might provide some information on this subject is suggested by Nelson’s (1969) examination of dorsal and ventral gill arch muscles in certain notopterid, mormyrid, osteoglossid and hiodontid species. Nelson’s results are somewhat equivocal (and some are modified by my study), but if the hyoid musculature is also taken into account, a rather different picture emerges.
Nomenclature. For the gill arch muscles I have followed the nomenclature used by Nelson (1967 and 1969) which, in turn, was based on the terminology of Vetter (1878) and Edgeworth (1935).
The hyoid muscles provide something of a nomenclatural problem. The major muscle connecting the hyoid bar with the lower jaw is generally called either pro- tractor hyoideus or geniohyoideus. (See Holmquist, 1g11; Dietz, 1912, Edgeworth, 1928 and 1935; Munshi, 1960, and Osse, 1969 for discussions of this problem). It is clear that on grounds of homology and ontogeny (Edgeworth, op. cit.) the muscle should not be called a geniohyoideus in teleost fishes.
Associated with this muscle there is usually a much smaller, transverse muscle, the so-called intermandibularis, which lies anteriorly between the rami of the jaws.
As Holmquist (of. cit.) suggested, and Edgeworth (1928) later demonstrated embryologically, the protractor hyoideus is a compound muscle derived from an intermandibularis component anteriorly (the so-called posterior intermandibularis as distinct from the transverse anterior intermandibularis), and an interhyoideus component posteriorly.
Among the osteoglossomorphs studied, the protractor hyoideus shows, at least superficially, varying degrees of complexity or unity. In many osteoglossoids it appears to be a single muscle, in hiodontids a single muscle obviously of compound origin, and in mormyrids and notopterids a group of distinct muscles. Thus, in osteoglossoids I shall refer to this muscle as the protractor hyoideus, and in the notopterids and mormyrids the components will be named (vz. interhyoideus and posterior intermandibularis). The use of the term “‘protractor’ here is purely
OSTEOGLOSSOMORPH HYOID MUSCULATURE 5
nominal and does not imply any functional attributes (see Osse, op. cit.). In most species the anterior intermandibularis is clearly identifiable despite its great variation in size and area of attachment.
There is general agreement on the nomenclature for that part of the constrictor hyoideus ventralis associated with the branchiostegal rays, namely the hyohyoideus, and that term is used in this paper.
lic. 1. Outline drawings, not to scale, of: A. Huiodon alosoides. B. Osteoglossum bicivrhosum. C. Scleropages leichardti. D. Heterotis niloticus. E. Avapaima gigas. F. Pantodon buchholzi. G. Papyrocranus afer. HH. Xenomystus nigri.
P. H. GREENWOOD
Fic. 2. Outline drawings, not to scale, of various mormyrid species: A. Mormyrus kannume. B. Mormyrus caschive. C. Mormyrus lacerda. D. Mormyrus hasselquisti. E. Cyphomyrus discorhynchus. F. Marcusenius cyprinoides. G. Marcusenius victoriae. H. Gnathonemus longibarbis. 1. Campylomormyrus elephas. J. Petvocephalus bane. K. Isichthys henryi. L. Mormyrops anguilloides. M. Hyperopisus bebe. N. Gym- narchus niloticus.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 7
MATERIALS AND METHODS
Species
Hiodon alosoides
Hiodon alosoides Osteoglossum bicirrhosum Scleropages leichardti Scleropages formosus Heterotis niloticus Arapaima gigas Pantodon buchhola Papyrocranus afer Papyrocranus afer Xenomystus nigri Xenomystus nigri Xenomystus nigrt Notopterus kapirat Notopterus kapirat Mormyrus kannume Mormyrus lacerda Mormyrus hasselquisti Cyphomyrus discorhynchus Marcusenius cyprinotdes Marcusenius victoriae Marcusenius victoriae Gnathonemus longibarbus Campylomormyrus elephas Campylomormyrus elephas Petrocephalus bane Petrocephalus bane Petrocephalus bane Petrocephalus catostoma Isichthys henryi Tsichthys henryt Mormyrops deliciosus Mormyrops anguilloides Hyperopisus bebe Hyperopisus bebe Gymnarchus niloticus Gymnarchus niloticus Gymnarchus niloticus
B.M.(N.H.) register number
1965.7.2 : I-3 1965.11.23 : 1-8 1926.10.27 : I-2 1966.9.23 : 5 1966.9.5 : I unregistered unregistered 1913.3.12 : I 1969.3.26 : 27 unregistered unregistered unregistered unregistered
1931.7.20 : Q-I7
1938.2.22 : II unregistered 1965.3.15 : 316-320 unregistered 1966.7.29 : 2-4 1g61.12.1 : 48-74 1961.6.13 : 16 1962.2.6 : 26 1928.5.24 : 2 1928.7.30 : 3-4 I919.9.10 : 61
1907.12.2 : 231-232
1905.3.15 : I-2 unregistered 1961.6.21 : 7-18 1958.9.18 : 5-6 unregistered 1969.3.26 : 28 unregistered 1948.6.30 : I-4 1969.3.25 : 34 1969.3.26 : 49 1948.6.30 : 21
1902.11.10 : 56
Dissections were made on the following specimens; for most species dried skeletons or alizarin transparencies were also available.
Standard length,
mm.
185, 143,
TI2,
150, 136 124, I10, 108, 105 286
225 145, 140, 140, 130
167, 155
210, 200
137, 125 132
134
150 104 112
76, 73, 72, 68, 60
200 270, 235 205 190, 195 379 330 290
240
145, 140, 140
A: AIM: Apn: ASCh:
Al
Bb 1-3:
BhTp: BrR: BrM: Bt:
CbI-V:
CbM: Ch: Che: ChUh: (ibe Cr
CTM:
bD: Gr:
HaBM:
HbI: HbIL: Hhy: Hhya:
Hp: Hpic: IH: IHhy: Ihyi:
Ihyim:
Pp. H. GREENWOOD
Species B.M.(N.H.) Standard length register number mm. Gymmnarchus niloticus 1953-7-10 : 5 400 Amia calva unregistered 48 (head only) Albula vulpes 1949.11.29 : 1-4 177, 157, 152 Elops saurus 1961.8.31 : 45 134 ABBREVIATIONS USED IN TEXT FIGURES Articular Thyl: interhyoideus muscle, lateral anterior intermandibularis muscle division aponeurosis Ihyvl: interhyoideus, ventrolateral muscle articular surface for ceratohyal on IOp: interoperculum 1st basibranchial (Nelson’s [1968] Le: ligament from basihyal tooth plate
terminology)
anterior transversus muscle basibranchial, 1st—3rd arch basihyal tooth plate branchiostegal ray branchiostegal membrane basihyal, and its tooth plate, with connective tissue cover ceratobranchial (arches I-V) muscle connecting 4th and 5th ceratobranchials
ceratohyal
cartilaginous part of ceratohyal muscle between ceratohyal and urohyal
cleithrum
connective tissue surrounding lateral and ventral edges of basi- hyal and its anterior tooth plate connective tissue mass covering posteroventral end of basihyal tooth plate and articulation of the urohyal
dentary
gill rakers
hypaxial body musculature hypobranchial (arch 1)
of second arch
hyohyoideus muscle
anterior portion of hyohyoideus muscle
hypohyal
hypohyal (cut through) interhyoideus muscle
inferior hyohyoideus muscle interhyoideus muscle, inner (medial) division
interhyoideus muscle, innermost division
(to urohyal) LUh: ligament from urohyal to hypohyals
Max: cut end of maxilla Mb: mental barbel MC: mental cartilage
Ob 1-3: obliquus muscle (1st—3rd gillarches)
PhCA and P: external and internal pharyngocleithralis muscles
PIM: posterior intermandibularis muscle
PIMm: lateral muscle bands of posterior intermandibularis muscle
PIMms: posterior muscle slips of posterior intermandibularis muscle
PIMt: tendinous portion of posterior intermandibularis muscle
POp: preoperculum
Proll; bony process from second hypo- branchial
QO: quadrate
R: rectus muscle
RC; rectus communis muscle
S: sulcus between posterior inter- mandibularis and interhyoideus muscles
Sc: lateral line sensory canal in dentary
SH: sternohyoideus muscle
SHI: sternohyoideus, lower division
SHu: sternohyoideus, upper division
SHhy: superior hyohyoideus muscle
SOpBr: subopercular branchiostegal rays
it teeth
TIhy: tendon from interhyoideus muscle
TIhy 1: tendon from lateral division of interhyoideus muscle
TSH: tendon from sternohyoid
TSHBb: tendon from sternohyoid muscle to second basibranchial
Uh: urohyal
T-Vi; gill arches
OSTEOGLOSSOMORPH HYOID MUSCULATURE 9
Family HIODONTIDAE Hiodon alosoides (Rafinesque) (Text-fig. 1A)
Protractor hyoideus complex (text-fig. 3). Although at first sight there appears to be only a single muscle forming the floor of the mouth, closer examination of fibre direction shows that it is a compound of: (1) a large posterior sheet (fibres running obliquely anteromedially towards the median aponeurosis) (2) a much smaller, oval sheet anteriorly (fibres transverse to postero-medial) and (3) on each side, joining the former two sheets, a narrow elongate muscle with almost longitudinally directec fibres.
The various components are tightly joined through narrow aponcuroses, and there is some exchange of fibres between the different muscles.
The posterior sheet originates on the ceratohyal, with a few fibres stemming from the bases of the fourth and fifth branchiostegal rays; a fine median aponeurosis is visible along the entire length of the muscle. A sheet of dense connective tissue joins the lateral margins of this muscle to the ventral margin of the dentary on each side. Anteriorly the muscle ends aponeurotically on a broad sheet of connective tissue extending between the rami of the jaw, and lying dorsal to the other muscles of the protractor hyoideus complex. The posterior sheet joins, aponeurotically, the oval median muscle and, on each side, the slender elongate muscles. Thus, it has no direct insertion onto the lower jaw. In contrast, both the median oval and the lateral rectangular muscles attach directly to the median ventral face of each dentary.
Nt
BrR ag XC \ ( ieee
Fic. 3. Hiodon alosoides: ventral hyoid muscles after removal of the skin.
10 Po HH. GREENWOOD
From its innervation (a branch of the hyohyoideus VII nerve) and its origin on the ceratohyal, I would identify the posterior muscle sheet as the conjoined left and right interhyoidei. The anterior complex of three muscles (all innervated by a branch of the mandibular V nerve) would seem to be, from their topographical positions, an unpaired median anterior intermandibularis (the oval muscle) and a posterior inter- mandibularis divided by the backward extension of the anterior intermandibularis.
There is a striking similarity between the protractor hyoideus complex in Hiodon and the condition found in Amia, Elops and Megalops (personal observations on A. calva, E. saurus and M. cyprinoides; also Allis, 1897; and Liem, 1967). In all, there is a single anterior intermandibularis, paired (but medially joined in Ama and in the elopoid genera) posterior intermandibulares (the geniohyoideus of Liem of. cit.), and an extensive, paired interhyoideus. Hiodon differs from the others chiefly in having the interhyoid components more compact and closely associated with each other. Liem (op. cit.) describes musculose insertions of all components onto the gular plate, but I was unable to confirm this in my own dissections. Indeed, the muscles seem to be completely free from the gular plate which is attached only to the connective tissue covering the muscles. The protractor hyoideus in all four genera differs markedly from that in Albula (personal observations; see also Holmquist, 1g11; Liem, op. cit.), which represents a relatively specialized condition (see also p. 44 for further comments on Albula and Gymnarchus).
Hyohyoideus. The superior (interbranchiostegal) part of this muscle is weakly developed, and is largely tendinous (text-fig. 3). The inferior part is much better developed. It originates on the first (7.e. lowermost) branchiostegal ray of each side and inserts, mainly, on the hypobranchial of the opposite side (left muscle lying below the right). From the medial side of each inferior hyohyoideus a short slip of muscle inserts, through a long shared tendon, onto the basihyal plate about half way along its length. The tendon is closely applied to the ventral face of the basihyal plate even before its actual point of insertion.
Sternohyoideus (text-fig. 4). The main, ventrally situated part of this muscle inserts directly onto the urohyal, and has its origin aponeurotically, from the ventral body musculature. No ventral part of the sternohyoid originates on the cleithrum, but there is a smaller dorsal component originating on the horizontal limb of that bone. From about the middle of this upper segment a broad-based, almost com- pletely tendinous slip runs forward and upwards to insert on the basibranchial of the second gill arch. Apart from this link, there is no connection between the sterno- hyoideus and the branchial skeleton. The possible significance of this tendon in the evolution of the tendon-bones and ventral bony processes associated with the second gill arch in all other osteoglossomorphs (see Greenwood, Rosen, Weitzman, and Myers, 1966; Nelson, 1968) will be discussed later (page 51).
Ventral gill arch muscles (text-fig. 4). Well-developed obliqui muscles are present on the first three gill arches. The rectus communis is a large muscle at its origin from the ceratobranchial of the fourth arch, but it becomes tendinous as it passes below the medial end of the second obliquus, and remains tendinous until its insertion on the second basibranchial. This tendon also attaches to a small process from the
OSTEOGLOSSOMORPH HYOID MUSCULATURE 11
second hypobranchial, and is very closely associated with, but distinct from the ligament joining the first basibranchial to the second hypobranchial. A poorly- defined rectus is present between the fourth ceratobranchial and the third hypo- branchial; it is barely distinguishable from the larger and laterally situated rectus communis. Well developed anterior and posterior transversi link the proximal ends of the fourth and fifth ceratobranchials respectively.
External and internal pharyngocleithrales are present. The origin of the external division is medial to the sternohyoideus, but the internal muscle originates from the cleithrum at the same level as the sternohyoideus, and superficially resembles a division of that muscle.
AT
PhCP PhCA
Uh TSHBb SH HaBM CL. sss)
Fic. 4. Hiodon alosoides: ventral gill arch musculature and sternohyoideus in left lateral
view. 4
12 P. H. GREENWOOD
Family OSTEOGLOSSIDAE Subfamily OSTEOGLOSSINAE (Nelson, 1968) Osteoglossum bicirrhosum Vandelli (Text-fig. 1B)
Protractor hyoideus: is a stout muscle, unpaired except posteriorly over its origins (on each ceratohyal and, in part, aponeurotically from the inferior hyohyoideus of each side). It inserts, through a very short tendinous portion, onto the anterior part of the dentary around the symphysis. This anterior section of the muscle is split, horizontally, by the passage of a stout, transversely aligned anterior inter- mandibularis; that part of the protractor lying above the intermandibularis is thicker.
There is no median longitudinal aponeurosis, but there is a distinct tendinous inscription running transversely at about the middle of the muscle.
Holmquist (1911) and, later, Edgeworth (1935, fig. 277) described the muscle as divided horizontally into a broader, dorsal, interhyordeus portion, and a narrower, ventral (and medial) posterior intermandibularis portion; both authors also figure the transverse inscription. Despite careful dissection and probing, I could not find any such horizontal division in the muscle. Indeed, in the two specimens I dissected (28-6 and 34-0 cm standard length) the muscle could better be interpreted as being transversely and vertically divided. For example, that section lying anterior to the inscription is innervated by a branch of the mandibular V nerve, but most of the muscle behind the inscription is supplied by a branch of the hyohyoideus VII. This would seem to imply that the anterior part is derived from the posterior inter- mandibularis, and the posterior part from the interhyoideus However, since the inscription does not penetrate deeply into the muscle (7.c. it is not a complete, plate- like aponeurosis) and because the branch of the trigeminal nerve extends behind it, I would not be prepared to delimit the component parts on adult morphology alone. My uncertainty is reinforced by the condition of the muscle in Pantodon buchholz (see page 16), where it seems to approach closely the condition described by Holm- quist for Osteoglossum bicirrhosum.
Hyohyoideus. The superior portion (between the branchiostegal rays) is moder- ately developed. The inferior portion, although narrow, is fully muscular. It originates entirely from the first branchiostegal ray and has a tendinous insertion mainly onto the hypohyal of the opposite side; a few fibres, however, have a tendi- nous insertion onto the hypohyal of their own side. (It may be noted that Holmquist [1911, fig. 11] shows the right inferior hyohyoideus overlapping the left, but in all specimens I have examined [and in all other osteoglossids] left overlaps right).
Sternohyoideus (text-fig. 5). The greater part of this muscle originates on the dorsal surface of the horizontal limb of the cleithrum; a small part stems from the anterior tip of the conjoined cleithra. The sternohyoideus inserts onto the urohyal, but that part passing below the first gill arch is closely attached to the hypobranchial
OSTEOGLOSSOMORPH HYOID MUSCULATURE 13
by a thick connective tissue fascia. The muscle is also closely attached to the inner aspect of the ventrally directed bony processes on the second hypobranchials; each process is, however, entirely superficial to the muscle.
Ventral gill arch muscles (text-fig. 5). Obliqui muscles (linking cerato- and hypo- branchial elements) are present on the first three gill arches. A small rectws muscle connects the fourth ceratohyal with the third hypobranchial.
No rectus communis is developed.
The proximal ends of the fourth and fifth ceratobranchials are joined, respectively, by the anterior and posterior transversi. The fifth ceratobranchial is also connected to the cleithrum by the strong external and internal pharyngocleithrales. The internal pharyngocleithralis is subdivided into a narrow posterior and a much broader anterior part. Both pharyngocleithrales have their origins medial to the sternohyoideus.
Scleropages leichardti Giinther (Text-fig. 1C) In most details, the musculature of S. leichardti is identical with that described for Osteoglossum bicirrhosum. Comments made above on the morphology of the
protractor hyoideus muscle and its components apply equally to the muscle in Scleropages.
HbI HbI Gr
PhCP
HaBM
Hpic Bb1 Pri SH ees
Fic. 5. Scleropages leichardti: ventral gill arch muscles and sternohyoideus in left lateral view. The hypohyal has been partly cut away.
14 P. H. GREENWOOD
Among the ventral gill arch muscles, the only intergeneric difference noted was that in Scleropages the proximal end of the rectus muscle is fused with the hypo- branchial head of the obliquus of the third arch (text-fig. 5).
Subfamily HETEROTINAE (Nelson, 1968) Heterotis niloticus (Cuvier) (Text-fig. 1D)
Protractor hyoideus. This stout muscle originates on the ceratohyal and lower three branchiostegal rays of each side, but is a single element over its anterior half. It inserts directly, through short left and right heads, on either side of the dentary symphysis. Slightly anterior to the point of tusion between left and right halves, there is a very distinct transverse inscription; no other division of the muscle can be detected by dissection. The anterior part of the protractor is supplied by a branch of the mandibular V nerve, the posterior part by a branch of the hyohyoideus VII. The latter branch emerges from the medial side of the branchiostegal membrane at the base of the first ray; the main nerve continues forward, and supplies the inferior hyohyoideus muscle.
Heterotis is unusual in having a strong connective tissue link between the pro- tractor hyoideus and the hypohyals (at a point near the union of the protractor’s two halves).
A small and narrow anterior intermandibularis lies dorsal to the protractor hyoideus and does not pass through any part of it (cf. Osteoglossum and Scleropages).
Sternohyoideus. This stout muscle originates almost entirely from the dorsal surface of the horizontal limb of the cleithrum; a few ventral fibres stem from the anterior tip of that bone. It inserts onto the urohyal but is also firmly attached to the hypobranchial of the second arch and, less intimately, to the first hypobranchial as well. The ventral processes from the second hypobranchials are partly buried in the sternohyoid.
Ventral gill arch muscles. The most outstanding feature of these muscles is the development of a rectus communis from the fourth ceratobranchial to the base of the process on the second hypobranchial; there is also a tendinous connection between this muscle and the third hypobranchial.
In all other respects the gill arch muscles (including the vectws) are like those of Osteoglossum and Scleropages. The pharyngocleithrales of Heterotis, however, are simpler since the internal muscle is undivided.
Arapaima gigas (Schinz) (Text-fig. IE)
Protractor hyoideus: is a short, largely paired muscle originating from the cerato- hyal and the basal parts of the second and third branchiostegal rays. Only about
OSTEOGLOSSOMORPH HYOID MUSCULATURE 15
the anterior third of the muscle is unpaired, the two halves meeting along a weak aponeurosis. Slightly anterior to this junction there is a transverse tendinous inscription.
The protractor inserts, through left and right musculose heads, on either side of the dentary symphysis. A small and narrow anterior intermandibularis passes through the muscle a little posterior to its insertion.
It is impossible, by inspection or dissection, to determine the extent of the protractor’s component muscles, nor is it possible to determine the manner of their fusion.
Hyohyoideus. The superficial part is moderately well-developed and largely muscular. The inferior division is short and relatively stout; it inserts through a long tendon onto the hypohyal of the opposite side.
Sternohyoideus. Undoubtedly the sternohyoideus is the most characteristic muscle of Avapaima gigas (see text-fig. 6). It is divided, horizontally, into a small anterior and dorsal division, and a much larger unpaired ventral part. The latter originates (as is usual) on the horizontal limb of the cleithrum, and inserts on either side of the peculiarly shaped urohyal. In cross-section, this bone is shaped like an inverted T, with the arms extended to such a degree that the anterior half of the lower sternohyoid is completely covered by bone. A broad ligament from either side of the urohyal attaches it, ventrally, to the ceratohyals and hypohyals. The small ventral process of the second hypohyal barely contacts the dorsal part of this ventral sternohyoid division. Indeed, the process in Avapaima is the shortest found in any osteoglossid, and has the least intimate contact with the sternohyoid.
Uh TSH Bt
Smm CbI BrM BrR Gr SHu Cb eee Fic. 6. Avapaima gigas: ventral view of the sternohyoid, to show its upper and lower divisions and their relationship with the urohyal, cleithrum and first hypobranchial. Semi-schematic; the lower jaw is not depicted. Cb=ceratohyal.
16 Pp. H. GREENWOOD
The dorsal division of the sternohyoid is paired. Each half originates on the ventral face of the first hypobranchial, and extends over almost the entire length of that bone (thus underlying the first obliquus muscle). The left and right halves meet, aponeurotically, in the midline; the broad tendon originating at this union inserts, somewhat asymmetrically, on the inner aspect of the ceratohyals near their tips. Although some part of the tendon inserts on the right ceratohyal, by far the greater part is attached to the left bone.
No other osteoglossid shows such clear-cut subdivision of the sternohyoid, or such asymmetry in its insertion. In fact, insertion of even part of the sternohyoid onto the ceratohyals is most unusual in teleosts. Nevertheless, it is difficult to identify the upper division of this muscle in Avapaima as other than part of the sterno- hyoideus. That other osteoglossids have a close association of the sternohyoid with the first and second hypobranchials, and that there is an incipient division of the muscle in Heterotis (see p. 14), all seem to support the recognition of a divided sternohyoid in Arapaima.
Ventral gill arch muscles: in Avapaima have the typical osteoglossid pattern ; there is no rectus communis.
Obliqui are present on the first three arches; the obliquus of the third arch is closely associated with the well-developed vectws running almost in the midline from the fourth ceratobranchial to the third hypobranchial. Anterior and posterior tyansversi are present, as are stout and undivided internal and external pharyngo- cleithrales.
Family PANTODONTIDAE Pantodon buchholzi Peters (Text-fig. 1F)
Protractor hyoideus (text-fig. 7): is a single muscle over its anterior half, but is paired posteriorly, with the left and right halves originating on the ceratohyal and lower three branchiostegal rays of their side.
The muscle is marked by a transverse inscription at the point where the two halves unite. At this point, the unpaired portion is visibly separable into a broad dorsal section (inserting on either side of the dentary symphysis) and a much narrower, more compact, median and ventral part which inserts, tendinously, onto the sym- physis itself. Slight pressure with a probe along the horizontal sulcus demarcating the two parts separates them back to the level of the transverse inscription. Beyond this point there is considerable interchange between the parts, and the sulcus itself is no longer distinct.
A well-developed, stout, transverse anterior intermandibularis lies between, and separates, the dorsal and ventral parts of the protractor anteriorly.
In many respects, the condition of the protractor hyoideus in Pantodon resembles that described for Osteoglosswm by Holmquist (op. czt.) and Edgeworth (1935); it will be recalled that I found a rather different arrangement in that genus (see p. 12).
OSTEOGLOSSOMORPH HYOID MUSCULATURE 17
Pantodon differs from Holmquist’s description of Osteoglossum in that the muscles cannot be separated posteriorly beyond the transverse inscription. Holmquist and Edgeworth identify the entire length of the median, ventral section of the muscle in Osteoglossum as a protractor hyoideus (i.e. a posterior intermandibu aris in the terminology used here), and the overlying, broader part as the interhyoideus component.
Identifying the components of the protractor hyoideus in Pantodon is not easy, particularly since the innervation cannot readily be traced within the muscle. Judging from the position of the upper and lower insertions, and from the fact that the ventromedial segment is so clearly circumscribed, I would identify it as the posterior intermandibularis; the much larger muscle above, lateral to and behind it would then be the interhyoideus portion. In other words, a situation like that in the protractor hyoideus of Salmo salar (see Holmquist, 1g11, and Dietz, 1912).
Hyohyoideus. Both the superior and inferior divisions are well-developed, the latter originating on the first branchiostegal ray, and inserting tendinously on the hypohyal of the opposite side.
Sternohyoideus. The lower third of this muscle originates on a broad aponeurosis with the ventral body musculature; the remainder stems from the dorsal surface of the horizontal limb of the cleithrum. Anteriorly, the sternohyoid inserts around the small urohyal which is completely embedded in the muscle. The medial face of
Fic. 7. Pantodon buchholzi: ventral hyoid musculature. A. In ventral view. B. In oblique ventro-lateral view. The transverse inscription (aponeurosis) is shown, but not labelled, in both views.
18 P. H. GREENWOOD
each process from the second hypobranchial is firmly but superficially attached to the sternohyoid, which is also closely attached by connective tissue to the ventral side of the first hypobranchial.
Ventral gill arch muscles. As in Heterotis (but not other osteoglossids) a distinct rectus communis is present; the small vectus from the fourth ceratobranchial to the third hypobranchial is distinct from the rectus communis but is closely applied to it. Obliqui are present on the first three arches; anterior and posterior tvansversi are well developed, and the pharyngocleithrales are simple but relatively stout.
Family NOTOPTERIDAE
Munshi has given a detailed description of the cranial muscles in the Asiatic species Notopterus chitala, and Nelson (1969) has listed the branchial muscles of the same species. In view of this previous work, I have concentrated on the two African species, Papyrocranus afer and Xenomvystus nigri (see also Greenwood, 1963 and Nelson op. cit.). Some comments on Munshi’s description of the hyoid muscles in N. chitala, and a general comparison of the hyoid and gill musculature in the three genera follow the separate accounts for Papyrocranus and Xenomystus.
Papyrocranus afer (Ginther) (Text-fig. 1G)
Ventral hyoid musculature (text-fig. 8). The most superficial (7.e. ventral) muscle has its origin, on each side, equally from the ceratohyal and from the basal part of branchiostegal rays 2 to 6. The two halves of this muscle join to form a single element over about the anterior half of their length.
Anteriorly, the muscle inserts onto the dentary through dorsal and ventral heads; the single, narrow and tendinous ventral insertion is onto the symphysis, while the broader, more musculose dorsal insertion is double and lies on either side of the symphysis.
This ventral muscle is innervated solely by a branch of the mandibular V nerve. Careful dissection shows that no branch of the hyohyoideus VII runs to it (see below under interhyoideus).
With respect to its innervation, the muscle differs from the topographically similar muscle (the protractor hyoideus) in osteoglossid and pantodontid fishes. Munshi (op. cit.) identifies the muscle in Notopterus as a posterior intermandibularis and I would agree with his identification, both on the grounds of the muscle’s innervation solely from the trigeminal nerve, and because distinct interhyoideus muscles (innervated from the facialis nerve) are also present. It will be recalled
. |
OSTEOGLOSSOMORPH HYOID MUSCULATURE 19
SHhy
AIM
IHhy BrM LA PIM B Che Ch IH BhTp Lom oe CTtcut) D
SHhy Hp
THhy(left) Ch L
Fic. 8. Papyrocranus afer: A. ventral hyoid muscles, after removal of the skin. B. ventro-lateral view of head to show the interhyoideus muscles. The right half of the lower jaw removed, as is the cheek and jaw musculature of that side.
20 P. H. GREENWOOD
that in osteoglossoids the protractor hyoideus has a double (trigeminal and facialis) innervation, and that in these fishes no separate interhyoideus is present.
A small and weak anterior intermandibularis muscle lies between the dorsal and ventral insertions of the posterior intermandibularis. Its fibres are transversely orientated, and attach to the median face of the dentary on either side of the symphysis.
Interhyoideus (text-fig. 8). This paired muscle appears to be further divided on each side into a shallow lower portion, and a deeper upper part (see fig. 8B). Anteriorly, however, the two parts have a common tendinous insertion onto the dorsal surface of the hypohyal of the same side. The upper division originates entirely from the ceratohyal, but the lower part originates from the ventral margin of the ceratohyal and, in minor part, from the heads of the first two branchiostega. rays. The interhyoideus is innervated by a branch of the hyohyoideus VII nervel This nerve enters the lower muscle division after emerging from behind the branchio- stegal rays and passing along the upper margin of the first ray.
Hyohyoideus. The superior division is rather poorly developed, and is almost completely tendinous. The inferior part is also tendinous but has its medial third strongly muscular (fig. 8B). This muscular section originates from the distal third of the first branchiostegal ray. The whole inferior hyohyoid inserts, tendinously, on the hypohyal of the opposite side (the left muscle passing ventral to the right).
Sternohyoideus (text-fig. 9). The large sternohyoideus originates entirely from the dorsal surface of the horizontal limb of the cleithrum. Ventrally, part of the muscle inserts, through a pair of tendons, onto the hypohyals and the anteroventral tip of each ceratohyal (a most unusual arrangement; see also Avapaima, page 15). However, the bulk of the sternohyoid inserts onto and around the small urohyal which is completely embedded in the muscle. (It should be noted that a pair of strong ligaments from the ventral face of the basihyal tooth-plate also attach to the urohyal and are, in consequence, partly covered by the sternohyoid.)
The median face and posterior margin of each ventral process (tendon bones) from the second basibranchial are firmly but superficially attached to the lateral face of the sternohyoid on each side. Anterior to this point the muscle closely approaches the first basibranchial but is not attached to it.
Ventral gill arch muscles (text-fig. 9) | Well-developed ob/iqui are present on the first three gill arches; the muscles of all three arches insert on the respective cerato- branchial, but those on the first and second arch have a double origin, from the hypo- and basibranchial.
A distinct vectus runs from the third hypobranchial to the ceratobranchial of the fourth arch where it comes into close contact with the head of a stout rectus communis connecting that arch with the base of the ventral process (tendon bone) articulating with the second basibranchial.
The prox mal ends of the fourth and fifth ceratobranchials are joined, respectively, by the broad anterior and posterior transverst.
The external and internal pharyngocleithrales are well-developed, simple muscles whose origins lie medial to the sternohyoideus.
OSTEOGLOSSOMORPH HYOID MUSCULARE 21
Xenomystus nigri (Ginther) (Text-fig. 1H)
In all major details the musculature of this species is like that of Papyrocranus afer. The most noticeable differences may be listed briefly.
Posterior intermandibularis. The origin is mainly from the ceratohyal, with only about one quarter stemming from the first branchiostegal ray. It inserts through a single broad head, an arrangement possibly correlated with the much weaker and rather ill-defined anterior intermanibularis in this species. As in Papyrocranus the posterior and anterior intermandibulares are innervated by a branch of the mandi- bular V nerve.
Hyohyoideus. The inferior divisions fuse in the midline before inserting onto the hypohyals.
Interhyoideus: in this species is a pair of undivided muscles fused anteroventrally, and inserting onto the hypohyals through a short tendon which wraps around the face of these bones.
Cbr HbI Hpic BhTp
LUh
SH Uh TSH Chileft)
Smm
Fic. 9. Papyrocranus afer: lateral view of gill arch muscles and sternohyoideus. The anterior part of the sternohyoid has been dissected away on the right side to show the embedded urohyal. The tendon (TSH) inserts onto the right ceratohyal (removed) and right hypohyal (partly removed).
22 P. H. GREENWOOD
Gill arch muscles. 1 was unable to locate a separate rectus in Xenomystus. The rectus communis is present and large; like the obliqui it appears to be relatively larger in this genus. As in Papyrocranus, the obliqui of the first two arches originate from both the hypo- and the basibranchials.
PAPYROCRANUS AND XENOMYSTUS COMPARED WITH NOTOPTERUS
Basically, the musculature in the three genera is very similar (see Munshi, 1960 for Notopterus). In the two Notopterus species I examined (N. chitala and N. kapirat), the interhyoideus of each side is undivided (i.e. like that of Xenomystus); in N. kapirat it is indistinguishable ventrally from the upper, anterior part of the inferior hyohyoideus. According to Munshi (oP. czt.) the left and right interhyoidei of N. chitala fuse medially. But, in his figure the muscles are shown as fan-like, horizon- tally disposed structures continuous with the inferior hyohyoidei, and running obliquely towards the midline; here the posterior portion ot the left interhyoideus overlaps that of the right muscle. In the single specimen I dissected, the inter- hyoidei are vertically aligned (as in the other species considered), insert separately and are quite distinct from the inferior hyohyoidei. In other words, a situation identical with that found in Papyrocranus.
I would contest Munshi’s identification of the muscle running from each basi- branchial process (tendon bone) to the urohyal. This he calls the pharyngohyoideus (= rectus communis in the nomenclature used above). However, the figure shows the muscle as only partly distinct from the sternohyoideus (Munshi’s rectus cervicus) below it, and no fibres are shown connecting with the upper posterior face of the tendon bones (as do those of the rectus communis). In the specimen I have exam- ined (as in N. kapirat, Papyrocranus and Xenomystus also) it is not possible to separate the fibres attached to the tendon bone from those of the underlying sterno- hyoid. Incontrast, the muscle I have identified as a rectus communis is distinguish- able from the sternohyoideus even at its insertion onto the tendon bone where the two muscles are closely apposed. Thus it seems likely that the muscle Munshi identifies as a rectus communis is, in fact, part of the sternohyoideus.
Family MORMYRIDAE
The bauplan of the hyoid and gill arch musculature is remarkably constant in the Mormyridae, despite the great range of variation in head shape and jaw form of these fishes. The more outstanding features of the mormyrid musculature may be summarized as follows :—
(i) The muscles of the protractor hyoideus complex show the greatest degree of individuality and subdivision of any osteoglossomorph fishes; the interhyoideus muscles are enlarged and insert onto the lower jaw.
(ii) There is considerable hypertrophy of the hyohyordeus with the consequent loss of distinct superior and inferior divisions; the left and right halves of the muscle
I ee
OSTEOGLOSSOMORPH HYOID MUSCULATURE 23
join medially and the muscle is attached to the interopercula and ceratohyals as well as to the urohyal. This specialized musculature is associated with the peculiar branchial specializations of all mormyrids. For example, there is no discrete and expandible branchiostegal membrane, the rays being buried in the hyohyoideus, and the whole mass covered by skin continuous with that of the body; in consequence, the opercluar aperture is greatly restricted and lies entirely above the branchiostegal rays.
(iii) In the ventral gill arch musculature the obliqui have extensive areas of attachment, and the weakly developed pharyngocleithrales lie external to the sterno- hyoideus. The anterior part of the sternohyoideus is closely associated with the corresponding ventral elements of the gill arches, although there is never a direct musculose or tendinous insertion onto these bones.
Surprisingly, little attention has been paid to the cephalic and branchial muscula- ture of mormyroid fishes. Holmquist (1911), drawing on his investigation of the hyoid musculature in Gymnarchus niloticus, noted certain peculiarities, but he was unable to extend his observations because of lack of comparative material. Nelson (1969) briefly commented on the gill arch muscles in Mormyridae and compared these with those of other osteoglossomorphs. The most comprehensive treatment is that of Bishai (1967) on Mormyrus caschive. Unfortunately, the nomenclature used by this author makes direct comparison very difficult, and I suspect that his description of the gill arch musculature is erroneous in many respects (see below, page 28).
Since the bauplan of the hyoid and ventral gill arch muscles is so similar in all mormyrids, I shall give a detailed description for one species only; deviations from this pattern will be noted for the other species examined.
Mormyrus kannume Forsk. (Text-fig. 2A)
The snout in this species is slender, moderately decurved, produced and tubular, with the mouth small and terminal.
Posterior intermandibularis (text-fig. 10A): is a fairly thick muscle with originates, aponeurotically, over the anterior part of the hyohyoideus, and directly from the interoperculum of each side. It inserts through two muscular heads on either side of the mental cartilage (that is, it does not attach directly to the dentary). At about its midpoint there is a faint, longitudinal aponeurosis which extends anteriorly; posterior to this point both halves of the muscle are contiguous medially throughout their lengths. The posterior intermandibularis is broad and extends laterally almost to the ventral margin of each dentary, to which it is attached by a connective tissue sheet.
Anterior intermandibularis (text-fig. 10B): lies immediately above the posterior division. It is a relatively thin but expansive muscle extending longitudinally from the level of the posterior interopercular margin to a little behind the mental cartilage.
P. H. GREENWOOD
Fic. 10. Moyrmyrus kannume: ventral hyoid muscles. A. After removal of the skin. B. Posterior intermandibularis removed. C. Anterior intermandibularis removed; right lateral division of the interhyoideus muscle removed.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 25
Laterally it inserts, directly, onto the interoperculum, articular and the dentary. All its fibres are transversely arranged; there is a faint but distinct median aponeurosis extending along the entire length of the muscle.
Both the anterior and posterior intermandibularis are innervated by a branch of the mandibular V nerve.
Interhyoideus (text-fig. 10C). A bilaterally paired muscle originating on the hyoid arch but inserting on the medial face of each dentary and articular, is identified as the interhyoideus because of its origin (the epi- and ceratohyal) and because of its innervation (a branch of the hyohyoideus VII).
The lateral (7.e. outer) muscle of each pair is the larger; it has a narrow origin from the upper part of the ceratohyal and part of the epihyal, and inserts tendinously onto the ventral margin of the dentary immediately lateral to the symphysis.
The inner and smaller muscle has a much wider origin (entirely from the cerato- hyal) ventral to that of the lateral division. It inserts on the medial face of the dentary, and also the articular, considerably behind and above that of the outer division.
Hyohyoideus (text-fig. 10): is a stout, thick muscle with the left and nght halves meeting medially along a fine aponeurosis. The branchiostegal rays are almost completely embedded in muscle; it is impossible to recognize separate superior and inferior divisions. The posterior part of the hyohyoideus is attached to the sub- opercular branchiostegal rays,! the middle section to the interoperculum, and the forward part to the anterior face and ventral margin of the ceratohyal laterally; medially the muscle is attached to the urohyal. All these connections are musculose.
Externally, the hyohyoideus is covered by skin which is continuous with that of the body and lower surface of the jaw. In other words, there is no separate branchio- stegal membrane. Internally, the medial face of the hyohyoideus is bound to the overlying ventral body musculature by an extensive connective tissue union.
Sternohyoideus: is a large muscle originating mainly from the dorsal face of the horizontal cleithrum, but with a small ventral portion continuous with the hypaxial body muscles. The sternohyoid inserts on the urohyal but its antero-dorsal section is closely associated, through a connective tissue fascia, with the second basi- branchial. The ventral processes from the second hypobranchials are, proximally, embedded in the sternohyoid; distally they lie outside the muscle but closely attached to its lateral face.
Ventral gill arch musculature (text-fig. 11). Well-developed obliqui are present on the first three gill arches; all are broad muscles extending onto the ceratobranchial well beyond its articulation with the small hypobranchial. From the first obliquus there is a distinct postero-medially directed slip of muscle which inserts onto the second hypobranchial. From the second obliquus a similar slip runs to, and inserts on, the second basibranchial. The third obliquus links only the cerato- and hypo- branchial of its arch.
1In mormyrids the upper two branchiostegal rays are blade-like bones which have lost their articu- lation with the hyoid arch. Instead, the two bones are firmly attached to one another, and the upper bone is immovably attached to the lower margin of the operculum. Together, these two rays form an apparently immovable pseudosuboperculum.
26 P. H. GREENWOOD
A moderately developed rectus communis extends between the fourth cerato- branchial and the proximal end of the ventral process from the second hypobranchial. The muscle is closely applied to, but distinct from, the dorsal margin of the sterno- hyoideus. A short, broad and nearly triangular muscle joins the fourth and fifth ceratobranchials near their proximal ends. I cannot be certain about the identity of this muscle. It could be a displaced obliquus of the fourth arch or, more likely, part of this obliquus since there is a large transversus associated with the arch (all this assuming that obliqui are, primitively, associated with each arch; see Nelson, 1967).
The anterior transversus is moderately broad; it links the ventral tips of the fourth ceratobranchials. The posterior transversus runs obliquely forward so that it has the shape of a V, the apex inserting on the cartilaginous block lying between the ventral tips of the third and fourth ceratobranchials (the fourth basibranchial of Nelson, 1968). The arms of the V are closely applied to the ventral surface of each cerato- branchial.
HbI CbI Ob2 RC Ob3
BhTp AsCH HbIL SH LoS}
Fic. 11. Moymyrus kannume: ventral gill arch muscles and sternohyoideus in left lateral view. Not all the connective tissue surrounding the basihyal and urohyal has been dissected away.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 27
External and internal pharyngocleithrales are present, but are not strongly developed. An unusual feature of these muscles in all mormyrids is their origin external to the sternohyoideus Both divisions, but especially the external one, have a deep and broad insertion on the fifth ceratobranchial.
Mormyrus caschive Linn. (Text-fig. 2B)
Bishai i(1967) has given a detailed account of the cranial and branchial muscles in this species Unfortunately, I have been unable to dissect a specimen, but since in osteologcal and other features M. caschive is very similar to M. kannume, it seems reasonable to assume that the musculature would also be similar.
Thus, it is surprising to find a number of departures from the M. kannume condi- tion in the ventral musculature of M. cashive as described by Bishai. Because muscle nomenclature used by this author differs from that usually employed, and because of the anatomical differences noted, I shall give a list of synonyms for the muscles involved and also comment on the apparent discrepancies between my observations on M. kannume and Bishai’s on M. caschive.
In general, I would agree with Bishai’s description of the superficial hyoid muscles. His depressor labii inferioris muscle is my posterior intermandibularis, and his inter- mandibularis is my anterior intermandibularis. The posterior intermandibularis of M. caschive (cf. text-fig. 10A with Bishai’s fig. 4) is narrower anteriorly and seems to extend further posteriorly, covering the entire hyohyoideus (i.e. the interbranchio- stegalis muscle of Bishai).
There is also substantial agreement in the arrangement of the deep ventral muscles (Bishai’s geniohyoideus internus and externus, which are my inner and outer divisions of the interhyoideus). According to Bishai’s figure 5, there are three divisions of this muscle, the innermost of which is not labelled. However, I suspect that the object which Bishai has labelled “‘geniohyoideus externus”’ is, in fact, a cut section of skin (or even cheek muscle) ; further, it seems that his “internus’”’ muscle should be labelled ‘“‘externus”, and that the short unidentified innermost muscle is the internal division of his geniohyoid series (i.e. my interhyoid series).
The same figure also shows a condition of the sternohyoideus which I find difficult to accept. However, I think the figure is explicable when one considers Bishai’s description of the ventral branchial muscles (op. cit., page 21, and fig. 8). Here the author describes three large muscles, originating on each side from the cleithrum, and inserting on the urohyal and the ventral processes of the second hypobranchials. These muscles are identified by Bishai as ‘‘anterior portions of the pharyngoclavi- cularis muscle’. Apart from an abuse of the term pharyngoclavicularis for muscles with these topographical relationships, it seems that Bishai failed to recognize their true identity as parts of the sternohyoideus (see page 25).
Bishai’s pharyngocleithralis posterior apparently consists of both the external and internal divisions of this muscle.
28 P. H. GREENWOOD
I cannot find separate muscles in M. kannume corresponding to Bishai’s pharyngo- arcualis anterior, and obliquus ventralis anterior and posterior (of the first gill arch, see his fig. 8). Indeed, it seems that he has misinterpreted the double-headed condition of the first obliquus muscle which inserts, mainly, on the first hypo- branchial but also has a slip passing to the second hypobranchial (see page 25).
Bishai does not describe a rectus communis muscle but his fourth obliquus ventralis superioris could well be part of a rectus communis (v7z. that portion near its origin on the fourth arch and below the third arch).
Likewise, Bishai’s obliquus ventralis superioris 5 seems to correspond to the small muscle, present in most mormyrids, which links the ceratobranchials of the fourth and fifth arches (see page 26).
The anterior transversus (i.e., the fourth transversus ventralis of Bishai) requires no comment, but I believe that Bishai has misidentified the posterior transversus, and called it the pharyngoarcualis posterior (see page 51 for a discussion of the posterior transversus in mormyrids).
Mormyrus lacerda Casteln. (Text-fig. 2C)
In this species the snout is relatively short, broad and but slightly decurved; the mouth is broad and terminal in position.
With few exceptions, the musculature is like that of Mormyrus kannume.
The anterior intermandibularis inserts onto the dentary and angular only; over its posterior third it fails to reach the lateral margins of the head. However, its postero- lateral tips are attached to the ceratohyal near the origin of the interhyoideus.
The posterior intermandibularis, relative to that of M. kannwme, is somewhat less substantial over its posterior half.
The pharyngocleithrales are complex. There are three distinct but contiguous heads on the cleithrum, all originating lateral to the sternohyoideus. At about the midpoint between girdle and ceratobranchial, the three separate muscles fuse, become tendinous and then, as a single element, become muscular again. Presum- ably this muscle should be considered as fused external and internal pharyngo- cleithrales.
Mormyrus hasselquisti Val. (Text-fig. 2D)
In this species the snout is short, broad and but very slightly decurved. The mouth is broad and terminal.
The musculature of M. hasselquisti is virtually identical with that of M. lacerda (text-fig. 12). The anterior half of the posterior intermandibularis is, however, not quite so broad. It is separated from the ventral margin of the articular and dentary by a distinct connective tissue band through which the underlying anterior inter- mandibularis can be seen.
OSTEOGLOSSOMORPH HYOID MUSCULATURE
Fic. 12. Mormyrus hasselquisti: ventral hyoid muscles. A. Right half of posterior intermandibularis removed. B. Right half of the anterior intermandibularis removed.
30 P. H. GREENWOOD
Cyphomyrus discorhynchus (Peters) (Text-fig. 2E)
The snout in this species is short, broad and strongly decurved, with the small mouth situated subterminally and its opening directed somewhat ventrally.
The genus Cyphomyrus, once part of the large and probably artificial assemblage of species in the genus Marcusenius, was separated out by Myers (1960).
There are several small but none the less characteristic features in the musculature of C. discorhynchus.
The posterior intermandibularis (text-fig. 13) is well developed, but anteriorly it does not extend to the lateral margins of the lower jaw. That section of the muscle originating on the interoperculum has its fibres more closely packed than are those in the larger section originating from an aponeurosis over the hyohyoideus muscle. The two parts of the intermandibularis can be separated easily by gentle traction.
TIhy AIM eS
MC PIM Hhy
Fic. 13. Cyphomyrus discorhynchus : oblique ventro-lateral view of the hyoid musculature, after removal of the skin.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 31
The anterior intermandibularis (text-fig. 13) is a short muscle showing some individual variability in length. In only one of the three fishes examined does this muscle extend anteriorly to the level of the posterior intermandibularis insertion ; in one fish it extends posteriorly to the interoperculum (to which it is attached), in another it does not quite reach that point, whilst in the third fish, the muscle is attached to the interoperculum on one side but does not contact that bone on the other side. In all three specimens a slip from the posterior margin ot the muscle passes upwards on each side to insert on the anterior face of the ceratohyal.
The major part of the sternohyoideus originates, aponeurotically, from the ventral body musculature; only a small part of the muscle originates from the cleithrum.
The ventral gill arch musculature is unusual in having, in addition to a rectus communis, a short but discrete muscle extending between the ventro-lateral ends of the second and third ceratobranchials.
Marcusenius cyprinoides (Linn.). (Text-fig. 2F)
This species is usually classified in the genus Gnathonemus, but Taverne’s (1968) recent investigations have shown that it and other short-snouted species are more properly included in the genus Marcusenius Gill as revised by Taverne (of which M. cyprinoides is the type species).
The snout of M. cyprinoides is relatively short and broad, and is not strongly decurved. The small mouth is terminal in position but has its opening directed somewhat dorsally. There is a noticeable submental protruberence into which the lower lip passes insensibly.
The posterior intermandibularis has the typical mormyrid origin from the inter- operculum and from above the hyohyoideus. It inserts, however, entirely onto the enlarged submental cartilage; that is, it has no direct connection with the dentary.
The anterior intermandibularis, in contrast with that of the other species described above, is greatly reduced in size. Its length is only slightly greater than that of the eye, and it is confined to the articular region of the lower jaw. It inserts onto the articular of each side just before the articular-quadrate joint. The medial section of this muscle is tendinous.
The paired interhyoideus muscles, and the hypertrophied hyohyoideus are typical.
The sternohyoideus originates mainly from the dorsal surface of the horizontal limb of the cleithrum, but a small ventral part arises, aponeurotically, from the hypaxial body musculature. As in other mormyrid genera, the anterior part of the sterno- hyoid is closely associated with the ventral region ot the first two gill arches. The ventral processes from the second hypobranchials are completely embedded in the sternohyoideus.
The ventral gill arch muscles are typical (see under M. kannume) except that there is, apparently, no short, triangular muscle connecting the fourth and fifth cerato- branchials.
32 P. H. GREENWOOD
Marcusenius victoriae (Worthington) (Text-fig. 2G)
This species closely resembles M. cyprinoides (and like that species was formerly included in the genus Gnathonemus).
The musculature too is virtually identical with that of M. cyprinoides, although the anterior intermandibularis (text-fig. 14) is a little further reduced in size, and the muscle connecting the fourth and fifth ceratobranchials is present.
Gnathonemus longibarbis (Hilgendorf) (Text-fig. 2H)
The snout is moderately short (especially as compared with other Gnathonemus species; see Taverne [1968]), the mouth small and terminal, and there is a long tubular submental barbel.
The posterior intermandibularis is an extensive muscle, originating, as is usual, over the hyohyoideus and from the interopercula. It has strong connective tissue attachments (in which some muscle fibres occur) to the ventral margin of the
I Smm
Fic. 14. Marcusenius victoriae: ventral hyoid musculature. The right half of the posterior intermandibularis muscle removed to show overlying muscles.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 33
articular, dentary and preoperculum. <Anteriorly, the muscle almost completely surrounds the massive, elongate submental cartilage, and reaches to its distal tip. A few fibres attach to the dentary at the base of the cartilage.
There is no trace of a distinct anterior (i.e. transverse) intermandibularis, nor can it be detected as a component of the posterior intermandibularis.
The tnterhyoidei are well developed but the bellies of the medial and lateral muscles on each side are not readily separated. Both divisions have distinct origins and insertions (like those in the species where the bodies of the muscles are distinct).
The hyohyoideus is typical (see page 25).
Most of the sternohyoideus originates on the cleithrum, but a small ventral portion is continuous with the hypaxial body musculature. As in other mormyrid species, the muscle has an intimate association with the ventral elements of the first two gill arches ; the processes from the second hypobranchials are embedded proximally but lie on the surface of the sternohyoid distally.
The ventral gill arch muscles are typical (see page 25 et seq.).
Campylomormyrus elephas (Blgr.) (Text-fig. 21)
Previously placed in the genus Gnathonemus, this species has, on osteological grounds, been reclassified, with others, in the genus Campylomormyrus (see Taverne, 1968).
The snout of C. elephas is greatly elongate, is tubular and strongly decurved; the tip lies well below the ventral head profile. The mouth is small and, relative to the snout, terminal. Because of the snout’s decurvature the mouth opening is directed ventrally. The lower lip is continuous with a short, root-like mental barbel.
Posterior intermandibularis (text-fig. 15A and B): is a complex muscle, paired posteriorly, single anteriorly. The single portion extends from about the level of the quadrate-articular joint to the tip of the mental cartilage on which it is inserted. At no point along its length does this part of the muscle insert onto the dentary or articular.
The paired part is subdivided and has, on each side, several sites of origin. A little behind the point where the two halves of the muscle unite, each is divisible into more or less readily distinguishable dorsally and posterodorsally directed branches (text-fig. 15A). The area between these divisions is, however, traversed by a few interconnecting fibres. The dorsal branch is attached to the dentary and the articular while the posterodorsal branch is attached only to the interoperculum. The points of attachment should be considered as the sites of origin for the muscle as a whole.
Near the insertion of the posterodorsal branch (on the interoperculum) it is joined by a third division which originates, aponeurotically, over the hyohyoideus. This third branch, compared with the hinder part of the posterior intermandibularis in
34 P. H. GREENWOOD
Fic. 15. Campylomormyrus elephas:
PIM
Thyl
Thyi
Hhy
PIM
Smm j
A. Ventral hyoid muscles in lateral view after
removal of the skin. B. The same, in ventral view, with the posterior three quarters of the left half of the posterior intermandibularis removed, and the left lateral division of
the interhyoideus displaced laterally.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 35
other mormyrids, is rather narrow and does not join its counterpart, although the two approach one another closely in the midline (see text-fig. 15A).
The anterior intermandibularis is, apparently, absent.
The interhyoideus (text-fig. 15B) like that of Gnathonemus, is incompletely sub- divided (see page 33).
The hyohyoideus conforms with the typical pattern for mormyrids.
The sternohyoideus originates, about equally, from the cleithrum and the hypaxial body muscles. It inserts onto the urohyal and has close connections with the second basibranchial and other ventral elements of the first two gill arches. The processes from the second hypobranchials are embedded proximally; distally each lies superficially on, but attached to the sternohyoid flank.
The ventral gill arch muscles are typical (see page 25); the muscle linking the fourth and fifth ceratobranchials is present.
Petrocephalus bane (Lacép.) (Text-fig. 2J)
The snout is short and broad, the mouth broad and subterminal, and lies almost immediately below the eye.
The text-figures for P. catostoma are, in general, applicable to this species as well.
Posterior intermandibularis. The two halves of this muscle are narrowly separated medially by a fine aponeurosis. The muscle originates from the first two branchio- stegal rays and from above the hyohyoideus. It inserts, tendinously, along the entire ventral margin of the short lower jaw from symphysis to articular-quadrate joint. The uppermost fibres of this muscle separate easily from the ventral fibres, and in one specimen there appears to be some exchange of fibres between the posterior intermandibularis and the overlying interhyoideus muscle (see below). Close contact between these muscles is enhanced by the complete absence of an anterior intermandibularis.
Petrocephalus bane, unlike the species described before, has only one interhyoideus present on each side. The muscle, despite its great relative depth has a narrow origin on the dorsal part of the anterior face of the ceratohyal. In contrast, it has a deep and extensive insertion on the inner aspect of the dentary and articular, linearly from a point near the symphysis almost to the joint with the quadrate, and vertically from dorsal to ventral margins of these bones.
The hyohyoideus is well-developed and of the usual form, except that only a very small part is inserted onto the urohyal.
The sternohyoideus originates about equally on the cleithrum and from the hypaxial body muscles.
The ventral gill arch muscles are typical, with a small interceratobranchial muscle between the fourth and fifth arches (see page 25 et seq.).
P. H. GREENWOOD
ntral hyoid muscles. A. After removal of the skin. intermandibularis muscle.
oma: ve B. After removal of the posterior
Fic. 16. Petrocephalus catost
OSTEOGLOSSOMORPH HYOID MUSCULATURE 37
Petrocephalus catostoma (Giinther)
In its superficial morphology this species closely resembles P. bane; the myology is likewise very similar.
The posterior intermandibularis (text-fig. 16A) of P. catostoma has a more extensive origin over the hyohyoideus, and the tendinous part of the insertion is wider in this species. Some anterior fibres are almost transversely arranged (text-fig. 16A), thus contrasting with the obliquely orientated fibres posteriorly. Since no distinct anterior tntermandibularis can be located in this species either, it is tempting to consider these transverse fibres as remnants of an anterior muscle now fused with the posterior part.
As in P. bane, only one interhyoideus (text-fig. 16B) is present on each side. These muscles are completely free from the underlying intermandibularis.
In all other respects the musculature of the two species can be considered identical, and typical of the family.
Isichthys henryi Gill (Text-fig. 2K)
The broad snout of this species 1s relatively short and is not at all decurved. The mouth is terminal, horizontal and rather broad.
The posterior intermandibularis (text-fig. 17) of I. henryi shows very considerable reduction when compared with that muscle in all the species described above. The
PIM
Fic. 17. Isichthys henryi: ventral hyoid muscles, after removal of the skin.
38 P. H. GREENWOOD
muscle is reduced to a narrow V-shaped band of fibres with the apex directed anteriorly; its origin is entirely from the interoperculum, and its tendinous insertion is onto the small mandibular cartilage lying above the interdentary symphysis.
Overlying the narrow posterior intermandibularis is the broad anterior inter- mandibularis (text-fig. 17) which extends across the lower jaw from the level of the posterior interopercular margin to immediately behind the interdentary symphysis. The muscle has a largely musculose insertion onto the ventral margins of the dentary and articular; its fibres are separated medially by a posteriorly broad but anteriorly narrow aponeurosis. Over the anterior quarter of the muscle the fibres are arranged almost transversely, but elsewhere they run obliquely forward.
The hyohyoideus (text-fig. 17) shows typical hypertrophy, and there are distinct lateral and medial interhyoideus muscles on each side.
The sternohyoideus originates about equally from the cleithrum and, aponeurotic- ally, from the hypaxial body musculature. As usual, the sternohyoid is closely associated with the ventral elements of the first two gill arches. The ventral processes ot the second hypobranchials he superficial to the lateral surface of the sternohyoid, but are closely bound to it by connective tissue.
The ventral gill arch musculature conforms to the typical pattern (see page 25); the small muscle between the fourth and fitth ceratobranchials is not clearly differen- tiated and seems to be a slip from the anterior transversus. The posterior transversus is a narrow muscle but is otherwise typical.
Mormyrops anguilloides (Linn.) (Text-fig. 2L)
The snout is relatively short in M. anguilloides, and is not decurved. The mouth is wide and terminal.
A specimen of the closely related M. deliciosus (Leach) was also dissected and its musculature found to be identical with that of M. anguwilloides.
Posterior intermandibularis (text-fig. 18A). The muscle thought to be a posterior intermandibularis (see below) is a largely tendinous sheet within which lies a narrow band of transverse fibres arranged on either side of a median aponeurosis. Pos- teriorly, a short, upwardly directed slip ot muscle runs to each interopercle; there is also a broader posterior extension of the muscular part which ends, aponeurotically, above the hyohyoideus (text-fig. I8A). Fibres in this part of the muscle are longi- tudinally orientated. The attachments to the interoperculum and the aponeurotic connection over the hyohyoideus should, presumably, be considered as the origins of the muscle; its insertion is directly onto the dentary near and on the symphysis (there being no mental cartilage in this species). The lateral and tendinous part of the muscle is attached to the interopercula, preopercula, the articulars and the dentaries,
OSTEOGLOSSOMORPH HYOID MUSCULATURE 39
Ihyi Ihy| / Hh iy y H hya Smm
Fic. 18. Mormyvops anguilloides: oblique ventro-lateral view of the ventral hyoid muscles. A. After removal of skin. B. After removal of posterior intermandibularis muscle C. As in B but with lateral division of left interhyoideus reflected to show the inner division of that muscle.
40 Pi Ho GREENWOOD
No discrete anterior intermandibularis is present. The possibility of this muscle having fused with the posterior division cannot be overruled because the muscle thought to be the posterior intermandibularis is fairly clearly demarcated into an anterior part with transverse fibres, and a posterior sector with longitudinal fibres (see above). If the posterior and anterior intermandibulares are fused, then Mormyrops presents a rare condition among the mormyrids (but see also Petro- cephalus catostoma).
The interhyoideus in Mormyrops also exhibits certain peculiarities (text-fig. 18B, C). As is usual, the lateral muscle of each pair is the larger. It originated dorsally on the anterior face of the ceratohyal, but it has two insertions, one directly onto the inner and ventral aspects of the dentary and articular, the other, ten- dinously, onto the anteroventral face of the dentary immediately lateral to the symphysis.
The smaller median muscle also originates on the anterior face of the ceratohyal but it inserts, aponeurotically, onto the median aspect of the lateral muscle. In other words, it has no direct attachment to the lower jaw. Indeed, the separation between the two muscles is very slight even near their origins and there is an exchange of fibres between them (a condition reminiscent of that in Gnathonemus and Campylomormuyrus [see pages 33 and 35 respectively)).
The hyohyoideus (text-fig. 18) shows typical hypertrophy and relationships with other structures (see page 25).
Most of the sternohyoideus originates as an extension of the ventral body muscula- ture; only a small part stems from the cleithrum. The processes from the second hypobranchials are fairly deeply embedded in muscle.
The ventral gill arch muscles are typical (see page 25), but the small muscle between the fourth and fifth ceratobranchials appears to be missing, and the slip from the first obliquus is greatly reduced in size.
The pharyngocleithrales are moderately well-developed, especially the external division.
Hyperopisus bebe (Lacép.) (Text-fig. 2M)
The snout is relatively short and not decurved, the mouth moderately broad, terminal in position and horizontally directed.
Hyperopisus bebe is unique among mormyrids in having greatly enlarged and molariform teeth on the parasphenoid and apposing basibranchial tooth plate. Not only are the teeth enlarged, but they occupy a much broader area of attachment in both places than is usual among species with small conical teeth (see figs. in Taverne, 1968).
The posterior intermandibularis (text-fig. 19gA) shows a degree of reduction com- parable with that in Jsichthys. From the interoperculum of each side a moderately
OSTEOGLOSSOMORPH HYOID MUSCULATURE 41
developed, strap-like muscle extends forward to insert on the small mental cartilage. Dorsally, the muscle has no contact with the ventral margin of the lower jaw. Between the arms of the strap-like muscle there is a broad tendinous sheet of tissue extending back to about the level of the posterior interopercular margin. Here it becomes continuous with the tendinous part of the overlying anterior intermandi- bularis. The margin of the tendon passes insensibly into a number of weak muscle- fibre bundles which, in turn have an aponeurotic connection with the hyohyoideus (see text-fig. I9A).
Immediately above the posterior intermandibularis is the well-defined and thick anterior intermandibularis. Except for a short distance posteriorly there is no median aponeurosis in this muscle, whose fibres extend from side to side of the lower jaw (text-fig. 1I9B). Insertion is onto the ventral margins of the dentary, articular, preoperculum and interoperculum, and, anteriorly, onto the mental cartilage. The anterior intermandibularis extends from the level of the vertical preopercular arm to a little behind the interdentary symphysis.
The interhyoideus muscles have a characteristic and complex form in Hyperopisus (text-fig. 1gC-D). The ventro-lateral muscle of each pair is the smaller and is largely tendinous (see text-fig. 1gC-D). It originates on the epihyal (with a very small part from the head of the fourth branchiostegal ray), and runs straight forward to insert on a low ventral process of the dentary.
The inner and dorsal muscle is divided, from near its origin, into two elements, one directed horizontally, the other (the innermost) running laterally and anteriorly (text-fig. 1gC-D). The medial subdivision is the smaller of the two. It arises from some of the dorsal and external fibres of the lateral subdivision near the upper part of its origin. From a narrow beginning, the medial subdivision fans out so that it comes to extend along the ventral margin of the urohyal (anterior, that is, to the insertion of the hyohyoideus). The left and right medial subdivisions meet in a fairly narrow aponeurosis over the urohyal; some fibres from the underlying anterior intermandibularis also are attached to this tendinous tissue. Anteriorly, the aponeurosis is extended as a short tendon which inserts onto the dentary immediately below the symphysis.
The lateral subdivision of the inner interhyoideus has a broad origin from almost the entire anterior face of the ceratohyal. Its fibres run outwards and forwards, crossing above those of the outer division. The muscle inserts, either directly or through a narrow tendinous margin, onto the inner aspect of the quadrate, articular, and anteriorly, the dentary (text-fig. r9C-D).
The hyohyoideus (text-fig. 1gC-D), although well-developed, gives the impression of being relatively less massive in this species. The anterior section is not as thick as in the other species described, and the interbranchiostegal portions are also weaker. As usual, the hyohyoideus inserts on the interoperculum, ceratohyal and urohyal, the latter insertion being relatively more extensive than in most of the other mormyrids examined.
The sternohyoideus is a short, deep muscle whose origin is entirely from the dorsal face of the cleithrum. The processes from the second hypobranchials curve out-
42 P. H. GREENWOOD
PIMm
PIMt
IOp
SOpBr
AIM
Apn
easy
Fic. 19. Hyperopisus bebe; ventral hyoid muscles. A. After removal of skin. B. After removal of posterior intermandibularis muscle and tendon complex. C. After removal of anterior intermandibularis muscle. D. Right ventro-lateral division of the inter- hyoideus muscle removed to show underlying innermost division of this muscle complex. Entire inner (medial) division of this complex also removed.
OSTEOGLOSSOMORPH HYOID MUSCULATURE
G Ihyvl Hhy D Ihyim Ihyvl POp IOp Hh ¥ SOpBr
43
44 P. H. GREENWOOD
wards and slightly backwards so as to lie almost horizontally ; each lies superficial to the sternohyoid but is closely bound to that muscle.
The ventral gill arch musculature is like that of Mormyrus (see page 25), except that there is no muscle linking the fourth and fifth ceratobranchials. The external pharyngocleithvalis has a short tendinous section at about the middle of its visible length, and the entire external division is virtually tendinous.
Family GYMNARCHIDAE Gymnarchus niloticus Cuvier (Text-fig. 2N)
The snout in Gymnarchus niloticus is relatively elongate, moderately broad and is not decurved. The terminal mouth is horizontally placed and its cleft extends much further posteriorly than in any mormyrid species. Gymnarchus is also unusual in having no teeth on the parasphenoid, and in lacking basibranchial tooth plates (see Nelson, 1968 and Taverne, 1970).
Holmquist (1g11) has given a fairly detailed account of certain muscles; his description and figures were later used by Edgeworth (1935). Neither author was able to carry out comparative studies on other mormyroids and it is thus under- standable how they came to emphasize the apparent similarities between Gymnarchus and Albula, in particular the hyperdevelopment of the transverse intermandibularis muscle.
Although these similarities do exist, and although Gymnarchus departs from the typical mormyrid condition, the similarities between it and Albula are less than those shared with the Mormyridae. In particular, one may note the shared specializations in the ventral gill arch musculature and the sternohyoideus.
The hyoid musculature (text-fig. 20). Immediately below the skin, and closely adherent to it, there is a broad, generally thick sheet of muscle which thins out posteriorly. It is attached to the operculum, sub- and interopercula, the preoper- culum, dentary and articular. Anteriorly, this muscle inserts around the ventral margin of the mandibular arcade; posteriorly it extends to the level of the branchial opening (text-fig. 20). Throughout its length the fibres of the muscle are trans- versely arranged and extend from side to side without any indication of a median aponeurosis. Innervation is from a branch of the mandibular V nerve.
Holmquist (0p. cit.) identifies this muscle as an intermandibularis; from the transverse arrangement of its fibres (text-fig. 20) it would seem to be homologous with the anterior intermandibularis of mormyrids. Since there is no trace of any muscle ventral to this sheet, and because there are no indications of longitudinal fibres in the sheet, one must conclude that a posterior intermandibularis is absent.
Immediately above the intermandibularis is a pair of muscles that insert, anteriorly, on the inner aspect of the dentary on either side of the symphysis (text-
OSTEOGLOSSOMORPH HYOID MUSCULATURE 45
fig. 20). Neither muscle meets the other at any point, although they are closely aligned anteriorly. The muscles originate in part from the epi- and ceratohyal, and in part from the first branchiostegal ray. In large individuals each muscle is incompletely divided, horizontally, so as to virtually form a small dorsal and larger ventral muscle (text-fig. 20). The ventral subdivision is that described above; the dorsal part originates on the epihyal and inserts on the inner face of the dentary about halfway along its length. The smallest fish examined (14:0 cm standard length) has no division, or even incipient division, in the muscle; consequently the single muscle on each side appears to have a double insertion, one anteriorly and the other laterally at about the middle of the dentary.
These muscles (or muscle) correspond, in most details, to the interhyoideus of mormyrids, and are identified as such in Gymnarchus.
One difference in the interhyoid of Gymnarchus is the considerable proportion of muscle originating from the first branchiostegal ray (text-fig. 20). In mormyrids the interhyoid originates exclusively from the hyoid bar, and there is complete separation between the interhyoid and the hyohyoideus. The situation in Gymnar- chus is very different. In fact, the muscle described above could well be a combined inter- and hyohyoideus. In Gymnarchus the only clearly recognizable hyohyoideus is the superior or interbranchiostegal ray portion (text-fig. 20) ; here it is represented by a few, well-spaced oblique fibres running between the branchiostegal rays. No distinct inferior hyohyoid can be recognized. Indeed, those fibres, which, topo-
Pic. 20. Gymmnarchus niloticus: ventral hyoid musculature. Anterior intermandibularis cut and reflected to the right; left interhyoideus reflected to the left.
46 P. H. GREENWOOD
graphically, would form the inferior hyohyoid merge with the fibres forming the body of the muscle I have identified as the interhyoideus. Thus, the situation in Gymnarchus has all the appearance of a secondary return to the primitive condition in which the constrictor hyoideus ventralis has not split into inter- and hyohyoid portions. That the condition is not strictly a primitive one is indicated by the separate left and right inter- cum-hyohyoideus muscles.
The poorly developed hyohyoideus of Gymnarchus contrasts markedly with the hypertrophy of that muscle in all mormyrids. Gymnarchus also differs in having the branchiostegal rays (and their associated musculature) free from the ventral body muscles. Nevertheless, the branchiostegal membrane (i.e. the interray muscles and tendons) still does not provide a ventrolateral floor to the branchial cavity. This is formed by a membranous sheet which runs, on each side, from the inner face of the corresponding ceratohyal to the upper part of the sternohyoid muscle. In effect, each branchiostegal membrane constitutes a lateral half pouch (opening medially but blind anteriorly) over the sternohyoid and hypaxial muscles. The intermandi- bularis muscle covers these pouches and their medial openings. This greater pouch is open posteriorly across the breadth of the body.
Identifying the hyoid muscles of Gymnarchus is further complicated by the presence of a pair of small muscles that lie immediately anterior to the sternohyoid
Uh Hp SH l
Fic.21. Gymnarchus niloticus; Sternohyoid and ceratohyal-urohyal muscles in left lateral view; semi-schematic.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 47
insertion (text-figs 20 & 21). Each muscle has its fibres arranged almost vertically, and 1s attached to the ventral margin of the ceratohyal above and to the urohyal below. It seems likely that these are the muscles which Holmquist (op. czt.) showed in his figure and which he tentatively identified as branchio-mandibulares (labelled :? genio-branchialis in Edgeworth’s [1935] reproduction of Holmquist’s figure). However, Holmquist does clearly state that the fibres run longitudinally.
Holmquist (op. cit.) commented that he had not seen such a muscle in any other teleosts, and neither have later workers described similar muscular arrangement.
Holmquist’s and Edgeworth’s identification of this muscle are tentative, but I have difficulty in understanding why they considered it to be a branchiomandi- bularis or geniobranchialis. Such a muscle is not found in teleosts, and morpho- logically it is one connecting the lower jaw with the hyobranchial skeleton. Be that as it may, the identification of the muscle pair still poses several questions. That each is attached to both the ceratobranchial and the urohyal suggests one of three things: that the muscle could be a subdivision of the interhyoideus (see Hyperopisus, p. 41), could be part of the sternohyoideus (see Avapaima, p. 15 and Papyrocranus, p. 20), or could be a separated segment of the hyohyoideus. Without ontogenetic studies it seems unlikely that an answer can be provided.
The sternohyoideus (text-figs 20 & 21) is a large muscle originating almost entirely from the dorsal face of the cleithrum, but with a few ventral fibres contributed by the hypaxial body musculature. Each ventral process from the second hypo- branchial is very large, stout, and orientated so as to lie horizontally along and within the upper border of the sternohyoid (see fig. 21). Because the cleithrum is almost vertical in position and without a forwardly directed arm, the ventral processes form the ventral margins to the branchial chamber; in mormyrids the dorsal surface of the horizontal arm of each cleithrum forms the margin. The great size and peculiar orientation of the ventral processes in Gymnarchus appears to be unique within the Mormyriformes.
Ventral gill arch muscles in Gymnarchus conform to the usual mormyrid pattern but show several modifications doubtless associated with the reduced branchial skeleton in this genus (Taverne, 1970; Nelson, 1968).
The obliquus of the first gill arch is a stout muscle with a double insertion onto the anterior and posteroventral margins of the ceratobranchial; its origin covers most of the ventral surface of the hypobranchial. The second obliquus is a much smaller muscle. There is no obliquus muscle on the third arch, whose ceratobranchial is closely associated with the rectus communis muscle linking the fourth ceratobranchial with the base of the ventral process from the second hypobranchial.
The pharyngocleithrales are relatively stout muscles which can only be seen after deep dissection because they originate on the medial side of the sternohyoid (cf. mormyrids where the origin is lateral to that muscle). The near vertical cleithrum also effects the alignment of these muscles which thus originate behind the fifth ceratobranchial.
The internal muscle of each pair is the larger and longer; it runs obliquely forward and upwards to insert on the ventral tip of the ceratobranchial. The external
48 P. H. GREENWOOD
muscle is flatter and thinner, and runs almost vertically upwards to insert on the ventral face of the bone near its median edge and posterior border.
The tvansversus of the fourth arch is a fairly broad muscle linking the ventral tips of the ceratobranchials. The transversus of the fifth arch, however, joins the cerato- branchials near their posterior (i.e. upper) borders, and is in close contact with the external pharyngocleithrales near their points of insertion. This arrangement contrasts with that in the Mormyridae, where the fifth transversus runs forwards and medially as a V-shaped muscle to attach to the cartilaginous basibranchial plate between the fourth and fifth arches.
SUMMARY AND DISCUSSION
For the purpose of this discussion, the term ventral hyoid musculature includes the anterior and posterior intermandibulares, and the interhyoideus muscle (with its subdivisions), either as separate entities or with some elements partly fused to form a protractor hyoideus (see Introduction, page 4). My use of the term protractor hyoideus differs somewhat from that employed by Edgeworth (1935, p. 101). This implies neither criticism nor acceptance of Edgeworth’s concepts, and is used merely for brevity’s sake (but see page 49 below).
Among the Osteoglossomorpha, two principal types of ventral hyoid musculature can be recognized. In one there is a protractor hyoideus (Hiodontidae, Osteo- glossidae and Pantodontidae), while in the other the posterior intermandibularis and the interhyoideus muscles are quite separate (Notopteridae and Mormyridae). Gymnarchus could, perhaps, be considered as constituting a third type, but I would prefer to consider its pattern as a modified mormyrid type (see below).
The pattern shown by Hiodon is the most unspecialized one, and compares closely with that of the holostean Ama and such primitive teleosts as Elops and Megalops. Compared with the protractor hyoideus in these fishes, that of Hzodon is a little more specialized because the component muscles show a greater degree of consolidation and unification (see page 9).
Apart from some slight intergeneric variations, the protractor hyoideus in the Osteoglossidae and Pantodontidae presents a uniform picture. Compared with Hiodon, consolidation of component muscles is carried even further. The compound muscle is, indeed, like that occurring in species of several euteleostean groups (e.g. Salmo, Abramis and Peristedion: see Edgeworth, 1935). Edgeworth (op. cit.), based on Holmquist’s (1911) studies, interprets the ventral hyoid muscles in Osteo- glossum as having the posterior intermandibularis “ . attached by intersection to the inner part of the Interhyoideus, so that externally there is an Interhyoideus which reaches the jaws and internally a Protractor hyoidei’. This description certainly fits the condition found in Pantodon (see page 16), but in the other osteo- glossoid genera (including Osteoglosswm) I can find no such clear-cut division into ventrally placed posterior intermandibularis and dorsal interhyoideus (see page 12). The impression gained is one of greater fusion between the two muscles (if indeed
OSTEOGLOSSOMORPH HYOID MUSCULATURE 19
they are arranged in a dorsoventral series). From adult anatomy it seems that an anteroposterior fusion is just as probable, the forward section being derived from the posterior intermandibularis muscle, the hinder part from the interhyoideus (see page 12). In other words, a condition fulfilling Edgeworth’s definition of“ ... a fully developed Protractor hyoidei . . . —a longitudinal muscle the anterior part of which is formed by the Intermandibularis posterior and the posterior part of the Interhyoideus’’.1
The unconsolidated ventral hyoid muscles in notopterid and mormyrid fishes represent a different line of specialization, and one apparently otherwise seen only in siluroid ostariophysans (see pages 18 & 25).
In both notopterids and mormyrids the most superficial muscle is the posterior intermandibularis. The Gymnarchidae have lost the posterior intermandibularis, and thus the superficial muscle is probably an expanded anterior intermandibularis (see page 44). In all three families the interhyoideus component is a deep muscle, relatively small in notopterids but enlarged (and generally subdivided) in mormyrids and gymnarchids. There is also a difference in the way in which the interhyoideus inserts. Notopterids have the insertion on the hypohyals, but in the two other families the muscle is attached to the medial face of the lower jaw.
There is little intergeneric variation in the hyoid musculature of notopterids but a considerable amount within the mormyrids, involving all elements, including the anterior intermandibularis muscle.
Recent studies on the anatomy and osteology of mormyrid and gymnarchid fishes all indicate that species of the genus Petrocephalus are probably the least specialized (see Nelson [1968] on branchial arches; Taverne [1968, 1969 and 1970] on osteology, and Orts [1967] on visceral and auditory anatomy). Petrocephalus is also outstand- ing for being the only mormyrid in which the interhyoid of each side remains un- divided. Of the other ventral muscles in Petrocephalus, the posterior intermandi- bularis is well-developed and expansive, but the anterior intermandibularis is absent (or, possibly, fused with the posterior muscle; see page 37). The anterior intermandibularis is present in other genera with relatively unspecialized cranial characters, for example Marcusenius which closely resembles Petrocephalus in many respects. Thus, the absence of an intermandibularis in Petrocephalus may be a specialization.
Equally, the broad and expansive anterior intermandibularis found in Mormyrus, Hyperopisus, Isichthys, and Gymnarchus is probably a derived condition, as would be the lack of this muscle in Gnathonemus and Campylomormyrus, genera showing great specialization in jaw and snout form. The peculiar condition of the intermandi- bularis muscle (or muscles) in Mormyrops (page 38) is difficult to interpret, and is unlike either the Mormyrus-Marcusenius or the Gnathonemus-Campylomormyrus
1Jf such definite ontogenetic differences exist in the way a ‘‘protractor hyoideus’’ is formed, then there would be every justification for recognizing the end products by different names.
I am less impressed by the validity of Edgeworth’s (op. cit.) two subdivisions for those fishes without a fully formed (sensu Edgeworth) protractor hyoideus (see Edgeworth, op. cit., pp. 1oo—-101). For instance, the differences described by Edgeworth between Perca (his division I) and Osteoglossum (divi- sion 2) seem to be more a difference of degree than of fundamental organisation.
50 Pp. H. GREENWOOD
types. The relatively short anterior intermandibularis of Marcusenius and Cypho- myrus should probably be taken as the least specialized mormyrid condition. Nevertheless, this type still represents a marked departure from the basic teleostean condition (as in Hiodon) or even that in the generality of the euteleosts.
The posterior intermandibularis is, in most mormyrid genera, well-developed and expansive. Although this form represents a derived condition as compared with the presumed basal teleostean type (e.g. Hiodon and Elops), it should probably be con- sidered the primitive condition for mormyrids. Specialization by reduction of this state is seen in Mormyrops, Isichthys and Hyperopisus. This shared specialization cannot alone be taken to imply any phyletic relationship between these genera, especially since the end-product in Mormyrops is unlike that of the other two genera. Different specializations of the posterior intermandibularis are seen in Gnathonemus (extension in connection with the hypertrophied mental barbel, see page 32) and Campylomormyrus (multiple sites of origin, presumably correlated with extreme snout elongation and decurvature; see page 33).
The complete absence of a posterior intermandibularis in Gymnarchus is unique among mormyroid fishes, and is associated with a number of other muscular specializa- tions in the hyoid and branchial systems (see page 44). The hyoid and gill arch musculature in this genus are so specialized that they provide few clear-cut phyletic pointers. About all that can be said regarding the relationships of Gymnarchus is that the presence in the skull of lateral ethmoids, paired orbitosphenoids, and a basisphenoid (see Taverne, 1970), together with the relatively unspecialized inner ear (Orts, 1967), all point to derivation from the Marcusenius—Petrocephalus— Mormyrus assemblage rather than the Gnathonemus—Isichthys-Mormyrops group.
The sum of characters, both specialized and generalized in Gynimarchus suggest that it was a fairly early departure from the main mormyroid stem. Its retention in a distinct, monotypic family is probably not justified on phyletic grounds; its placement in a subfamily (Gymnarchinae) of the Mormyridae, however, would more accurately reflect its relationships.
The ventral branchial muscles in the Osteoglossomorpha are, on the criteria discussed by Nelson (1967), moderately specialized, being comparable with those of Polymixia in his series Elops—Aulopus—Polymixia—E pinephalus. The presence of obliqui inferiores in the dorsal branchial musculature of mormyrids and osteo- glossids is, according to Nelson (1969), a secondary and advanced feature not present in any member of his Elops—Epinephalus series.
In Nelson’s argument, the development of a rectus communis (and the degree of its antero-posterior extent) is taken to be indicative of specialization. This muscle in Osteoglossomorpha provides something of a puzzle. It is developed in all mem- bers of the superorder except in three of the five osteoglossid genera, namely Osteo- glossum, Scleropages, and Arapaima.1_ When present the muscle extends from the fourth arch to the second (attaching to the basibranchial in Hiodon but to the hypo-
1Hence, because these were the species he examined, Nelson’s (1969) observation that a rectus com- munis is not developed in Osteoglossidae; it is, on the other hand, well-developed in Heterotis niloticus and Pantodon buchholzi.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 51
branchial in all others). In terms of its volume, the muscle is largest in Hiodon although its anterior third is entirely tendinous.
Apart from the large rectus communis, the ventral branchial muscles in Hzodon show no outstanding characters. As in the notopterids (except Nenomystus) and osteoglossids, Hiodon has a rectus muscle between the third ceratobranchial and the second hypobranchial, and the well-developed pharyngocleithrales originate medial to the sternohyoideus.
An interesting feature of the notopterids is the way in which the heads of the ventral obliqui muscles on the first two arches attach to both the basibranchials and hypobranchials. The same muscles in mormyrids also have a complex origin, the first attaching to the second hypobranchial as well as to the first, and the second obliquus extending below the hypobranchial of that arch to reach the second basi- branchial.
No rectus muscle is developed in mormyrids, but in one species (Cyphomyrus discorhynchus) a stout muscle links the ventral ends of the second and third cerato- branchials. Another peculiar muscle, present in most mormyrids (but not in Mormyrops, Hyperopisus or Gymnarchus, and weak in Isichthys) extends between the ventral ends of the fourth and fifth ceratobranchials (see page 26). This muscle could be a segment of the anterior transversus muscle. The posterior transversus in all mormyrids except Gymnarchus (page 48) is V-shaped with its apex attached to the cartilaginous fourth basibranchial.
The mormyrid pharyngocleithrales are, in general, weak and often tendinous muscles with a deep insertion onto the fifth ceratobranchial, and an origin lateral to the sternohyoideus. Gymmnarchus, by contrast, conforms with the more usual teleost condition in which the muscles originate lateral to the sternohyoid.
All Osteoglossomorpha, with the exception of Hiodon, have a pair of ventrally directed bony processes associated with the ventral end of the second gill arches (see Greenwood et al., 1966 and Nelson, 1968). The processes are closely associated with the sternohyoideus muscle to which they are closely attached or in which they are embedded.
The sternohyoid of Hiodon is particularly interesting in this regard, and suggests a way by which the typical osteoglossomorph hypobranchial process might have evolved. In Hiodon the main body of the sternohyoid arises, aponeurotically, on the hypaxial body muscles and is ventral to the cleithrum. A small dorsal part of the muscle on each side, however, originates from the cleithrum. Like the ventral part, it inserts of the urohyal, but from its middorsal region a broad slip of near- tendinous muscle runs forwards and inserts on the second basibranchial (fig. 4). The possible significance of this unusually discrete connection is best appreciated when one recalls the ventral processes (or tendon bones) in notopterids. In these fishes the tendon bones articulate with the second basibranchial (and not the hypo- branchial as in other osteoglosomorphs), and are closely attached to the dorsal part of the sternohyoid.
The sternohyoid in Notopteridae originates entirely from the cleithrum. Its insertion is unusual since it has tendinous connections both with the hypohyals and
52 P. H. GREENWOOD
with the ceratohyals; furthermore, the muscle completely surrounds the small urohyal. Since the urohyal has strong ligamentous connections with the basi- branchial tooth plate, the sternohyoid has connections, albeit indirect, with that bone as well.
Among the osteoglossids there is some fairly marked variation in sternohyoideus relationships. In all genera, however, the ventral hypobranchial processes are firmly attached to the muscle, which is also closely associated, by connective tissue fascia, with the first hypobranchials.. Avapaima gigas has a complex sternohyoid (see page 15), subdivided into paired anterodorsal and unpaired ventral portions, the former inserting, through a common tendon, mainly onto the left ceratohyal. It is interesting to note that Heterotis (the other member of the subfamily Hetero- tinae) shows incipient longitudinal division of the sternohyoid anterodorsally (see page 14).
The mormyrid sternohyoideus shows little intergeneric variation, except in the relative proportions of the muscle originating from the cleithrum and from the hypaxial body musculature. All genera have a close association, through mem- branous fascia, between the muscle and the second basibranchial, and with the ventromedial elements of the first two gill arches (including, of course, the bony processes of the second hypobranchials).
The most outstanding feature of the sternohyoid in Gymmnarchus is its relationship with the hypertrophied ventral processes. These elongate and robust bones are no longer ventrally directed, but lie horizontally and are embedded in the dorsolateral margin of the muscle (see page 47). The peculiar ceratohyal-urohyal muscle in Gymnarchus (see page 46) may be another specialization of the sternohyoid, but the relationships of the two muscles are still far from clear.
The hyohyoideus muscles in all Osteoglossomorpha, except the Mormyridae, conform to the usual teleost pattern.
The Mormyridae are outstanding for the extent to which the hyohyoidei are hypertrophied and so arranged that the ‘“branchiostegal membrane’’ is virtually immovable (see page 23). The branchiostegal rays are deeply embedded in the thick superior hyohyodei which, in turn, are aponeurotically connected in the mid- line and are broadly inserted along almost the entire length of the ceratohyals. Because the inner aspect of each “‘branchiostegal membrane” is attached to the skin covering the ventral body musculature above and medial to them, they completely occlude the ventral opening to the peribranchial chamber. As a result, this opening is greatly reduced in its vertical extent.
Although it is not possible to carry out a reliable functional analysis on the basis of morbid anatomy, the arrangement of the hyohyoideus and sternohyoideus muscles in mormyrids suggests the development of a strong branchial pump. Asa means of ingesting small prey, such a device would accord well with the weak jaws, small mouth and “‘parasphenoid-basihyal bite’ of these fishes (see Nelson, 1968).
As with other myological features, the hyohyoideus muscles in Gymnarchus depart markedly from the typical mormyrid type (page 46). In this genus, the branchio- stegal membrane is free from the overlying skin of the body, and the hyohyoidei are
St ee
OSTEOGLOSSOMORPH HYOID MUSCULATURE 53
reduced to a few fibres between the branchiostegal rays. Furthermore, these muscles are continuous with the more dorsally located interhyoidei (see page 45). Since Gymnarchus is the only mormyrid lacking parasphenoid teeth and tooth-bearing dermal plates on the basibranchials, it is very likely that the differences in muscula- ture are associated with different feeding methods. That the gape of the mouth is relatively much larger than in the Mormyrinae, may also be significant.
Phyletic relations within the swperorder. The ventral gill arch musculature provides little information on this subject. The presence of a rectus communis in some osteoglossoids (see page 18) rather negates the importance of this character in the analysis of gill musculature given by Nelson (1969). Nelson, on the basis of evidence then available noted the presence of this muscle in notopterids and mormyrids, and its absence in osteoglossids,
The hyoid musculature, on the other hand, allows a clear-cut division to be made into fishes with a protractor hyoideus (Osteoglossidae and Pantodontidae), and those in which the posterior intermandibularis and the interhyoideus are distinct muscles (Mormyridae [including Gymnarchus| and Notopteridae). The Hiodontidae have a primitive teleostean type of hyoid musculature with respect to which both other types must be considered specialized, albeit along different lines. The osteoglossoid specialization is a common one among teleosts (see Holmquist, rg11; Dietz, 1912; Takahasi, 1925; and Edgeworth, 1935). The notopterid-mormyrid type has other- wise been recorded only in the catfishes (Siluroidei, Ostariophysi).! This siluroid type (Holmquist, 1911; Takahasi, op. cit.; Munshi, 1960; also personal observations on Ictalurus nebulosus, Parasilurus aristotelensis and Bagrus docmac) is basically like that of the notoperids, but its specializations are of a type not found in the latter fishes (or the mormyrids either) since they are associated with the mandibular barbels. Because, on so many characters, the siluroids are not allied to the osteo- glossomorphs, the resemblance in hyoid musculature can only be convergent.
Convergence is an unlikely explanation for the similarities between notopterid and mormyrid musculature; the two groups are related in so many other characters (Greenwood et al, 1966; Nelson, 1968 and 1969).
The relationships of the mormyrid fishes within the Osteoglossomorpha are still uncertain. Greenwood et al., (op. cit.), give the Mormyridae ordinal status co- ordinate with all the other osteoglossomorphs (including the Hiodontidae). McAllister (1968) preferred subordinal status, coordinate with the Osteoglossoidei and Notopteroidei, in an order (Mormyriformes) which excluded the Hiodontidae. In a paper reviewing gill arch skeletons within the Osteoglossomorpha, Nelson (1968) pointed out that most authors have given undue emphasis to certain mormyrid characteristics (especially the brain and electric organs). As a result of this stress on differences, the phyletic relations, he thought, are obscured by the hierarchical categories chosen for the Mormyridae. The gill arch studies made by Nelson led him to rearrange the osteoglossomorphs into two suborders in an order, the Osteo-
1Munshi’s (1960) description of the clupeoid Hilsa ilisha gives the impression that, in this species too, the posterior intermandibularis is the principal (and most superficial) hyoid muscle. The interhyoideus, he notes, is absent. I have dissected a specimen of this species and find that, like other clupeo ids (see Kirchhoff, 1958) the interhyoideus and posterior intermandibularis are fused to forma protractor hyoideus.
’ 54 P. H. GREENWOOD
glossiformes. In one suborder (Notopteroidei) he placed those families in which there are either no ventral processes on the second gill arch (Hiodontidae) or the processes are tendon bones articulating with the second basibranchials (Notopter- idae). In the other suborder (Mormyroidei) Nelson placed those families in which the processes are bony and fused with the second hypobranchials (Mormyridae, Gymnarchidae, Pantodontidae and Osteoglossidae). Nelson’s later studies (1969) on osteoglossomorph infraorbital bones (and some other features) gave no grounds for altering his earlier classification.
I agree with Nelson’s (1968) comments on the undue weight given to certain mormyrid characters, but I consider that other evidence indicates closer relationship between the notopterids and mormyrids than between the latter family and the osteoglossids. This evidence (some yet unpublished) is concerned mainly with specializations of the inner ear shared by the notopterids and mormyrids. A specialized inner ear also occurs in the hiodontids (see Greenwood, 1970; and unpublished). All three groups have an otophysic connection (of varying complex- ity) but none exists in the osteoglossids. Other indicators of relationship between the notopterids and hiodontids are discussed by Nelson (1969, p. 27).
The ventral hyoid musculature of Hiodon is too unspecialized to be of value as a phyletic indicator, but the shared specializations of this musculature in mormyrids and notopterids (see above, page 53) seems to reinforce other characters suggesting a relationship between these families. The very different specialization shown by the hyoid muscles of the Osteoglossidae and Pantodontidae implies that these families together represent a distinct lineage.
Until I have completed certain other studies on osteoglossomorph anatomy it would be premature to put forward a revised formal classification of the superorder. However, it seems very likely that any new classification will group (within a single order) the Hiodontidae, Mormyridae and Notopteridae in one category (probably subordinal) and the Osteoglossidae and Pantodontidae in a second, and coordinate category.
ACKNOWLEDGEMENTS
It is with considerable pleasure that I express my gratitude to Dr. Karel Liem of the Field Museum, Chicago, who during his recent visit to London has spent much time discussing various myological problems with me. His great interest in this work has added considerably to the pleasure of its execution, and I have benefited considerably from his experience. To Sharon Chambers go my thanks for the painstaking care she has taken in preparing the text-figures; and to my assistant Gordon Howes, my gratitude for his help in many aspects of the work. Finally, I must thank Miss Elizabeth Todd (Queen Elizabeth College, London University) for making available to me her unpublished observations on the musculature of several clupeoid fishes.
REFERENCES Auuis, E. P. 1897. The cranial muscles and cranial and first spinal nerves in Ama calva. J.
Morph., 12 : 487-808. BisHar, R. M. 1967. Cranial muscles of Mormyrus caschive (L.): Anat. Anz., 121 : 12-25.
OSTEOGLOSSOMORPH HYOID MUSCULATURE 55
Dietz, P. A. 10912. Vergelijkende anatomie van de kaak- en kieuweboogspieren dey Teleoste. Leiden.
EpGewortH, F. H. 1928. II. The development of some of the cranial muscles of ganoid fishes. Phil. Trans. R. Soc. Ser. B, 217 : 39-89.
EpGewortH, F. H. 1935. The cranial muscles of vertebvates. Cambridge University Press.
GREENWOOD, P. H. 1963. The swimbladder in African Notopteridae (Pisces) and its bearing on the taxonomy of the family. Bull. Bry. Mus. nat. Hist. (Zool.), 11 : 377-412.
GREENWoop, P. H. 1970. On the genus Lycoptera and its relationship with the family Hiodontidae (Pisces, Osteoglossomorpha). Bull. Br. Mus. nat. Hist. (Zool.), 19 : 257-285.
GREENWOOD, P. H., Rosen, D. R., WeitzMAN, S. H., & Myers, G.S. 1966. Phyletic studies of teleostean fishes, with a provisional classification of living forms. Bull. Am. Mus. nat. Hist., 131 : 339-456.
Hormouist, O. 1911. Studien in der von N.N. Trigeminus und Facialis innervierten Musku- latur der Knochenfische. II. Zur vergleichenden Morphologie der M.M. Intermandibularis, Protractor Hyoidei und Hyohyoideus. Acta Univ. lund., n.s., 7 : 47-79.
KircuHorr, H. 1958. Functionell-anatomische Untersuchung des Visceralapparates von Clupea harengus L. Zool. Jb. (Anat.), 76 : 461-540.
Liem, K. F. 1967. A morphological study of Luciocephalus pulcher, with notes on gular elements in other recent teleosts. J. Morph., 121 : 103-133.
McALListER, D. E. 1968. Evolution of branchiostegals and classification of teleostome fishes. Bull. natn. Mus. Can., Biol. ser., 77 no. 221 : 1-239.
Munsu1, J. D. 1960. The cranial muscles of some freshwater fishes. Indian J. Zootomy,
1, No. 2 : 1-76.
Myers, G. S. 1960. The mormyrid genera Hippopotamyrus and Cyphomyrus. Stanford ichthyol. Bull., 7, no. 4 : 123-125.
NeEtson, G. J. 1967. Branchial muscles in some generalised teleostean fishes. Acta zool., Stockh., 48 : 277-288.
Netson, G. J. 1968. Gill arches of teleostean fishes of the division Osteoglossomorpha J. Linn. Soc. (Zool.), 47 : 261-277.
NEtson, G. J. 1969. Infraorbital bones and their bearing on the phylogeny and geography of osteoglossomorph fishes. Am. Mus. Novit., No. 2394 : 1-37.
Orts, S. 1967. Contribution a l’anatomie comparee et a la systematique des mormyroides. Mem. Acad. vy. Sct. outremey (Cl. Sci. nat. med. 8°), N.S., 17 : 1-87.
OssE, J. W. M. 1969. Functional morphology of the head of the perch (Perca fluviatilis L.) : an electromyographic study. Neth. J. Zool., 19 (3) : 289-3092.
Takanasi, N. 1925. On the homology of the cranial muscles of the cypriniform fishes. J. Morph., 40 : 1-109.
TAVERNE, L. 1968. Ostéologie du genre Gnathonemus Gill sensu stricto [Gnathonemus petersii (Gthr.) et espéces voisines] (Pisces Mormyriformes). Annis. Mus. y. Afr. cent. Ser. 8°., no. 170 : 1-91.
TAVERNE, L. 1969. Etude ostéologique des genres Boulengeromyrus Taverne et Géry, Genyomyrus Boulenger, Petrocephalus Marcusen (Pisces Mormyriformes). Annls. Mus. yr. Afr. cent. Ser. 8°., no. 174 : 1-85.
TaVERNE,L. 1970. Note sur l’ostéologie du genre Gymnarchus Cuvier (Pisces Mormyriformes). Bull. acad. y. Belg. Cl. Sci., (5), 55 : 63-78.
VeTTER, B. 1878. Untersuchungen zur vergleichenden Anatomie der Kiemen-und Kiefer- musculatur der Fische. Jena. Z. Naturw., 12 : 431-550.
P. H. GREENWoop D.Sc.
Department of Zoology
BritisH Museum (NatTuRAL History) CROMWELL Roap
Lonpon SW7 5BD
A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY)
Kay, E. Attson. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. — 1965. (Out of Print.) £3.75.
WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier and Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4.
Taytor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure of Mineralogy at the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates 77 Text-figures. 1969. £4.50.
Haynes, J. R. Cardigan Bay recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. (In press.)
_
ahhe 5
BC Printed in England by Staples Printers Limited at their Kettering, Northants, establishment Te
f CLUPEOID -FISHES DESCRIBED - BY FRANCIS DAY
mPa es ge Sia Me F 2 x eed « H POYy Lm bets ii7Z ‘ Ser yer 5 ES ‘ ise
P. K. TALWAR AND P. J. P. WHITEHEAD
Bo yd eae >
BULLETIN OF 1 MUSEUM (NATURAL HISTORY) ee eee Vol. 22 No. 2
Vise en - LONDON : 1971
ergy
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY
BY PURNESH KUMAR TALWAR AND PETER JAMES PALMER WHITEHEAD
Pp. 57-85, 2 Plates, 3 Text-figures
BULLETIN OF : THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 22 No. 2
LONDON : 1971
THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series.
Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year.
In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department.
This paper is Vol. 22, No. 2 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals.
World List abbreviation Bull. Br. Mus. nat. Hist. (Zool.).
@©Trustees of the British Museum (Natural History), 1971
TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY)
Issued 31 December 197% Price £1.20
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY
By P. K. TALWAR & P. J. P. WHITEHEAD
ABSTRACT
Francis Day described six new clupeoid fishes: Spratelloides malabaricus, Clupea sindensis, Clupea variegata, Chatoessus modestus, Pellona sladeni and Engraulis auratus. For the first of these the new genus Dayella is proposed. Except for the last, all are considered valid. Lectotypes have been chosen from among Day’s figured specimens now in the Zoological Survey in Calcutta, except in the case of E. auratus for which a British Museum specimen has been chosen. The history of Day’s fish collections is briefly outlined; the Calcutta specimens are considered to be the most important, followed by those in Sydney, Vienna, Leiden, Berlin, Florence and Chicago. Apart from some small collections prior to 1870, the British Museum received only the remainders (in 1889).
INTRODUCTION
Francis Day (1829-1889) listed 55 clupeoid species in his Fishes of India (1875-8), of which 47 were illustrated, and 46 species are here recognized as valid (Table 1), He described six new species of clupeoid fishes, of which all but one are valid. Day’s descriptions and figures are generally good but many diagnostic features essential to modern clupeoid systematics are omitted, Unfortunately, Day did not specify which specimens were used in his original descriptions and a major problem has been to decide in which institution his types are deposited. The full history of Day’s collections is complex and will be described elsewhere (Whitehead & Talwar, in preparation) but a brief resumé can be given here.
Day’s first ichthyological work, the Fishes of Malabar (1865) resulted from his stay in Cochin (1859-64) and from this time stemmed small collections sent to Albert Giinther at the British Museum. Day subsequently investigated the fisheries in almost every large river and along most of the coast of India and Burma, making large collections and finally teturning to England in 1874 to work on his specimens and write his monumental Fishes of India (1875-1888). Unfortunately, a series of bitter quarrels developed between Day and Giinther, with the result that Day donated or sold much of his collection to other museums, the British Museum once again receiving material only in the year of Day’s death. The following is a sum- mary of the distribution of Day’s specimens:
1864-1870 British Museum (r15 lots, c. 400 specimens. Day types specified in letters but not in Register)
1865 East India Museum, London (7 species, including Engraulis auratus)
60 PK. FALWAR & PP. i. Po WHITEHEAD
1875-1879 Rijksmuseum van Natuurlijke Historie, Leiden (11 lots, c. 500 specimens. Day types claimed in Register)
1876-7 Indian Museum, Calcutta (figured specimens, now in Zoological Survey of India)
1874-1882 Zoologisches Museum, Berlin (many lots, 296 specimens. Day types claimed in Register)
1881-1884 Florence (three lots, 333 specimens, 3 types claimed in Register)
1883 Australian Museum, Sydney (Day collection from International Fisheries Exhibition, London. Day, Bleeker and Blyth types claimed by Whitley, 1958)
1886 Naturhistorisches Museum, Vienna (1000 specimens, 815 species— Day and Bleeker types claimed in Annual Report)
1889 British Museum (c. 5,000 specimens. Day types subsequently recognized) ; also Leningrad (see p. 85)
1899 Field Museum of Natural History, Chicago (452 Day specimens
from British Museum sent in exchange by Boulenger)
The Calcutta specimens include those used by Day in illustrating the Fishes of India (specified as such in Registration Book) and these are being listed by Talwar & Chakrapany (in press). We have concluded that, unless a valid lectotype has been chosen already, the figured Calcutta specimens are the most suitable. In his final letter of reconciliation to Giinther, Day specified that his type collection of Indian fishes went to Calcutta, his No. 2 to Sydney, No. 3 to Vienna, while Florence, Berlin and Leiden had large numbers of specimens. Thus the British Museum received his remainders, except for types in the pre-1870 lots. This order should be followed in making a lectotype selection.
In the descriptions given here, measurements follow those of previous clupeoid studies (e.g. Whitehead, Boeseman & Wheeler, 1966). Synonomies include references based on Day material or those relevant to the discussion of Day’s species. All further synonyms are given by Whitehead (in press). The following abbreviations are used:
AMS Australian Museum, Sydney
BMNH British Museum (Natural History), London FMNH Field Museum of Natural History, Chicago MNHN Muséum National d’Histoire Naturelle, Paris NHMV _ Naturhistorisches Museum, Vienna
RMNH Rijksmuseum van Natuurlijke Historie, Leiden ZMB Zoologisches Museum, Berlin
ZSI Zoological Survey of India, Calcutta
ACKNOWLEDGEMENTS
For information on the Day collections sent to Leiden, Berlin, Vienna, Florence, Sydney and Chicago, and for the loan of clupeoid specimens from Leiden, Paris and Berlin, we are indebted to Dr M. Boeseman of the Rijksmuseum van Natuurlijke Historie, Dr C. Karrer of the Zoologisches Museum, Dr P. Kahsbauer of the Natur- historisches Museum, Prof. Leo Pardi of the Museo Zoologico dell’Universita,
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 61
Dr J. R. Paxton of the Australian Museum, Dr L. P. Woods of the Field Museum of Natural History in Chicago, and Dr M.-L. Bauchot of the Muséum National d'Histoire Naturelle in Paris. The senior author also wishes to thank the Director, Dr A. P. Kapur, and the Superintending Zoologist, Dr A. G. K. Menon, of the Zoological Survey of India, for providing necessary facilities.
Family CLUPEIDAE (Subfamily CLUPEINAE)
I. Clupea sindensis Day, 1878 = Sardinella sindensis (Day, 1878)
(Plate ra)
Clupea sindensis Day, 1878, The Fishes of India: 638, pl. 163 (2) (Seychelles, Sind, Bombay; figure (life-size) on Karachi specimen, shown 95:5 mm S.L.); Idem, 1889, Fauna British India, Fishes, 1 : 374 (repeat).
Sardinella sindensis: Chan, 1965, Jap. J. Ichthyol., 13 (1-3) : 11, fig. 21 (key, 44 specimens ex Philippines); Whitehead (in press), Symp. Indian Ocean Adj. Seas. Mar. biol. Ass. India (key, synopsis, fig.).
MATERIAL. a. ZSI.2630, a fish 95-5 mm S.L., ex Karachi (stated to be basis for Day’s figure— LECTOTYPE)
b. ZSI.2614, a fish 90-0 mm S.L., ex Karachi (PARALECTOTYPE)
c. AMS.B7642, a fish 118-5 mm S.L. (140-5 mm tot. 1.), ex Bombay (claimed as type by Whitley, 1958—? PARALECTOTYPE)
NHMYV (no specimens)
RMNH (no specimens)
ZMB (no specimens)
BMNH.1889.2.1.1919-24, five fishes 77-5-99°5 mm S.L., ex Sind (registered as Clupea venenosa; one specimen, a skeleton)
ga rho a
Whitley (1958) gave wholesale endorsement to the type designations made in the published list of Day material bought by the Australian Museum in 1883 (Anon., 1885, 1886). This material, which had been shown at the International Fisheries Exhibition in London in 1883, had already been catalogued (Day, 1883) and, al- though some specimens were marked as the types of Bleeker and Blyth species, none was indicated as a Day type. Day appears to have described the present species on more than one specimen and while the Sydney specimen may be part of the syntypical series, we feel it preferable to designate as lectotype the one speci- men that definitely contributed to the original description, vzz. the Calcutta figured specimen, particularly since Day himself drew the figure.
Day (1878, 1889) tentatively included Meletta venenosa Valenciennes in his synonymy, hence the inclusion of the Seychelles in his distribution of the species. The Valenciennes species is Herklotsichthys punctatus (Rippell) (Whitehead, 1967 :
62 15 Jee ANINIBMIINGRS Go 12) odes AVIS CID ODL NID)
35). The British Museum specimens were not relabelled ‘sindensis’ until examined by Regan (1917a); it is unlikely, therefore, that they were used by Day in his description.
Description. Based on the LECTOTYPE, a fish 95:5 mm S.L., ex Karachi, ZSI1.2630 (basis for pl. 163 (2) of Fishes of India) (in parenthesis are given measure- ments for the Calcutta PARALECTOTYPE, ZSI.2614).
Br. St. 6, D iii 13 (14), Piz4, Vi7, Ali 16, g.r. 36 + 65 (37 + 63), scutes 18 + 14, scales in lateral series 42 (43), transverse II, pre-dorsal 15 (?).
In percentages of standard length: body depth 25-7 (23:9), head length 26-7 (22:5); snout length 6-5 (6-9), eye diameter 6-3 (6-9), upper jaw length 9-9 (10-6), lower jaw length 12-0 (10-6); pectoral fin length 16-2 (15-0), pelvic fin length 9-9 (8-1), length of anal fin base 15-7 (14:5); pre-dorsal distance 46-I (43-3), pre-pelvic distance 48-2 (48-9), pre-anal distance 76-4 (77°8).
Body fairly compressed, its width about 2} times in its depth, the latter more or less equal to head length; belly keeled, scutes partly concealed by scales on either side. Snout equal to or a little greater than eye diameter. Upper jaw reaching to vertical from anterior third of eye; two supra-maxillae, the Ist (anterior) about 5 times as long as deep, the 2nd (posterior) with upper and lower expanded parts similar in shape and size, the whole almost circular; no hypomaxilla; expanded portion of maxilla with faint longitudinal ridges, lower edge of maxilla with fine denticulations posteriorly. Lower jaw profile rising steeply, its depth half its length. Pre-maxillae and vomer edentulous, but fine teeth on either side of dentary sym- physis, a median line of conical teeth on tongue and fine teeth on palatines and ecto- and endo-pterygoids.
Gillrakers fine and slender, close-set, the longest about 4 of eye diameter and equal to length of corresponding gill filaments. Pseudobranch present, exposed, with a dozen filaments, its length about equal to eye diameter. Cleithral lobe and bilobed dermal outgrowths from cleithrum well developed. Operculum about twice as high as wide, its lower margin almost horizontal; sub-operculum rectangular. Opercular series and cheek covered by adipose tissue overlying ramifications of sensory canal system. Fronto-parietal region with cuneiform area bearing 8 (9) longitudinal striae; supra-orbitals with about four longitudinal striae.
Dorsal fin origin much nearer to snout than to caudal base; lower part of fin invested in scaly sheath. Pectoral fin tips failing to reach pelvic base by more than one eye diameter, failing to reach vertical from dorsal origin by 14 eye diameters; no axillary scale but scales above first ray truncated to leave shallow depression for reception of fin. Pelvic fin base below middle of dorsal base, nearer to pectoral base than to anal origin; axillary scale present, almost length of fin. Anal fin slightly nearer to caudal base than to pelvic base; last two rays enlarged, about twice length of antepenultimate ray.
Scales: unexposed portion of scale with one major and four (anterior scales) to six (posterior) minor vertical striae, the former continuous, the latter interrupted at centre of scale; exposed portion of scales with eroded and slightly crenellated posterior border, faint horizontal ridging and small perforations. Pre-dorsal
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 63
medial ridge covered by overlapping scale rows on either side. Alar scales absent (probably lost; present in some specimens of BMNH.1889.2.1.1919-24).
Colour: in alcohol, upper } of body slate-coloured, remainder of flanks silvery- gold. Fins hyaline, but dark spot at base of anterior dorsal rays. Inside face of operculum somewhat dusky.
Note. Sardinella sindensis, together with S. gibbosa (Bleeker), can be separated from other species of Sardinella by its slightly higher post-pelvic scute count (15-16, rarely 14 or 17-18; cf. 12-14, rarely 11 or 15—see key in Whitehead, in press). This slight distinction held true in 44 and 159 specimens (respectively) examined by Chan (1965), and also in British Museum material, and it is unfortunate that both lectotype and paralectotype of S. sindensis have the lower count of 14. One out of five other Day specimens (BMNH.1889.2.1.1919-24) has 14 post-pelvic scutes. If scute number is diagnostic, then S. sindensis can be separated from S. gzbbosa by its slightly higher range for gillraker numbers (58-72 at 69-122 mm S.L.; cf. 43-63 at go-150 mm S.L.—figures from Whitehead, in press). Specimens with only 14 post-pelvic scutes can be distinguished from S. albella (Valenciennes) and S. fimbriata (Valenciennes) by their more slender body (24:5-27:8% of S.L. (Chan, 1965); cf. 32-35 and 28-34°%, respectively—Whitehead, in press). Sardinella brachysoma Bleeker and S. zunasi (Bleeker) are also slightly deeper species which are further distinguished by the numerous overlapping or continuous vertical striae on the posterior scales.
(Subfamily PELLONULINAE) DAYELLA gen. nov.
TYPE SPECIES: Spratelloides malabaricus Day.
Diacnosis: clupeid fishes with 5-6 branchiostegal rays, a short anal fin (less than 20 rays), small unkeeled pre-pelvic scutes but no post-pelvic scutes, eight pelvic rays, a single (posterior) supra-maxilla, gillrakers present on posterior face of 3rd epibranchial, and posterior frontal fontanelles occluded in adults. A single species known.
2. Spratelloides malabaricus Day, 1873 = Dayella malabarica (Day, 1873)
Spratelloides malabaricus Day, 1873, Proc. zool. Soc. Lond.: 240 (‘Sea, ascending rivers in Malabar, and attaining 3 inches in length’); Idem, 1878, Fishes of India: 648, pl. 161 (5) (‘Western Coasts of India, in rivers and estuaries’; up to 3 inches, figure (? life size) 55:3 mm S.L.); Idem, 1889, Fauna British India, Fishes, 1 : 400, fig. 124 (repeat).
MATERIAL. a. ZSI.2246, a fish 51-0 mm S.L., ex Malabar (stated to be basis for Day’s figure although 4:3 mm shorter—LECTOTY PE) b. RMNH.2726, a fish 58 mm S.L., ex Malabar—PARALECTOTYPE
64 P. K. TAEWARN& PP. J.P’ WHITEBEAD
c. BMNH.1889.2.1.2048, a fish 47-3 mm S.L., ex Malabar, stained with alizarin —PARALECTOTYPE
d. Zool. Inst. Leningrad, 8220, a fish 48°3 mm S.L., ex Canara —PARALEC- TOTYPE
The following specimens are Ehirava fluviatilis.
e. AMS.B8288, a fish 44:0 mm S.L. (52:0 mm tot. 1.), ev Malabar (claimed as
type by Whitley, 1958)
f. NMV (no specimens)
g. ZMB.10413, three fishes 28-9-35-4 mm S.L., ex Malabar.
h. BMNH.1889.2.1.2051, two fishes 40-0-49:I mm S.L., ex Malabar (removed
from jar containing BMNH. paralectotype)
i. BMNH.1889.2.1.2050, one fish 53-0 mm S.L., ex Malabar (also removed from
BMNH paralectotype jar)
j. BMNH.1889.2.1.2052-5, four fishes 46:5-56:2 mm S.L., ex Canara (one fish 50-0 mm stained with alizarin; three fishes donated to the Musée Royale de l'Afrique Centrale, Turvuren)
. FMNH.2379, a fish 49:0 mm S.L., ex Canara (numbered 240 and donated by G.A. Boulenger from BMNH collection)
an
The two species included in the Day material are superficially very similar and hitherto Deraniyagala’s fluviatilis has been considered a synonym of Day’s mala- baricus (e.g. in Whitehead, 1963). It was not until the single true specimen of Dayella malabarica in the British Museum was stained with alizarin and re-examined that the Day material was found to be mixed. Although specimens of E. fluviatilis predominate, Day’s original description seems to have been based on D. malabarica since Day states that the ‘dorsal commences slightly before the origin of the ventral’. In D. malabarica the dorsal origin is well before the pelvic base, the latter lying below the first branched dorsal ray. In E. fluviatilis the pelvic base is before, below or only just behind the first wnbranched dorsalray. The statement is repeated in Day’s second description and in his figure (Day, 1878 : pl. 161) the dorsal origin is clearly well before the pelvic base.
The specimen in Calcutta is slightly smaller than Day’s figure but is presumed to have been the model and is here chosen as lectotype of Spratelloides malabaricus. The holotype of Ehivava fluviatilis Deraniyagala is in London (BMNH.1929.7.1.1) ; the specimen in the Zoological Survey of India (ZSI.F11043/1), claimed as a paratype by Menon & Yazdani (1963), is from Moratua (Western Province of Ceylon), a locality not mentioned in the original description.
DESCRIPTION. Based on the LECTOTYPE, a fish 51-0 mm S.L., ex Malabar, ZSI1.2246 (basis for pl. 161 (5) ot Fishes of India). Figures for the Leiden and British Museum specimens (49:5 and 47:3 mm S.L.—lots b and c above) are given in parenthesis.
Bye SUG (6) (Gy 5) ID) tht aese (Gere, sea), 12) at Fe (Ger), aes), Wah, (Ge, a) eX vob) aeY (02105), (C nr. (n.r., 10 + 9g), g.r. II + 27 (10 + 24, Io + 27), scutes ? o (4, 1), scales in lateral series 38 (n.r., 36), transverse g (n.r., n.r.), vertebrae 40 (BMNH alizarin specimen).
THE
CLUPEOID FISHES DESCRIBED BY FRANCIS DAY
Fic. 1. Pre-pelvic scutes in three Indo-Pacific pellonulines. a. Dayella malabarica, 473mm S.L., BMNH.1889.2.1.2048. b. Dayella malabarica, 49-5 mmS.L., RMNH.8585. c. Gilchristella aestuarius, 52:3 mm S.L., BMNH.1915.7.6.3. d. Ehivava fluviatilis,
491mm S.L.,
BMNH.1889.2.1.2051.
65
66 PO oK. TALWAR & PB. J: PS WHITEREAD
In percentages of standard length: body depth 22-5 (20-0, 20-0), head length 27°7 (24-0, 26-0); snout length 8-3 (7-5, 7-4), eye diameter 8-3 (8-0, 7-8), post-orbital distance 9-9 (8-5, 8-7), length of upper jaw 9-9 (8-7, 9-7), length of lower jaw 14:7 (12-3, 13-1); pectoral fin length 16-7 (17-1, 15-8), pelvic fin length 13-7 (13-3, 14:5), length of anal base 18-6 (15-3, 14°5); pre-dorsal distance 49-0 (47:9, 48:5), pre-pelvic distance 52-9 (50-4, 49°5), pre-anal distance 78-4 (77:3, 75:0).
Body fairly compressed, its width almost 3 times in its depth, the latter a little less than head length; belly rounded, fully scaled but the scales underlain by 4 (Leiden) or 1 (BMNH) plate-like scutes bearing rudimentary lateral arms (fig. 1a, b), the scutes not reaching back to the main pelvic scute (which has normal lateral arms). Snout equal to or a little shorter than eye diameter. Jaws unequal, the lower projecting slightly. Upper jaw reaching to vertical from anterior eye border or anterior pupil margin, ventral expanded portion of maxilla beginning abruptly and not tapered smoothly into slender anterior limb of bone, the entire edge of the expanded portion finely denticulated (fig. 2); a single (posterior) supra-maxilla, the lower part of the expanded portion deeper and longer than the upper part (Harengula shape), its depth about + eye diameter. Lower jaw rising fairly steeply in the first third of its length; 6-7 small conical teeth on either side of symphysis. A single row of small conical teeth on pre-maxillae, separated by a median diastema. Fine teeth on tongue, scattered on antero-median process of palatine and along outer edge of that bone.
Fic, 2. Dayella malabarica, upper and lower jaws, alizarin stained specimen, 47-3 mm S.L., BMNH.1889.2.1.2048.
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 67
Gillrakers fine, slender, 24 times in eye diameter and a little longer than corres- ponding gill filaments; about 7 short, stumpy gillrakers on posterior face of 3rd eipibranchial. Pseudobranch present, exposed, about } eye diameter, with about a dozen filaments. Cleithral lobe present at lower angle of gill opening, fairly well developed. Operculum about 14 times as deep as broad, its posterior margin with a deep indentation, its lower margin rising at an angle of about 15° to the horizontal; suboperculum rectangular except for rounding of postero-ventral angle. Cutaneous sensory canals branching over entire opercular series and cheek. Fronto- parietal rea smooth, posterior frontal fontanelles retained and minute, I-o mm in length (oa-5 mm in BMNH specimen—fig. 3c; virtually occluded in Leiden specimen).
Dorsal fin origin nearer to snout tip than to caudal base by ? eye diameter and a little in advance of vertical from pelvic base. Pectoral fin tips failing to reach pelvic base by 1 (14) eye diameters; no axillary scale. Pelvic fin base below first branched dorsal ray and about equidistant between pectoral base and anal origin; axillary scale present, about } of fin length (Leiden specimens). Anal fin origin a little nearer to caudal base than to pelvic base; last two rays normal, not separated from rest of fin.
Scales: deeper than broad, with distinct anterior ‘shoulders’; unexposed portion with a single continuous striation, preceded by 3-4 short radiating striae (absent on anterior scales, joined to form a loop on some posterior scales) ; exposed portion of scale without striae, its posterior margin slightly eroded and produced medially.
Colour: in alcohol, uniform light brown with a faint silvery midlateral stripe not quite as broad as eye; fins hyaline.
Note. The new genus Dayella is a member of the pellonuline complex that comprises the monotypic Indo-Pacific genera Ehivava, Gilchristella, Sauvagella and Spratellomorpha. These five genera are distinguished from all other Indo- Pacific pellonulines by their lack of post-pelvic scutes; in addition, pre-pelvic scutes are either absent or extremely poorly developed. The separation of the five species at generic level is arguable but the features that distinguish them are non-meristic and appear to be of some significance in other groups of clupeids. Dayella can be identified from the following key.
INDO-PACIFIC PELLONULINAE THAT LACK POST-PELVIC SCUTES
I. Anal fin entire, last two rays not separate A. Gillrakers present on posterior face of 3rd epibranchial I. Posterior frontal fontanelles minute (4 eye diameter) or completely occluded in adults (fig. 3c); pre-pelvic scutes 1-4, rudimentary, lateral arms barely developed (fig. ra, b); pelvic base well behind dorsal origin; gillrakers 24-27 Dayella malabarica (Day) 2. Posterior frontal fontanelles larger (4-4 eye diameter), retained in adults (fig. 3a, b); pelvic base before dorsal origin a. No pre-pelvic scutes; gillrakers Ig (at 40 mm S.L.) Sauvagella madagascariensis (Sauvage)
68 PK. TALWAR & P. J. P) WHITEHEAD
b. Up to 9 pre-pelvic scutes, with thin lateral arms i. Anterior arm of supra-occipital broadening anteriorly (fig. 3b); gillrakers 39-45 Gilchristella aestuarius (Gilchrist) ii. Anterior arm of supra-occipital very slender anteriorly (fig. 3a); gillrakers 40-60 Gilchristella sp. (see below) B. Gillrakers absent on posterior face of 3rd epibranchial; posterior frontal fontanelles large, anterior arm of supra-occipital broadening anteriorly (fig. 3d); pre-pelvic scutes present, poorly developed but with distinct lateral arms (fig. 1d); pees base below or before dorsal origin; gillrakers 26-30 . : : F Ehirava fluviatilis Deraniyagala
II. Anal fin split, the Hast two rays separate from rest of fin; gillrakers present
on posterior face of 3rd epibranchial; posterior frontal fontanelles large,
probably retained in adults, similar to those of Ehirava; gillrakers 26-31 Spratellomorpha bianalis (Bertin)
The four previously described Indo-Pacific genera were formerly placed in the round herrings or Dussumieriidae (Whitehead, 1963). The subsequent discovery of partially scuted (Laeviscutella, Sierrathrissa) or non-scuted (Congothrissa) forms amongst the otherwise fully scuted West African Pellonulinae suggested that both Indo-Pacific and African genera were members of a clupeid group that showed progressive stages in scute loss. This appeared to be correlated with trends towards reduction in supra-maxillae and numbers of branchiostegal rays, together with a retention of the posterior frontal fontanelles by adults (Poll, Whitehead & Hopson, 1965). The non-scuted genera Spratelloides (Indo-Pacific) and Jenkinsia (Western Atlantic) may eventually join this group, although their very characteristic W-shaped pelvic scute seems to link them with the ‘true’ round herrings Dussuwmieria and Etrumeus—whose high and presumably primitive branchiostegal count implies yet another route to scute loss (or perhaps the primitive absence of scutes, at least in this branch of the clupeids). For the present, the five poorly or non-scuted genera shown in the key above are placed in the tribe Ehiravini of the subfamily Pellonulinae.
In Dayella, the scutes are more rudimentary than in any other clupeid genus. They are thin, difficult to find in unstained material and those with small lateral arms could easily be mistaken for scales. Their resemblance to scales is increased by the relatively large size of the expanded portion of the scute when compared with those of other genera (fig. ta-d). The variation in shape of these scutes implies that these are structures on the way to being lost and not an early stage in the evolution of scutes.
Dayella appears to be most closely allied to Gilchristella, Sauvagella and Ehirava, differing from them chiefly in its occluded posterior frontal fontanelles and its less advanced pelvic base. The absence of gillrakers on the posterior face of the 3rd eipibranchial seems to hold some significance elsewhere in the Clupeidae, but this
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 69
may not mean that Ehivava is necessarily remote from the other genera of the group. The status of Sauvagella madagascariensis is uncertain. Re-examination of a syntype of this species (40 mm S.L., MNHN.3794) has confirmed that even the rudimentary scutes of Dayella are not present. This fish has only 109 gillrakers and the posterior frontal fontanelles are large, suggesting that it is a juvenile; gillrakers may increase with size. In a redescription of the species (Whitehead, 1963), eleven South African specimens were included, from the Buffalo river, Cape Province (BMNH.1878.1.22.25 and 33-43) and the stated gillraker count of 40-56 referred to these specimens only. Careful removal of the belly scales now shows that these fishes have up to 7 thin and barely apparent scutes with fairly long lateral arms. Thus, they are clearly distinct from Sauvagella (as far as can be judged from the very small types) and for the moment they appear to be an undescribed species ot Gilchristella. From G. aestuarius they differ, however, in having a very slender
P
Fic. 3. Posterior frontal fontanelles in four Indo-Pacific pellonulines, dorsal view showing frontals (fr.), parietals (p.) and anterior arm of supra-occipital (s. occ.), the fontanelles black. a. Gilchristella sp., 52-6 mm S.L., anomalous Buffalo River specimen, BMNH.1878.1.22.25. b. Gilchristella aestuarius, 54-5 mm S.L., BMNH.19109.9.12.3. c. Dayella malabarica, 47-3 mm S.L., BMNH.1889.2.1.2048. d. Ehivava fluviatilis, 50:0 mm S.L., BMNH.1889.2.1.2052.
70 TPE Is ABYMEM VINE (4 125 fg 1, \WiISWlsh olen d/N1D)
anterior arm of the supra-occipital (fig. 3a) and a higher gillraker count (40-60 cf. 39-45). In general, the members of the Ehirava complex are rather poorly known and would repay further study when more specimens are available.
(Subfamily ALOSINAE) 3. Clupea variegata Day, 1869 = Gudusia variegata (Day, 1869) (Plate 1c)
Clupea variegata Day, 1869, Proc. zool. Soc. Lond.: 623 (Irrawaddy and its branches; many specimens, to 7 inches); Idem, 1878, Fishes of India: 639, pl. 161 (4) (repeat; figure of fish 152 mm S.L., presumably life-size); Idem, 1889, Fauna British India, Fishes, 1 : 375 (repeat).
Gudusia variegata: Regan, 1917, Ann. Mag. nat. Hist., (8) 19 : 308 (on single Day specimen in British Museum); Motwani, Jayaram & Sehgal, 1962, Tvop. Ecol., 3 (1-2) : 17-43 (Brahmaputra at Jogighopa, Goalpara District); Whitehead, 1965, Bull. By. Mus. nat. Hist. (Zool.), 12 (4) : 150, fig. 11 (Day specimen redescribed; Day’s figure reproduced) ; Idem, (in press), Symp. Indian Ocean Adj. Seas. Mar. biol. Ass. India (key, note on synonymy).
? Clupea suhia Chaudhuri, 1912, Rec. Indian Mus., 7 : 436, pl. 38 (1) (river Gandak in Saran, Bihar).
? Gudusia godanahiai Srivastava, 1968, Fishes Easteyn Uttar Pradesh: 6, fig. 4a, b (Gorakhpur, Uttar Pradesh).
MATERIAL. a. ZSI.2245, a fish 150:0 mm S.L., ex Irrawaddy river (stated to be basis for Day’s figure—LECTOTY PE) b. ZSI (Duplicate Cat.) 43, a fish 158 mm S.L., ex Bassein river (labelled Clupea burmanica) c. ZSI. (Duplicate Cat.) 168, a fish 78 mm S.L., ev Mandalay, coll. Major E. B. Sladen (labelled Clupea burmanica) d. AMS.B7676, a fish 158-5 mm S.L. (191 mm tot. 1.), ex Bassein (claimed as type by Whitley, 1958. NHMV (no specimens) RMNH.2586, a fish 106-4 mm S.L., ex Bassein. ZMB (no specimens) BMNH.1870.6.14.38, a fish 155 mm S.L., ex Bassein.
Bog rh ©
For reasons given under the previous two species, the figured specimen in Calcutta is chosen as lectotype. The two specimens labelled ‘buymanica’ ate G. variegata but are probably not syntypes since their meristic counts exceed the values given in the original description (but are consistent with the ranges given in the Fishes of India). Day did not publish the name burmanica but may have initially intended to use it for this species. The British Museum specimen, although presented in 1870 and thus the first of this species to be given away by Day, has too low an anal and pectoral count (111 22 and i113; cf. 111 26 and i15) to have figured in the original
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 71
description, which in any case seems to have been based on a single specimen (although many were taken). For this reason, the Sydney and Leiden specimens cannot be regarded as paralectotypes.
DEscRIPTION. Based on the LECTOTYPE, a fish 150-0 mm S.L., ex Irrawaddy river, ZSI.2245 (basis for pl. 161 (4) of Fishes of India).
Br. St. 6, Diii 12, Pi 15, Vi7, A iii 26, g.r. 210 (approx.), scutes Ig + II, scales in lateral series go, transverse 35.
In percentages of standard length: body depth 44-3, head length 29:3; snout length 5-5, eye diameter 6-7, length of upper jaw 11-7, length of lower jaw 13-0, operculum height 14-3, its breadth 8-3; pectoral fin length 20-0, pelvic fin length 10-0, length of anal fin base 23-0; pre-dorsal distance 52-0, pre-pelvic distance 54-7, pre-anal distance 74:0.
Body strongly compressed, its width 33 times in its depth, the latter 1} times head length; belly keeled, tips of scutes projecting slightly from sheath of scales on either side. Snout shorter than eye diameter; pre-orbital distance (including eye) } of post-orbital distance. Lower jaw included when mouth closed: pre- maxillae rising steeply to form very distinct notch in upper jaw. Maxilla reaching to just beyond vertical from eye centre, expanded portion without longitudinal ridges or striae and smooth along lower edge; two supra-maxillae, the rst (anterior) about 6 times as long as deep and almost equal in length to eye diameter, the 2nd (posterior) with slender anterior shaft and lower part of expanded portion larger than upper. Lower jaw rising rather gently in the first third of its length, its depth about 3 times in its length; no teeth present. No teeth in upper jaw nor within mouth.
Gillrakers fine, straight or slightly curved, slightly longer than corresponding filaments; filaments of anterior hemibranch of 1st arch 3-? length of those of posterior hemibranch; many fine gillrakers present on posterior face of 3rd epibranchial. Medio-pharyngobranchial present, about 2 eye diameter, bearing many short gillrakers. Pseudobranch present, attenuated, about 1# times eye diameter, with distinct ventral ridge and a groove below it. Cleithral lobe barely developed, hardly breaking outline of gill opening. Operculum not quite twice as deep as broad, its lower edge rising steeply; sub-operculum crescentic; lower third of an- terior operculum margin not overlapped by pre-operculum, leaving a small triangular area covered only by skin. Cutaneous sensory canals branching through the adipose tissue covering the suborbitals, operculum, sub-operculum and scales behind head. Adipose eye-lid with vertical slit exposing 2 of pupil. Dorsal surface of head covered by fairly thick skin but a pair of cuneiform fronto-parietal areas with about six longitudinal striae left exposed.
Dorsal fin origin slightly nearer to snout tip than to base of caudal; a very low scaly sheath along base. Pectoral fin tips failing to reach pelvic base by about 2 eye diameter; axillary scale present, half length of fin. Pelvic fin base below vertical from dorsal origin and a little nearer to pectoral base than to anal origin; axillary scale present, half length of fin. Anal origin equidistant between pelvic and caudal bases; anal base longer than pectoral fins and greater than the distance
72 P. K. TALWAR & P. J. P. WHITEHEAD
snout tip to posterior border of pre-operculum. Caudal fin (broken) slightly longer than head length, lower lobe longer than upper.
Scales: almost circular, but becoming more elongate on posterior part of body; a single vertical striation continuous across scale, preceded by o (anterior scales) to 3 (posterior scales) short and irregular striae interrupted at centre of scale; exposed border irregular, becoming pectinate in posterior scales. Minute scales covering caudal except for hind border.
Colour: in alcohol, back brown, flanks golden; a series of brown spots along upper flank, some expanded vertically, those behind dorsal extending right across back; a dark humeral spot.
Note. The genus Gudusia (often misspelt Gadusia in the literature) at present includes two species, G. chapra (Ham. Buch.) and G. variegata. The latter has a deeper body (depth greater than 40% of S.L.), a shorter head (head length less than 28% of S.L.), and more anal finrays (111 22-26; cf. iii 19-22). In addition, G. variegata has a very prominent series of black spots along the flank, whereas G. chapra is usually described as having a dark shoulder spot, sometimes absent, and faint or no spots along the flank (Whitehead, 1965; Srivastava, 1968). Rather few specimens of G. variegata have been described, however, and some of these were misidentifications. Thus, specimens identified as G. variegata from Akyab by Lloyd (1907) include at least one fish (ZSI.1491/1) that is Hilsa kelee, while juveniles reported as G. variegata from the Mandalay fish market by Jenkins (1910)"include specimens (ZSI.1770/1) of a species of Hilsa (Tenualosa).
Gudusia variegata is usually considered a Burmese species (Regan, 1917b; Fowler, Ig4I : 635), but both Chandhuri’s Clwpea suhia and Srivastava’s Gudusia godanahiar (Ganges drainage) had very prominent black spots along the flanks which were, indeed, the main reason for distinguishing these nominal species from the sympatric and unspotted G. chapra (not a sexual feature according to Srivastava, 1968). Gudusia variegata has also been reported from the Brahmaputra, by Motwani et alii (1962), presumably because of the strong pattern of spots since G. chapra was also recorded from the same area (no descriptions given, however). If these Indian records truly relate to G. variegata, then some modification must be made to the key since Srivastava’s G. godanahiai were rather slender (depth 33-3-38:7% of S.L.) and thus within the range of G. chapra (31-0-40:0% in 30 specimens—Whitehead, 1965: fig. 13); the head length (27-4-31-7% of S.L.) of Srivastava’s specimens, however, agreed with current definitions of G. variegata, but the low anal count of iii 20 was that of G. chapra. Srivastava distinguished his new species by the presence of 14 pectoral rays (13 in his G. chapra), but Regan (1917b) recorded 13-14 pectoral rays in the British Museum material of G. chapra and there is probably overlap between the two species. For the present, the status of G. variegata and its possible synonyms must remain uncertain until more material has been examined.
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 73 (Subfamily DOROSOMATINAE)
4. Chatoessus modestus Day, 1869 = Gonialosa modesta (Day, 1869)
Chatoessus modestus Day, 1869, Proc. zool. Soc. Lond.: 622 (Bassein river as high as Een-gay-gyee Lake; many specimens, up to 5} inches); Idem, 1878, Fishes of India: 633, pl. 160 (1) (also Selwein at Moulmein; figure of fish 100-r mm S.L., ? life size); Idem, 1889, Fauna British India, Fishes, 1 : 386 (repeat).
Gonialosa modesta: Regan, 1917, Ann. Mag. nat. Hist., (8) 19 : 315 (on Day material in British Museum); Menon & Yazdani, 1963, Rec. 2001. Surv. India, 61 : 98.
MATERIAL.
a. ZSI.2695, a fish 98-0 mm S.L., ex Bassein river (stated to be basis for Day’s figure—LECTOTY PE)
b. ZSI.F8022/1 and 8023/1, two fishes 58-o-r01-0 mm S.L., ex Bassein river (= G. manminna)
c. AMS.B7637, a fish 105-0 mm S.L. (127:5 mm tot. 1.), ex Burma (claimed as
type by Whitley)
NHMYV (no specimens)
RMNH.2585, a fish 116-0 mm S.L., ex Moulmein (claimed as type in register)
ZMB (no specimens)
g. BMNH.1880.2.1.1879, a fish 82-6 mm S.L., ex Burma
Menon & Yazdani (1963) erroneously listed the first of the three Zoological Survey fishes as holotype, and the other two as paratypes, but Day did not indicate a holo- type nor did he give an exact length measurement. The first Calcutta specimen, used for Day’s figure, is here designated lectotype; the other two are G. manminna and thus do not agree with the original description, which in any case shows no ranges for meristic and morphometric values, suggesting that only a single fish was measured (although many were caught). For this reason, the Sydney, Leiden and London specimens are not regarded as paralectotypes.
DEscRIPTION. Based on the LECTOTYPE, a fish 98-0 mm S.L., ex Bassein river, | ZSI.2695 (basis for pl. 160 (1) of Fishes of India). Br. St. 6, D iii 13, Pir5, Vi7, A iii 25, g.r. 150 (approx.), scutes 17 + 12, scales in lateral series 47, transverse 17.
In percentages of standard length: body depth 48-5, head length 27-6; snout length 6-4, eye diameter 7:9, post-orbital distance 13-3, length of upper jaw 7:1, length of lower jaw 9-7; pectoral fin length 23-0, pelvic fin length 10-2, length of anal base 26-8; pre-dorsal distance 52-0, pre-pelvic distance 50-0, pre-anal distance 72°4.
Body compressed, its width 4,3, times in its depth, the latter almost twice head length; belly keeled, tips of scutes projecting below scaly sheath; profile of back concave beyond nape, rising abruptly to dorsal origin, belly profile evenly convex. Snout shorter than eye diameter. Mouth sub-terminal, transverse, with snout projecting strongly; pre-maxillae meeting at an angle to form a distinct notch in
tho A
74 Pp. K. TALWAR & P. J. P. WHITEHEAD
upper jaw. Maxilla slender, slightly expanded and curved downwards distally, reaching to vertical from anterior border of eye; a single narrow supra-maxilla, its length 2-5 mm and depth 0-5 mm. Lower jaw with dentaries meeting at an obtuse angle, the edge of each dentary flared or reflected outwards in front of tips of maxillae. No teeth in jaws.
Gillrakers very fine, close-set, shorter than corresponding gill filaments and slightly more than } eye diameter; gill filaments of anterior hemibranch equal to those of posterior hemibranch. Numerous fine and close-set gillrakers on posterior face of 3rd epibranchial. Pseudobranch present, exposed, its length almost one eye diameter; about 20 filaments present. Cleithral lobe present, breaking outline of gill opening but not strongly developed. Operculum 1} times as deep as broad, its lower border rising steeply (about 40°); suboperculum long and narrow, its posterior border rounded. Cutaneous sensory canals branching over cheek, oper- cular series and nape. Adipose eye-lid with vertical slit exposing }4 of eye. Dorsal surface of head with a pair of cuneiform fronto-parietal areas bearing 6 longi- tudinal striae, the two areas linked posteriorly by a transverse bony ridge.
Dorsal fin origin slightly nearer to snout than to base of caudal fin; final finray not elongated. Pectoral fin tips reaching to beyond pelvic base; axillary scale present, } length of fin. Pelvic fin base in front of dorsal origin and nearer to pectoral base than to anal origin; axillary scale present, § length of fin. Anal origin nearer to pelvic than to caudal base. Lower lobe of caudal larger than upper.
Scales: unexposed portion with one major vertical striation and 3 (anterior scales) to 4-5 (posterior) minor vertical striae interrupted at scale centre. Exposed portion without striae, posterior margin of scale not eroded, perforated or fimbriated.
Colour: in alcohol, upper parts of body light brown, lower parts silvery; a dark humeral spot present. Fins hyaline.
Note. The two species of Gonialosa, G. modesta and G. manminna (Ham. Buch.), have been separated on body depth (40-50% of S.L. in modesta ; 30-39 % in manminna) and number of scales along the flank (45-47 and 55-65 respectively—Whitehead, 1962; in press). Gudusia manminna (from the Ganges and Brahmaputra and their tributaries) is fairly well represented in collections and the literature, but G. modesta (recorded only from Burma) is not. Larger collections may show modesta to be merely a subspecies of G. manminna.
(Subfamily PRISTIGASTERINAE)
5. Pellona sladeni Day, 1869 = Ilisha sladeni (Day, 1869)
Pellona sladeni Day, 1869, Proc. zool. Soc. Lond.: 623 (Irrawaddy at Mandalay; specimens up to 7 inches); Idem, 1878, Fishes of India: 645, pl. 164 (1) (repeat; ‘A single example obtained, 7 inches in length’; figure shows fish of 146 mm S.L.); Idem, 1880, Fauna British India, Fishes, 1 : 383 (repeat).
Ilisha sladeni: Norman, 1923, Ann. Mag. nat. Hist., (9), 11 : 6 (Day specimen and two others described).
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 75
MATERIAL.
a. ZSI.2672, a fish 210 mm S.L. (ca. Io inches tot. 1.), ex Irrawaddy (stated to be basis for Day’s figure—LECTOTY PE)
b. ZSI (Duplicate Cat.) 298, a fish 189 mm S.L. (ca. 8} inches tot. 1.), ex Mandalay (? PARALECTOTY PE)
AMS (no specimens)
NHMV (no specimens)
RMNH (no specimens)
ZMB (no specimens)
BMNH.1870.6.14.36, a fish 209 mm S.L. (almost ro inches tot. 1.), ex Mandalay.
mM ro Ae
As in previous cases, Day clearly collected more than one specimen, but he gave no ranges for meristic or morphometric values in his original description. In the Fishes of India, however, he stated ‘single example obtained’; he may perhaps have been referring to the fish that he himself had drawn for the Fishes of India, since it was now eight years since he had given his other specimen to the British Museum. All three extant specimens are larger than either the figure or the maximum length stated (7 inches). They also differ in having more pre-pelvic scutes (23 or 24; cf. 20) and more pectoral rays (14 or 15; cf. 11). Day altered his scute count to 23 in the Fishes of India (but not the pectoral count) and it must be presumed that the earlier counts were errors. The larger Calcutta fish, the figured specimen, is chosen as lectotype, on the assumption either that Day lost the original (smaller) specimen or that the maximum length of 7 inches was also an error.
DEscRIPTION. Based on the LECTOTYPE, a fish 210 mm S.L., ex Irrawaddy river, ZSI.2672 (basis for pl. 164 (x) of Fishes of India) (measurements of ZSI para- lectotype given in parenthesis).
Br. St. 6, D iii 10, Pi 13, Vi6, A iii 4I, g.1. Io + I + 21, scutes 23 + 10, scales in lateral series 48, transverse 10.
In percentages of standard length: body depth 21-9 (25-4), head length, 24:5 (27:5); snout length 5-5 (5-8), eye diameter 6-0 (6-6), length of upper jaw 11-0 (11-9), length of lower jaw 11-4 (13-0); pectoral fin length 19-8 (21-7), pelvic fin length 4°8 (6-6), length of anal fin base 30-0 (29-4) ; pre-dorsal distance 59°I (59:3), pre-pelvic distance 39-I (44-0), pre-anal distance 65:2 (67-2).
Body strongly compressed, its width 4 times in its depth, belly strongly keeled, the tips of the scutes projecting below scaly sheath, especially behind pelvic fin base; anterior four scutes on isthmus. Dorsal profile slightly concave before nape, ventral profile evenly convex, the two almost parallel between pectoral base and dorsal origin. Snout a little shorter than eye diameter. Lower jaw strongly projecting, about 4 eye diameter beyond snout when mouth closed. Maxilla teaching to vertical from anterior pupil border, fine denticulations along its lower edge; no hypo-maxilla; two supra-maxillae, the rst (anterior) 5 times longer than deep and about 1} eye diameter, the 2nd (posterior) with lower lobe of expanded portion much larger than upper. A single series of fine teeth on pre-maxillae, with median diastema, small conical teeth present on either side of dentary symphysis. No teeth on vomer but fine teeth on tongue, palatines and ecto- and endo-pterygoids.
76 Pr Ke DAL WAR Te Ps Po WHITE EE AD
Gillrakers fairly slender, the longest + eye diameter and 14 times length of cor- responding gill filaments; no gillrakers on posterior face of 3rd epibranchial. Pseudo- branch present, exposed, its length } eye diameter; ventral margin not ridged. No cleithral lobe. Operculum elongated posteriorly, its ventral margin equal to its height and rising at an angle of about 20°; sub-operculum eliptical, long and narrow, its height 3,4, times in its width; lower border of sub-operculum and hind border of inter-operculum almost parallel to upper profile of head, to leave a broad triangular area below (bounded posteriorly by base of pectoral fin). Dorsal surface of head with two prominent longitudinal striae, diverging slightly posteriorly, flanked by two small lateral striae over eyes.
Dorsal fin origin set far back on body, equidistant between caudal base and posterior margin of operculum. Pectoral fin tips reaching almost to tips of pelvics; axillary scale present, } length of fin. Pelvic base nearer to pectoral base than to anal origin by 2 eye diameters; axillary scale present, about 4 length of fin. Anal fin origin below vertical from anterior third of dorsal base; base of fin covered by low scaly sheath.
Scales: unexposed portion with a single complete W-shaped vertical striation, preceded by 3 (anterior scales) to 6 (posterior) shorter striae interrupted at centre of scale; exposed portion with about 16 very short radiating striae at edge of scale, not discernible in posterior scales.
Colour: in alcohol, upper 4 of body brown, rest of flanks silvery; fins hyaline, hind margin ot caudal dusky. Inner face of operculum slightly dusky.
Note. Ilisha sladeni closely resembles J. pristigastroides (Bleeker), with which it has been synonymized (Whitehead, 1970), but comparison of Day’s material with Bleeker’s type in the British Museum (1867.11.28.12—redescribed in Whitehead et alii, 1966) suggests that I. sladeni is distinct. In both species the anal origin is well before the vertical from the midpoint of the dorsal base, a feature used in keys to separate these species from all other Ilisha (Whitehead et alit, loc. cit.; Whitehead, 1970). The type of I. pristigastroides is a smaller fish (151 mm S.L.; cf. 189-210 mm in the Day material), but this does not account for its deeper body (30:8% of S.L.; cf. 21-9, 25-4 and 22-4% in the Day material) since a larger Bleeker specimen (302 mm S.L., BMNH.1867.11.28.9) is still deeper-bodied (30:9% of S.L.) than a similar large specimen of J. sladeni (25-9% ina fish of 308 mm S.L., ex Sittang river, Burma, BMNH.1891.11.30.402). The Bleeker type also has a relatively longer anal base (41:1% of S.L.; cf. 30-0, 29:4 and 30-8%), the anal origin being set further back on the body (equidistant between caudal base and eye centre; cf. nearer to caudal base than to pectoral base), and the dorsal origin is also corres- pondingly less far back on the body. In spite of the more compact body in I. pristigastroides, there are more pre-pelvic scutes (26) than in the elongate J. sladeni (23-24). All these features also serve to separate the two larger specimens mentioned above.
A striking feature of I. sladeni, to some extent shared by the type of J. pristi- gastrotdes, is the very elongate appearance of the head, shown well in Day’s drawing (see Pl. 2). This is partly due to the length of the head but more particularly to
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 77
the more squat operculum (its height just over 24 times in head length; cf. 2-2} times in other species). Also, the depth of the head, taken at right angles to the profile at the occiput, is much less (1 ? times in head length) than in other species (1j- 14 times).
A further difference between I. sladeni and I. pristigastroides is in the form of the swimbladder. In J. sladeni the swimbladder terminates at the posterior end of the body cavity (BMNH specimens of 209 and 308 mm S.L.), whereas in I. pristigastroides there is a postcoelomic, tapering prolongation down the right side of the body lateral to both the haemal spines and the anal pterygiophores and reach- ing as far as the level of the 15th branched anal ray. The condition in J. sladeni appears to be unique amongst Indo-Pacific members of J/isha but it is found in the South American J. furthii (and probably also in the related I. amazonica—no BMNH specimens). The asymmetrical postcoelomic prolongation in J. pristi- gastroides is similar to that found in the Indo-Pacific J. elongata and I. megaloptera; in I. africana (West Africa) and I. indica (Indian Ocean) the prolongation of the swimbladder is bifid.
Family ENGRAULIDAE
6. Engraulis auratus Day, 1865
= Thryssa dussumieri (Valenciennes, 1848)
(Plate 2)
Engraulis dussumieri Valenciennes, 1848, Hist. Nat. Poiss., 21: 69 (no locality; putative neotype described by Whitehead, 1967 : 142); Day, 1878, Fishes of India: 627, pl. 158 (4) (E. auratus in synonymy); Idem, 1889, Fauna British India, Fishes, 1 : 391.
Thryssa dussumieri: Whitehead (in press), Symp. Indian Ocean Adj. Seas. Mar. biol. Ass. India (key, synopsis, fig.).
Engraulis auratus Day, 1865, Proc. zool. Soc. Lond.: 312 (Cochin on Malabar coast; on specimen 438; inches = 117 mm); Idem, 1865, Fishes of Malabar: 238, pl. 19 (2) (repeat; fig. (? life-size) 44 inches = 114-7 mm tot. 1.).
MATERIAL.
a. ZSI (no specimens)
b. AMS (no specimens and none in Great Fisheries Exhibition Catalogue by Day, 1883)
NHMV (no specimens)
RMNH (no specimens)
ZMB.10412, a fish 89-7 mm S.L. (109-2 mm tot. 1.), ex Bombay
BMNH.1867.5.30.13, a fish 83:1 mm S.L. (99-4 mm tot. 1., caudal lobes damaged, estimated 103-3 mm), ex Madras, coll. Day
BMNH.1889.2.1.1779, a fish 90-6 mm S.L. (112-7 mm tot. 1., caudal complete), ex Malabar, coll. Day (outside label altered from E. auratus to mystax Gthr.) (label inside jar, Engraulis auratus Malabar), LECTOTYPE
ho ao
ga
78 P. K. TALWAR & P. J. P. WHITEHEAD
h. BMNH.1889.2.1.1780, a fish 55:2 mm S.L. (68-9 mm tot. 1., caudal complete), ex Canara, coll. Day (label in jar, Engraulis auratus Canara)
In the Fishes of Malabar (p. vi) it is stated that a specimen of E. auratus had been deposited in the East India Museum (but apparently not in the British Museum). Manuscript catalogues and lists of zoological material presented to the museum of the East India Company are now in the British Museum (Natural History). One list of fishes 1s headed “The following families are from Day’s Malabar Fishes’ and it includes Engraulis auratus from Malabar (preceded by the number 5, which seems to be an indication of the number of specimens). When the India Museum was dispersed in 1879, at least two of Day’s presentations (birds and fish skins) were sent back to Day. Since no fishes were given to the British Museum at that time, Day probably also received back his spirit specimens, including Engraulis auratus. It is possible, therefore, that the specimen (or one of them) is that now in London and presented in 1889 (specimen g above).
There is a discrepancy of 2-3 mm between the length of the single specimen described in the original description and the length of the figure, which suggests that the figure was not exactly life-size. Since all four extant specimens are too small and there can be no certainty which, if any, contributed to the original des- cription, we have chosen the British Museum Malabar specimen (g above) as lecto- type.
DEscRIPTION. Based on the LECTOTYPE, a fish 90-6 mm S.L. (112-7 mm tot. 1.) in good condition, ex Malabar, BMNH.1889.2.1.1779.
Br. St. 11, DI ili 10, Pi rz, Vi 6, A iii 32, g.r. 15 + 18, scutes 15 + 7.
In percentages of standard length: body depth 27-8, head length 26-9; snout length 3°6, eye diameter 6-4, length of upper jaw 41-1, length of lower jaw 19:3; pectoral fin length 18-1, pelvic fin length 11-1, length of anal base 33-0; pre-dorsal distance 49:0, pre-pelvic distance 42-2, pre-anal distance 59-3.
Body compressed, its width 3 times in its depth, the latter only slightly greater than head length; belly not strongly keeled, all but tips of scutes concealed by scaly sheath; head profile rising steeply from snout to nape and then more gradually to dorsal origin, belly profile evenly convex. Snout short, about = eye diameter. Upper jaw very long, the right maxilla pointed posteriorly, reaching to pelvic base and about # along pectoral fin (tip of left maxilla broken off); expanded portion of maxilla tapering rapidly behind 2nd supra-maxilla but with a membrane along upper edge; 2nd (posterior) supra-maxilla with upper part of expanded portion larger than lower; no Ist (anterior) supra-maxilla. Dentary symphysis below mid- point between eye and tip of snout; articulation of lower jaw 4 eye diameter behind 2nd supra-maxilla. A single series of fine conical teeth on dentaries, pre-maxillae and along lower edges of maxillae except near tip; two (right) and three (left) conical teeth on vomer; fine granular teeth on tongue, palatines and on endo- and ecto-pterygoids.
Gillrakers fine, slender, the longest 14 times length of corresponding gill filaments and almost equal to eye diameter; 8 short, triangular rakers on posterior face of 3rd epibranchial: gillraker serrae on both Ist and 2nd arches in distinct clumps,
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 79
with the longest serae in the middle of each clump. Pseudobranch present, con- cealed by skin but with small posterior opening. Isthmus silvery, not bearing anterior members of scute series, tapering evenly to just behind posterior margin of branchiostegal membrane. Operculum 3} times as deep as broad, its posterior margin evenly rounded and not completely covering gill opening. Dorsal surface of head covered by thick layer of skin with numerous pores; posterior frontal fontanelles present, long and narrow, 2:1 mm by 0-5 mm (right).
Dorsal fin origin nearer to snout than to caudal base by 1 eye diameter; fin preceded by a small scute-like plate bearing a retrorse spine. Pectoral fin reaching just over 3 along pelvic fin; axillary scale present, just over half length of fin. Pelvic fin base I eye diameter before vertical from dorsal origin and much nearer to pectoral base than to anal origin; axillary scale present, # length of fin; a second triangular scale present, below fin, } length of fin. Anal origin 3 eye dia- meter behind vertical from last dorsal ray. Caudal peduncle a little deeper than long.
Scales : distinct anterior and posterior ‘shoulders’ to scale; unexposed portion with 10-12 irregular vertical striae, not interrupted at centre of scale, exposed portion with one semicircular striation but more frequently reticulated, especially in posterior scales, the reticulations finally covering the whole scale. Many elongate scales at base of caudal but no true alar scales.
Colour: general body colour silvery/gold, but brown where scales lost ; dark brown venulose humeral area with peppering of dark pigment across back (as in Day’s figure—see Pl. 2c). Fins hyaline except for narrow dark posterior border to caudal.
Note. The long maxilla, absence of the rst supra-maxilla and distinct clumping of the gillraker serrae are characteristic of T. hryssa dussumieri (descriptions in Whitehead, 1967 : 142, fig. 14c and 1968 : 23, fig. 2a), and Day (1878) later placed his auratus in the synonymy of that species.
Pp. K. TALWAR & P. j- P. WHITEHEAD
80
(vruanjord san (Ztgr “Te A) * se osye) * * vivIvAoy] vsojVNIS| (I) zQI gtO ayy vadnia Legt “eA Aajswso1a, vyjaUipéavs — 9€9 4ajsvsoia) vagnig (vsouauan se) bz-6161'1'z (gZgr ‘Aeq) 6rSz7HNWY +SSorgWz 6ggsr'HWNaA zbola'sNV of9z [SZ sisuapurs vyjauipavs (z) €g1 gto stsuapurs vagn{D (4¥gr “ye A) * * * * * * vywiquiy vyjauripaws §(€) 191 LEO vywuquy vadna zSgi “IIE * * vuoskyovag vyjauripavs (€) €gt S€g wmosdyavaq vadnjo (6zgr “ang) * * * vanunjau vyauipavs (S$) tgr 9€9 vpnvriv vadnia (6zgr “and) * DANUDIAUL DIJIUIPADS — tg vanunjam vagdnia Legr “eA * * * * * sdaa1suo) vyjauripawg (z) 191 LEQ sdansuo, vadnja (4€g1 “ddnay) * (stsuasanjou se) * * snyojaund sXyjyoisjopysazy (1) E91 9£9 vazunyy vadnjg (eeutedny9) Legit “Tea 1yjassvy (sapiosdoja se) * * vinav viaauunssnq (S$) 991 Ltg Diuatuinssnq Lrgr “TRA * . * * * pinoy visatunssnq (bt) 991 Lhg DvnIv viAaLUNssNng (aeurierurmnssnq) avdiadn To (SZZ1 “yssi0,q) * x * * * qvaop snaquazoary (€) 991 zSQ quaop snAquaz0Ai14D AVGININADONIHOD urpiog. PUUeTA uopuoTy Saupss e}WNIeg vipuy fo saysty uapiay “sn sn “sn “sn *f@AINS uor}eOyIyUSP] ‘sy «6 oded ut seeds
“snusy{Qy [007 -3styanyzeN Jug. “qsny [007
(sadA} ajqissod 10f uaAt3 Joquinu Uor}eI4sI8e1) WOTDa][OO ut Juesaid,
uepley pue ulpieg ‘euusr,~ ‘uopuo7y ‘AaupAs ‘ezQn9]e9 Ul [eLIazeu Aeq pue wpuy fo saysry sAeq ut saysy proadnyo ayy,
I alavy
81
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY
98St HN
9z4z HNWY
€rbtorgwz
*
(£g -d ose aes)
(voyasip se)
* 6Lgr1z “6881' HNWA
(sn1j409 se)
*
gt'bi-9 ‘oLgr HNWA
S—ofoz'1°z
“638 HNWA gboz'1'z “638l HNWA
(voysjtp se) *
* * LEolqSWV $69z'ISZ
* *
9Lola'SWV SbzzISZ
88784 SNV
9hzz SZ
(6€gr “ureas) woipur DYysuT (6€gr “urems) vaipur nysyy (6€gr “ureas) vILpuL vYSUT (o€gr “uueg) Dywsuoja vysyy (o€gr “uuag) vywsuoja vysyy
Zhgr “ea vjayIpp vuonjag Leer “Tea
wjayrpip vuonag
(zzgi “yong ‘urezz) DPUNIv_YI DULOJSOJUOPOU Fr (69g1 ‘Aeq) DISapoul vsojv1U0+ (zzgr “yong ‘wezz) Durnuunu vsoj]MI1U0y (S621 ‘yoorq) SNSDU VSO]DIVIMa
(69gr ‘feq) Djvsa1Ava wisnpNy
(zzgr “yong ‘wezq) vadvys visnpny
(4g “yea) 40} DSILH
(zzgi “yong ‘urezz) YSU OSIVET
(6zgr “ang) 99124 DSIUT
6z61 ‘ereseAtueiaq syyoiany vavary a
(€Zgr ‘Aeq) DIVADQvIVUL VIaKD
ZZBI “Yong “wey DUAOQOS DILAOD
(§)
(S)
I9QI
zg
sho
zbg
vutoskyonaq vuojag snyoum vuoyag vIupur wuojjag Yynvuayssay Duojag vyosuoja vuoYag tuanaoy vuonag
vjaynip vuoyag (seuriezses1yst1q)
vpunovy9 snssaojvyD Snysapout snssaojvy 9 Dunuunue snssaojvyD
Snspu snssaojvyD (aeuryeurI0so10q)
vywsarava vadnt9 vadoys vadnia 104 vadnya
vysyr vadnyo vyansouvy vadrig
(eeutsory)
SNILADQDIDUL
sapropjaywads (eeurpnuojeg)
DUAOGOS VILAOD
PAK. TALWAR & PP. j. P. WHITEBEAD
82
*
(snqvann) ziFol'gqwz
* (snqoanv) E1rof'S
‘Lost HNWA
*
(uqjassna se)
* gf'b19
‘ofgl HNWA
*
zL9zISZ
(oSgr ‘10}ue5) sdazinasq vuurdryas
(zzgr “yong ‘urezyz) vsvyd vuurdyas
(grgr “TeA) Ayo vuurdyas
(6rgr “111q) sisuajpmuiDy vsstAy T
(S6Zr “yoorq) DILADQujvUL DSSAAY
(S€gr ‘Aviy) 1UuopuUDy vSshAy T
(grgr “Te A) taatunssnp vsshAy T
(zzgr “yong “ureyz) vavang vsshay J.
(rogr “uy9s) xvjstua vsstay T
(zg “ssnorg) SUAJSOAIJAS DSSKAY T.
(SZZ1 “yss10,7) DULYIAVG DUISSIAY T,
(€ogr ory) UUwosAaMmor snsoYygajojS
(€zgr ‘sseyy ura) snovpur snaoydajoys
(zSgr “r4Iq) 14q snsoydajojs
1€gr ‘Avi DUDIaSSNA DPUOIDY
(6zgr “ang) asoopan} snaazgoyssidc,
(6981 ‘Aeq) 1mapyjs DYSYT
(6€gr ‘urems) vaaqojwsaut DYsty
(6€gr ‘“urems) vaajqoppsaum vysyy
829
fto
Sto
sdarinadq synvasug vavay SYNDATUA Ayo} synvssug sisuajpuuDy synvasud snaLnqnjpur
synvAsUq
UUWONUMDY SYNDATUT
lalunssnp SYNnvASUT vavand sinvasuq xvjshut synvssug SUISOAYAS SINDASUT punjaog synvasugq SnuDuossauuLor sunvasuq snoipUr synvAsuy
14q SYNDATUT HVATITNVYONA
DUDIIJASSNA DPUOIDAT Aoozany snsazqoyjsidQ luapyys DUuOTIaIT vaasyy Duojag
vaajgopsaum vuonag
83
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY
setoads z€ ‘uaploT
saroeds gr
“LLgr ‘1'z'6ggr ‘TN jo suaurtoads s,Aeq Jo auQ “snswu- py 04 asor
soroeds €z ‘PUUST A
soroeds 1b ‘uopuoT
satoads oF
saroads 6€ Aeq Aq peztuSosex asoy ysduoure saroads prea ob
shgi “Tea taaluinssnp v1y109
grgr “Tea ¢wpypucar vytog
grgr “Tea wouter vytog
(zzgr “yong ‘wrezz) YVADIUIDA D1Y10D
(zzgi “yong -urezz) YVADIUMIDA DYL0D
(tzgr “yong ‘wezz) YVAVIUMIDA 117109
(z) 661
Wd ‘ereur) xa ‘sofoods mou sry} st wasn Ysytg 94} UT susvu: N 9 ‘psojvjvuan Jo seeds Mou v Suiqiiosep st UOS[aN YJoIeD “Id : aj0N
eq Aq pays seroeds proednyo S¢
taaunssnp v1Y4109
stsuaausog vYyt0D
wouter vyiog
sypuusasvaponb mi109 S1dojuna vY10D
YVADIULDA DI10D)
84 P. K. TALWAR & P. J. P. WHITEHEAD
REFERENCES
ANON. 1885. Australian Museum (Report of the Trustees for 1884). Govt. Printer, Sydney, 46 pp. App. XI (p. 29)—report by E. P. Ramsey; (pp. 42—46)—list of Day specimens. 1886. Australian Museum (Report of the Trustees for 1885). Govt. Printer, Sydney, 17 pp. (p. 5)—list of Day’s specimens, cases II and III. Cuan, W. L. 1965. <A systematic revision of the Indo-Pacific clupeoid fishes of the genus Sardinella (Family Clupeidae). Jap. J. Ichthyol., 12 (3-6) : 104-118; Ibid., 13 (1-3) : 1-39. Day, F. 1878. The fishes of India, being a natural history of the fishes known to inhabit the seas and freshwaters of India, Burma, and Ceylon. William Dawson & Sons, London, Pt IV, pp. i-xx, 553-778 and pls. 134-195. 1883. Gyveat International Fisheries Exhibition, London, 1883. Catalogue of the exhibits in the Indian Section. London, 198 pp. (list of Day’s own specimens, pp. 176-197). 1889. The fauna of British India, including Ceylon and Burma. Fishes, 1. Taylor & Francis, London, 548 pp. FowLer, H. W. 1941. Contributions to the biology of the Philippine Archipelago and adjacent regions. Bull. U.S. natn. Mus., 13 (100) : 1-879. Jenkins, J. T. 1910. Notes on fish from India and Persia, with descriptions of new species. Rec. Indian Mus., 5 : 123-140. Lioyp, R. E. 1907. Notes on a collection of marketable fish from Akyab, with descriptions of a new species of Lactarius. Rec. Indian Mus., 1 : 219-231. Menon, A. G. K. & Yazpani, G. M. 1963. Catalogue of type specimens in the Zoological Survey of India. Rec. zool. Surv. India, 61 : 91-190. Mortwanl, M. P., Jayaram, K. C. & SeHcaL, K. L. 1962. Fish and fisheries of Brahmaputra river system, Assam. Tvop. Ecol., 3 (1-2) : 17-43. Pott, K., WHITEHEAD, P. J. P. & Hopson, A. J. 1965. A new genus and species of clupeoid fish from West Africa. Bull. Acad. rv. Belg. Cl. Sci., (5) 51 : 277-292. Recan, C. T. 1917a. A revision of the clupeoid fishes of the genera Sardinella, Harengula, etc. Ann. Mag. nat. Hist., (8) 19 : 377-395. 1917b. A revision of the clupeoid fishes of the genera Pomolobus, Brevoortia and Dorosoma and their allies. Ann. Mag. nat. Hist., (8) 19 : 297-316. Srivastava, G. J. 1968. Fishes of eastern Uttay Pvadesh. Vishwavidyalaya Prakashan, Varanasi, India, i—xxii, 1-163. Tatwar, P. K. & CHAKRAPANY, S. (Inpress). Catalogue of the fishes figured in Day’s Fishes of India and deposited in the Zoological Survey of India. Rec. zool. Surv. India WHITEHEAD, P. J. P. 1962. <A review of the Indo-Pacific gizzard shad genera Nematalosa, Clupanodon and Konosivus (Pisces: Dorosomatidae). Bull. Br. Mus. nat. Hist. (Zool.), 9 (2) : 87-102. 1963. <A revision of the recent round herrings (Pisces: Dussumieriidae). Bull. By. Mus. nat. Hist. (Zool.), 10 (6) : 305-380. 1965. A preliminary revision of the Indo-Pacific Alosinae (Pisces: Clupeidae). Bull. Br. Mus. nat. Hist. (Zool.), 12 (4) : 115-156. 1967. The clupeoid fishes described by Lacepede, Cuvier and Valenciennes. Bull. Br. Mus. nat. Hist. (Zool.), Suppl. 2 : 1-180. 1968. Indian Ocean anchovies collected by the Anton Bruun and Te Vega, 1963-64. J. Mar. biol. Ass. India, 9 (1) : 13-37. 1970. ‘The clupeoid fishes described by Steindachner. Bull. Br. Mus. nat. Hist. (Zool.), 20 (1) : 1-46. (in press). A synopsis of the clupeoid fishes of India. Symp. Indian Ocean Adj. Seas. Mar. biol. Ass. India. WHITEHEAD, P. J. P., Borsreman, M. & WHEELER, A. 1966. The types of Bleeker’s Indo- Pacific elopoid and clupeoid species. Zool. Verhandl. Leiden, No. 84 : 1-152. WuitLey, G. P. 10958. List of type specimens of recent fishes in the Australian Museum, Sydney. (Unpublished typescript, 40 pp.)
THE CLUPEOID FISHES DESCRIBED BY FRANCIS DAY 85
ADDENDUM In 1889 the Zoological Institute in Leningrad received a large collection of Day Fishes, comprising 357 specimens (284 species). These were a gift from the British Museum (Natural History), presumably taken from the large collection presented to the Museum just before Day’s death. Amongst these specimens are 32 clupeoids, including a paralectotype of Spratelloides malabaricus (see p. 64), making a total of 1g clupeoid species.
P. K. Tatwar, Ph.D. ZOOLOGICAL SURVEY OF INDIA 34, CHITTARANJAN AVENUE CatcuTtTa-12, INDIA
P. J. P. WHITEHEAD, B.A. Department of Zoology
British Museum (NaturaL History) CROMWELL Roap
Lonpon, SW7 5BD
PLATE 1
a. Clupea sindensis, Fishes of India, pl. 163 (2), F. Day del., 119 mm tot. 1. (= Sardinella sindensis).
b. Spratelloides malabaricus, Fishes of India, pl. 161 (5), C. Achilles del. et lith., 69 mm tot. 1 (= Dayella malabarica).
c. Clupea variegata, Fishes of India, pl. 161 (4), C. Achilles del. et lith., 197 mm tot. 1. (= Gudusia variegata).
PAE, x
Mus. nat. Hist. (Zool.)
Bull. Br.
PI AME, ¥2 Chatoessus modestus, Fishes of India, pl. 160 (1), C. Achilles del. et lith., 132 mm tot. 1. (= Gonialosa modesta). . Pellona sladeni, Fishes of India, pl. 164 (1), F. Day del., 177 mm tot. 1. (= Ilisha sladeni).
Engraulis auratus, Fishes of Malabar, pl. 19 (2), F. Day del. et sculp., 115 mm tot. 1. (= Thryssa dussumiert).
PLATE 2
Mus. nat. Hist, (Zool.) 22, 2
Bull. Br.
A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY)
Kay, E. Atison. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75.
WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier and Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. {4.
Taytor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure of Mineralogy at the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates, 77 Text-figures. 1969. £4.50.
Haynes, J. R. Cardigan Bay recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. (In press.)
, Br Printed in England by Staples Printers Limited at their Kettering, Northants, establishment
3B. E. BROOKER
ee BULLETIN OF EUM (NATURAL HISTORY) 2 | Vol. 22 No. 3 LONDON : I97I
of
FINE STRUCTURE OF BODO SALTANS AND
BODO CAUDATUS (ZOOMASTIGOPHORA :
PROTOZOA) AND THEIR AFFINITIES WITH THE TRYPANOSOMATIDAE
BY
BRIAN EDWARD BROOKER
Nuffield Institute of Comparative Medicine, London
Pp. 87-102; 6 Plates, 1 Text-figure
BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 22 No. 3 LONDON : 1971
THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical series.
Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year.
In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department.
This paper is Vol. 22, No. 3 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals.
World List abbreviation Bull. Br. Mus. nat. Hist. (Zool.).
© Trustees of the British Museum (Natural History), 1971
TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY)
Issued 31 December, 1971 Price {1:20
FINE STRUCTURE OF BODO SALTANS AND
BODO CAUDATUS (ZOOMASTIGOPHORA :
PROTOZOA) AND THEIR AFFINITIES WITH THE TRYPANOSOMATIDAE
By B. E. BROOKER
CONTENTS Page SYNOPSIS : 3 : : é 3 5 : E > z 89 INTRODUCTION c : A “ 2 5 : : : 89 MATERIALS AND METHODS . 5 : : : * 3 : : 90 RESULTS ; 4 0 C 0 6 5 6 5 : " 90 Light microscopy O 2 : 7 3 : : é go Bodo saltan. 5 5 : : 4 F a 90 Bodo caudatus . : 6 c g é é 92 Electron microscopy . é é A c ° ¢ ¢ a 92 Flagellar pocket . 5 : 0 : 92 Basal bodies and flagella. . : 7 3 92 Alimentary system . é 3 é c ¢ 93 Microtubular systems : c c : 5 94 Kinetoplast-mitochondrion c 5 < : 95 Nucleus . 6 ; 2 6 4 4 96 Endocytoplasmic bacteria . . 5 c ° 96 Cytoplasmic membrane systems . ¢ : : 96 DIscussIoN. 5 0 ¢ : ¢ : : : 5 5 97 ACKNOWLEDGEMENT A : : 5 6 . : A : IOL REFERENCES . 0 : 6 ; 5 5 : : 6 IoI SYNOPSIS
Characters specific to each of two species of Bodo are described for the first time. Bodo saltans possesses cytoplasmic bacteria, hair-like appendages (mastigonemes) on the anterior flagellum and circumbuccal lappets surrounding the opening of the alimentary system. In B. caudatus, an electron dense band separates the kinetoplast from the basal bodies of the flagella. In addition, clear differences exist between the microtubular systems associated with the buccal cavity and cytopharynx. It is suggested that the mastigonemes and circumbuccal lappets of B. saltans are responsible for the capture of food organisms.
In both species the alimentary system is a membrane-lined tube surrounded by a number of microtubules. The single mitochondrion is dilated in the vicinity of the basal bodies and contains a prominent kinetoplast. Since both these organelle systems closely resemble those found in members of the related Trypanosomatidae, the possible origin of this family from Bodo or a Bodo-like flagellate is discussed.
INTRODUCTION
Bodo is a cosmopolitan flagellate found in fresh and brackish waters and in some soils. This and other members ot the Bodonidae are of interest chiefly because of their close relationship to the economically and medically important trypanosomes.
go B. E. BROOKER
This relationship is based on the presence of a mass of DNA—the kinetoplast— situated in a dilatation of the single mitochondrion. In trypanosomes, this organelle has been regarded as a genetic system containing the information required for the synthesis of mitochondrial enzymes (Steinert, 1g60) but in Bodo its function has not been examined. Of the many species of Bodo which have been described, Bodo saltans Ehrenberg, 1831 and Bodo caudatus Stein, 1878 are probably the two most commonly found. They occur in quite different habitats for whereas B. caudatus tends to be coprozoic, B. salians is usually found in freshwater. Separation of the two species depends on such characters as body shape and size, length of the flagella and the position of the nucleus relative to that of the kinetoplast. However, the present fine structural study describes a number of clearly defined qualitative differences between B. saltans and B. caudatus which would not have been visible to the earlier light microscopists.
MATERIALS AND METHODS
Bodo saltans was isolated from a sample of fresh water taken from a pond near Slapton Ley, Devon. From this isolate, clone cultures were established on 0-1% w/v ‘Oxoid’ dehydrated liver intusion (pH 4-6) and maintained at 25°C.
Bodo caudatus was isolated from an infusion of pig faeces which was obtained from Winches Farm, near St. Albans, Hertfordshire. Clone cultures from this isolate were maintained at 25°C on 0:2% w/v ‘Oxoid’ beef extract (pH 5-8). In both cases, cultures were agnotobiotic. For light and electron microscopy, cultures were harvested after 3 days growth.
Light microscopy.—Phase contrast observations were made using a Leitz Ortholux microscope fitted with a Heine condenser. Flagellates were examined either alive or after fixation with 1°/, osmium tetroxide. Smears fixed in Schaudinn’s fluid were stained with iron haematoxylin and examined by bright field microscopy.
Electron microscopy.—Flagellates were collected by centrifugation at 1,000 r.p.m. for 10 minutes and the resulting pellet fixed for 5 minutes at room temperature in 1% osmium tetroxide buffered to pH 7-4 either with o-1 M Sorensen’s phosphate buffer or 0-1 M veronal acetate (Michaelis). Before dehydration, the pellet was treated with 1°% uranyl acetate in 25% ethanol for 30 minutes. After dehydration in ethanol-water mixtures and absolute ethanol, the pellet was treated with propylene oxide or toluene and embedded in Araldite. Sections were cut using a Porter-Blum MT2 ultramicrotome and stained in lead citrate prior to examination in an EM 6B electron microscope.
Some flagellates were fixed as before, washed in distilled water and dried onto grids. They were then placed in a coating unit and shadowed with gold/palladium at an angle of 30°. Negatively stained preparations were made using sodium phosphotungstate at pH 7-0.
RESULTS
LicHt mMiIcroscopy.—Bodo saltans.—The body ot this flagellate is oval in shape (5-8 um long and 2-5 um wide) and has two flagella of unequal length which arise from the bottom of a depression (the reservoir or flagellar pocket) near the anterior
STRUCTURE OF BODO SALTANS AND B. CAUDATUS gl
end of the cell (Fig. 1). During locomotion in which the body rotates on its own axis, the shorter anterior flagellum is extremely active while the posterior flagellum remains stationary or undergoes slight movement. When stationary, the flagellate often attaches itself to some object by the tip of its trailing flagellum and may then exhibit rapid oscillations. This behaviour is characteristic of B. saltans. The kinetoplast lies near to the basal bodies of the flagella and the nucleus, which con-
\ WH tah | )
LAPPETS
FLAGELLAR POCKET
KINETOPLAST
CYTOPHARYNX BUCCAL CAVITY
] y \ 2, CONTRACTILE Yr \ VACUOLE \
CYTOPLASMIC BACTERIA
MITOCHONDRION
FLAGELLA
NUCLEUS
FOOD VACUOLE
Fic. 1. Diagram of Bodo saltans showing the arrangement of the main organelles.
tains a conspicuous central karyosome, is central or mid-ventral in position. The single mitochondrion is markedly siderophilic and appears as a thin filament which originates from one side of the kinetoplast, describes a figure of eight or a loop within the cell and terminates at the opposite side of the kinetoplast. A single, round contractile vacuole appears at intervals at the anterior end of the living cell. It is situated just below and to one side of the flagellar pocket into which it periodic- ally discharges its contents. The posterior half of the flagellate is occupied by a variable number of food vacuoles.
92 B. E. BROOKER
The opening of the alimentary system, the buccal cavity, is marked by a small vacuole on the ventral surface near the anterior end of the flagellate. Ingestion of bacteria is a very rapid process and can be observed satisfactorily only during the short periods of quiescence when it is attached to the substratum by its posterior flagellum. During ingestion, a bacterium is drawn into the buccal cavity and rapidly passes ventro-dorsally along a path which corresponds exactly with that of the cytopharynx (see later). A large vacuole then appears at a point which prob- ably corresponds to the end of the cytopharynx and the bacterium passes into it. This food vacuole then slowly moves to the posterior end of the cell.
Bodo caudatus. When harvested in the logarithmic phase of growth, B. caudatus is long and narrow with a convex dorsal and concave ventral surface (8-14 ym long and 4-6 um wide). The two flagella are of unequal length and arise trom the bottom of the flagellar pocket near the anterior end of the cell. The anterior flagellum is the shorter but most active during locomotion. Situated in the anterior half of the cell, the nucleus frequently lies very close to the kinetoplast. As in B. saltans, the single mitochondrion is continuous at both its ends with the kineto- plast and describes a loop-like circuit of the cell. Almost the entire posterior half of the flagellate is occupied by food vacuoles and an active contractile vacuole lies close to the flagellar pocket. Although anteriorly a small buccal cavity is clearly visible, the ingestion ot bacteria is very difficult to observe because of the relentless swimming habit of this flagellate.
ELECTRON MICROscopy.—Flagellar Pocket. In both species of Bodo, the flagellar pocket is a lateral depression at the anterior end of the body. Because the two flagella emerge from the cell at the bottom of this structure (Fig. 1, Pl. 1, Fig. A), Pitelka (1961) referred to it as the circumflagellar depression. The flagellar pocket is lined with a unit membrane continuous with that covering the rest of the cell and below this, there lie a number of pellicular microtubules. Further details of these are given below.
Basal bodies and flagella. The two basal bodies are embedded in the cytoplasm at the base of the flagellar pocket and are therefore antero-lateral in position. They are structurally differentiated into 2 regions, an intracytoplasmic proximal portion and an extracytoplasmic transition zone which connects the intracytoplasmic portion to the flagellar shaft and is bounded by the flagellar membrane (PI. 1, Fig. A). The proximal end of the basal bodies lies very close to the surface of the kinetoplast capsule (Pl. 1, Fig. A). The gap between these two structures is some- what variable in width but is generally in the order of 100 nm and direct contact has never been observed. Under some conditions of fixation, a reticulum of fine filaments (8 nm wide) fills each basal body.
The junction of the basal body transition zone with the flagellum is marked by two transverse basal plates which are thickened peripherally (Pl. 1, Fig. B). The two central tubules of the axoneme originate just above the distal basal plate and the paraxial rod—a structure which runs parallel to the axoneme for about three quarters of its length—arises from a lateral extension of the proximal basal plate. The basal plate extensions of each flagellum face one another and the overlying flagellar
STRUCTURE OF BODO SALTANS AND B. CAUDATUS 93
membranes are joined by a wide but very thin extracellular striated band (PI. 1, Figs B and C). Of the three closely spaced striations, only the central one (5 nm wide) is prominent. Those lying on either side of it are more diffuse and can be satisfactorily resolved only in longitudinal sections of the band.
The transition zone of the basal bodies consists of 9 peripheral doublets of tubules each of which is joined to the adjacent flagellar membrane by a narrow connective. The intracytoplasmic portion is composed of g tubule triplets and is open proximally. The proximal ends ot the basal bodies are connected by three striated rootlets (Pl. 1, Fig. D). Two of these arise from a single triplet on the basal body of the posterior flagellum and diverge slightly as they approach and insert onto the other basal body. A third rootlet running approximately parallel to the other two also connects the basal bodies. All three rootlets have a major period of 50 nm. The prominent bands delimiting the major repeating unit are 25 nm wide and between these lies a narrower 5 nm wide band. There is in addition, a ‘Y’ shaped rootlet which arises from the basal body of the anterior flagellum and passes antero-laterally to insert onto a group of 2-3 short microtubules running parallel to the cytopharynx (PI. r, Fig. D).
Both flagella contain the familiar ‘9+2’ arrangement of tubules and a paraxial rod (Pl. 1, Fig. E). After running parallel to the axoneme for about three-quarters of its length, the paraxial rod tapers before terminating. Details of its structure are difficult to resolve but it appears to have a lattice-like architecture. The anterior flagellum of Bodo saltans, unlike that of B. caudatus, bears mastigonemes. These hair-like appendages are arranged in bundles along one side of the flagellum and have been reported elsewhere (Brooker, 1965). Metal-shadowed preparations of B. saltans show that the posterior flagellum bears a number of parallel transverse striations whose separation is 14 nm (PI. 1, Fig. F). They are found only after the flagellum has emerged from the flagellar pocket and extend distally for a distance of 3um. In B. caudatus similar striations first appear at a level corresponding to the junction of the basal body with the flagellum but because they extend distally only for a distance of about 1 ym, they are not visible in shadowed material. In sectioned material, the striations are seen as periodic thickenings (14 nm wide) of both leaflets of the flagellar membrane (PI. 1, Fig. E, Pl. 2, Fig. A). Such thickenings occupy about 25% of the circumference of the flagellar membrane (PI. 1, Fig. E). On passing distally, this percentage gradually decreases so that in shadowed prepara- tions of B. saltans, the array of striations appears to taper to an end (PI. 1, Fig. F).
Alimentary system. The ingestion of food particles by Bodo takes place by way of permanent oral structures—the buccal cavity and cytopharynx—which lie to the right of the flagellar pocket (Pl. 2, Fig. B). The position at which the buccal cavity opens on to the surface of the cell differs in the two species. In B. saltans, it opens . antero-ventrally (Pl. 2, Fig. C) but in B. caudatus it is found at the extreme anterior end of the flagellate (Pl. 2, Fig. D). Very rare sections in which a bacterium is found in the buccal cavity (PI. 1, Fig. A) suggest that this organelle is capable of considerable distension. Frequently, the membrane lining the buccal cavity has a pronounced cell coat (Pl. 2, Fig. C) which takes the form of bundles of fine filaments projecting perpendicularly from the membrane. A number of structures, referred
94 B. E. BROOKER
to here as circumbuccal lappets, surround the opening of the buccal cavity of B. saltans (Pl. 2, Fig. B). They are flat, triangular projections (Pl. 3, Fig. A) which are joined at their bases to form a ring. Each lappet is composed of many fine filaments and arises from an intracellular band at the margin of the buccal cavity. These structures appear to be totally absent from B. caudatus.
From the left side of the buccal cavity arises an elongate tube lined by a unit membrane and surrounded by a number of microtubules (Pl. 2, Fig. B). In accord- ance with the terminology used for the oral apparatus of ciliates, this organelle will be referred to as the cytopharynx. In Bodo saltans the cytopharynx approaches the kinetoplast, passes underneath and to one side of it and on reaching the dorsal surface of the body, curves to one side (Pl. 3, Fig. C). In B. caudatus, it passes posteriorly just below the cell membrane (Pl. 2, Fig. D). Although the cytopharynx terminates just posterior to the kinetoplast, the microtubules associated with it frequently reach the posterior margin of the nucleus before terminating. In both species, the diameter of the lumen decreases distal to the buccal cavity. Numerous vesicles of variable diameter are always found underneath and to the right side of the cytopharynx in B. saltans (Pl. 2, Fig. B, Pl. 3, Fig. C). Although similar vesicles are occasionally found in B. caudatus, they are generally smaller and less numerous. Their origin is not clear, but many longitudinal sections of the cyto- pharynx suggest that they arise as invaginations of the cytopharyngeal membrane.
Both light and electron microscope observations suggest that ingestion of food particles takes place at the blind end of the cytopharynx. Since the diameter of the cytopharynx is at all points along its length smaller than that of food organisms found in the food vacuoles, ingestion must be accompanied by considerable dis- tension of the cytopharynx. Ingestion results in the formation of a number of food vacuoles which migrate to the posterior half of the cell. The food vacuoles are bound by a single unit membrane and may contain one or more food organisms (Pl. 5, Fig. D). When digestion is complete, only a diffuse mass of undigestible material remains in the vacuole. Since such vacuoles do not accumulate in the cytoplasm, it seems likely that undigestible material is voided by the coalescence of the cell membrane with that of the food vacuole. However, this has never been observed.
Microtubular systems. Microtubules of external diameter 20-25 nm are associated with several organelles in the anterior half of the flagellate. Many, but not all, of the tubules which are associated with the flagellar pocket, cytopharynx and body, appear to arise from the proximal ends of the two basal bodies.
Bodo saltans. The microtubules which are to surround the cytopharynx pass along the flagellar pocket in two groups. At the opening of the flagellar pocket, three members of each group approach each other (Pl. 3, Fig. B) and continue as microtubule doublets. They curve over the narrow bridge ot cytoplasm separating the flagellar pocket from the alimentary system and enter the walls of the buccal cavity where they adopt a ‘U’ shaped configuration. From the left side of this cavity, the microtubules follow the course of the cytopharynx and become arranged around it in two groups. Associated with the floor and left side of the cytopharynx is a group of 5 microtubules and with the roof a group of 3 (Pl. 3, Fig. D). The right
STRUCTURE OF BODO SALTANS AND B. CAUDATUS 95
side of the organelle is free of tubules. The middle 3 tubules of the group of 5 are double and joined on their left side to the cytopharyngeal membrane by a short connective. Although 8 microtubules normally run parallel to the cytopharynx, 9 or Io are sometimes found (Pl. 3, Fig. E). The microtubules of each group are connected by a number of fine filaments which occur at intervals (13 nm) along their length (Pl. 4, Fig. A). Beyond the blind end of the cytopharynx, the connectives of the microtubule doublets disappear but the arrangement of the two groups of tubules remains constant until they terminate near the dorsal surface of the flagellate.
Beneath the cell membrane at the anterior end of the cell lie a number of pellicular microtubules. Although some of these arise from the basal bodies of the flagella, many appear to have their origin near the opening of the flagellar pocket. From here, the microtubules spiral round the anterior end ot the cell and on approaching the right side of the buccal cavity dip below the cell surface (Pl. 4, Fig. B). They then describe a semi-circular path as a curved band of 15-20 tubules (Pl. 4, Fig. C). Passing below the buccal cavity (Pl. 3, Fig. C) they move anteriorly and terminate in the cytoplasm between the cytopharynx and the flagellar pocket.
Bodo caudatus. The arrangement of microtubules associated with the buccal cavity resembles that described for B. saltans. The tubules which are to surround the alimentary system pass along the wall of the flagellar pocket and enter the buccal cavity. Here, transverse sections show that the tubules are arranged in a line along the left wall (Pl. 4, Fig. D, Pl. 5, Fig. A). As the tubules follow the cytopharynx from the left side of the buccal cavity, they distribute themselves around the cytopharyngeal membrane (PI. 5, Fig. B) in 2 overlapping groups. One group associated with the roof of the cytopharynx usually contains 4 microtubules; the other group, which lies next to the floor and left side of the organelle, may con- tain 4, 5 or 6 tubules three of which are double and joined as in B. saltans to the cytopharyngeal membrane by a short connective. Such connectives disappear a short distance from the buccal cavity.
Another set of microtubules emerges from the flagellar pocket, spirals round the anterior end of the cell and comes to occupy the right hand wall of the buccal cavity just below the lining membrane (PI. 4, Fig. D). The 15-20 microtubules in this set are joined by intertubular connectives and for most of their length travel posteriorly parallel to the buccal cavity and cytopharynx. A limb of the single mitochondrion which passes beneath the buccal cavity modifies the path taken by some of the tubules. As they travel posteriorly, the tubules gradually move away from the cytopharynx (Pl. 5, Figs A, B) and towards the cell membrane until, at a level beyond the blind end of the cytopharynx, they are seen as a row of pellicular micro- tubules. These tubules travel some distance posteriorly before terminating. In addition to this major set, an equally conspicuous row of very short parallel tubules is found on either side of the buccal cavity (Pl. 4, Fig. D).
Kinetoplast-mitochondrion.—Lying just below the cell membrane, the single mitochondrion describes a loop-like circuit of the cell (Fig. 1). In Bodo saltans, this organelle travels above and parallel to the cytopharynx, describes a semi-circular path around the right side of the buccal cavity (Pl. 5, Fig. D) and passes to the extreme posterior end of the cell before returning anteriorly to the basal body region.
96 B. E. BROOKER
The whole circuit of the mitochondrion frequently takes the form of a figure of eight. The form of the mitochondrion in B. caudatus is similar to that of B. saltans except that in the posterior region of the former it is frequently seen as a branching structure.
The cristae are predominantly plate-like and arise from the inner mitochondrial membrane. The granular matrix of the mitochondrion is dense but sometimes contains irregularly shaped bodies of much greater electron density (PI. 5, Fig. C).
In the region of the basal bodies, a prominent spherical dilatation of the mito- chondrial tube houses the kinetoplast. This will be referred to as the kinetoplast capsule. The kinetoplast is composed of a complex reticulum of fine filaments (2°5-3-0 nm thick) and contains a large number of irregularly distributed electron dense nodes (PI. 6, Fig. A). It is separated from the wall of the mitochondrion by a number of cristae embedded in a thick layer of mitochondrial matrix (Pl. 6, Fig. A). The basal bodies of the flagella lie very close to the surface of the mitochondrial kinetoplast capsule but are never seen in contact with it. In Bodo caudatus, the basal bodies are separated from the capsule by a flat electron dense pad (PI. 1, Fig. A).
Nucleus.—tThe relative position of the nucleus within the:cell is one of the features by which the two species of Bodo may be separated. In B. saltans it is found mid- ventrally some distance from the kinetoplast capsule whilst in B. caudatus, it is always found very close to this part of the mitochondrion. The nucleus is bound by a nuclear envelope composed of two membranes of which the outer is continuous with the granular endoplasmic reticulum. A prominent and finely granular nucleolus of variable shape occupies the centre of the nucleus (Pl. 5, Fig. D). In B. saltans a layer of condensed chromatin is often seen attached to the inner membrane of the nuclear envelope but in B. caudatus, this component of the nucleus appears to be absent or at least is not visualized by the techniques used in this study.
Endocytoplasmic bacteria.—Structures have been observed in the cytoplasm of Bodo saltans which, on morphological grounds, have been tentatively identified as bacilliform bacteria (approximately I ym long and 0-3 wm wide). They are found in all individuals of the flagellate population and are always situated in the anterior half of the cell (Pl. 2, Fig. C, Pl. 4, Fig. C). Although the largest number of bacteria seen in one cell profile is four, the total population is probably much larger. They can be easily distinguished from food organisms since the latter are separated from the cytoplasm of the flagellate by the membrane of the food vacuole and are usually found at the posterior end of the cell (Pl. 5, Fig. D).
Each bacterial cell is bounded by a cell membrane 8 nm thick which bears on its outer surface a thin layer of filamentous material in direct contact with the host cytoplasm. Extensions of a layer of dense material lying beneath the cell membrane project into the electron lucent central portion of the bacterium which is traversed by fine fibrils. Deep, mid-length constrictions of the bacteria suggestive of division are commonly encountered and appear independently of the host cell division cycle (Pl. 6, Fig. C).
Cytoplasmic membrane systems.—The contractile vacuole is situated on the left side of the flagellate just below and to one side of the flagellar pocket (Pl. 4, Fig. D). Surrounding the vacuole is a number of vesicles and tubules which serial sections
STRUCTURE OF BODO SALTANS AND B. CAUDATUS 97
show to be continuous with the lumen of the vacuole (PI. 6, Fig. B). After systole, the membrane lining the vacuole appears rounded in section. At discharge, the membranes between the flagellar pocket and contractile vacuole coalesce and the contents of the vacuole are discharged into the flagellar pocket. The thin layer of cytoplasm between the membranes of the flagellar pocket and contractile vacuole is traversed by a number of concentrically arranged circular septa. Coated vesicles are commonly encountered in the vicinity of the vacuole and often they are seen with their membrane confluent with that of the vacuole.
The Golgi apparatus is situated directly below the cytopharynx (Pl. 4, Fig. C, Pl. 5, Fig. D) and to the right side of the contractile vacuole and is composed of a stack of 6-10 compressed saccules. Although vesicles of both the smooth and coated type are actively proliferated from the margins of all the Golgi saccules, there is frequently a notable concentration in the region of the distal saccule.
Cisternae of granular endoplasmic reticulum arise from the outer membrane of the nuclear envelope and ramify throughout the cell. One limb of this reticulum is permanently associated with the posterior margin of the cytopharynx and runs parallel to it for most of its length (Pl. 3, Fig. D).
DISCUSSION
Since, at the level of the light microscope, Bodo saltans and B. caudatus appear to possess the same major organelle systems, separation of the two species is based on differences in the spatial arrangement of these organelles and on differences in the size and shape of the body. Whilst confirming the validity of such criteria, the present study has shown that separation is also possible using characters which are beyond the resolution of the light microscope. Thus in B. saltans, circumbuccal lappets, endocytoplasmic bacteria and mastigonemes on the anterior flagellum are consistently present but are never found in B. caudatus. Similarly, the dense layer of material which separates the basal bodies of the flagella from the kinetoplast capsule is present only in B. caudatus. Although Pitelka (1961) was only able to make a tentative identification of the flagellate she studied, it is clear that it was B. saltans for her pictures show the cytoplasmic bacteria and circumbuccal lappets of this species.
The alimentary system and the microtubules associated with it have been briefly reported by Pitelka (1961). She pointed out that the non-contractile rostral vacuole described by Hollande (1942) corresponded to the cup-like depression at the opening of the alimentary system, a structure which has been referred to as the buccal cavity in the present study. Sections of the alimentary system containing partially ingested bacteria suggest that the buccal cavity and cytopharynx are capable of considerable distension. This conclusion is supported by the observations made by Sinton (1912) on a flagellate which he referred to as Prowazekia urinaria but which, from his description, was probably Bodo caudatus. He observed that the flagellate was able to ingest not only large bacteria but also red blood cells and that on these occasions the buccal cavity was capable of being greatly distended. As Sinton describes it, the path taken by the bacteria through the cell during ingestion
98 B. E. BROOKER
corresponds exactly to the course of the cytopharynx described in the present study. A similar conclusion has been drawn from the light microscope observations of feeding in B. saltans. These results are contrary to Pitelka’s assertion that bacteria do not pass along the cytopharynx. Although the function of the microtubules associated with the alimentary system is unknown, it is possible that they confer a degree of elasticity on the organelle which enables it to return to its normal shape and size once ingestion is complete. Schuster (1968) suggested a similar function for the cytopharyngeal tubules of the cryptomonad flagellate Cyathomonas truncata and extended this proposal to the case of Bodo.
The capture of prey by Bodo caudatus has been described by Sinton (1912). According to this author, the distal portion of the anterior flagellum is capable of grasping bacteria and propelling them to the opening of the buccal cavity by coiling movements. Infolding movements of the edges of the buccal cavity then initiate ingestion. This mechanism is possible in B. caudatus only because the buccal cavity opens anteriorly and is therefore ideally situated to receive bacteria carried to it by the anterior flagellum. Because the buccal cavity of B. saltans opens antero- ventrally, the mechanism described above for B. caudatus does not appear adequate to explain the capture of food organisms. Instead, a mechanism involving the mastigonemes of the anterior flagellum is proposed. Although the mastigonemes do not appear to play a major role in locomotion (Holwill, 1966), it is possible that during the oar-like movements of the anterior flagellum (Holwill, 1966) they exert a component force in the direction of the buccal cavity which sweeps food organisms towards it. Such a mechanism may bring bacteria to the vicity of the buccal cavity but the initiation of ingestion probably depends on movements of the margins of the buccal cavity as described by Sinton (1912) for B. caudatus. In this process, participation by the circumbuccal lappets may be important. It is visualized therefore that the mastigonemes and the circumbuccal lappets are functionally integrated to form a system responsible for the capture and ingestion of food organisms. Schuster (1968) has described a filamentous fringe surrounding the opening of the cytopharynx of Cyathomonas which, like the circumbuccal lappets of B. saltans, has its origin beneath the cell membrane and is believed to assist in feeding.
In her study of Bodo saltans, Pitelka (1961) suggested that the cytopharynx was a modified intracytoplasmic flagellum. This suggestion was based on the observation that the cytopharynx is surrounded by 9 microtubules and arises close to the surface of the kinetoplast near the basal bodies of the flagella. However, it has been shown here that the cytopharynx passes beyond the kinetoplast capsule and although occasionally surrounded by 9g or ro tubules, 8 is the usual number. In view of these findings, the homology attempted by Pitelka must be considered doubtful.
Vesicles of various sizes are associated with the cytopharynx of Bodo (Pitelka, 1961), Ichthyobodo(Costia)necator (Joyon and Lom 1966, 1969) and Cyathomonas truncata (Mignot, 1965; Schuster, 1968). In the case of Cyathomonas, Mignot (1965) believed that these vesicles arise from the Golgi apparatus and Schuster (1968) has suggested that they contain digestive enzymes which are ultimately emptied into the food vacuoles. In Bodo however, profiles showing an undulatory cytopharyngeal
STRUCTURE OF BODO SALTANS AND B. CAUDATUS 99
membrane strongly suggest that the vesicles arise by pinocytosis although in the absence of tracer experiments this can only be conjecture.
Although in most respects the flagella of Bodo closely resemble those described from other kinetoplastid flagellates (Pyne, 1960; Anderson and Ellis, 1965; Vicker- man, 1969), they do possess two features, namely the striations of the posterior flagellum and the extracellular interflagellar connective, which are not shared by other members of this group. The significance of these structures is not known, but an extracellular connection between the 2 flagella may go some way to explaining synchrony of these organelles during movement. The paraxial rod of both flagella arises from the proximal basal plate but in the closely related trypanosomatids it only becomes recognizable a short way along the flagellum.
The mitochondrion was probably first visualized by Whitmore (1911) who described a fibril from Bodo asiaticus (= B. caudatus) which ran from the kinetoplast to the posterior end of the cell. Alexeieff (1912) observed this ‘fibrille sidérophile’ in B. caudatus and B. edax but reported that it could only be seen in flagellates from young cultures. A detailed study of this structure was made by Hollande (1936, 1942). He noted that it was a constant feature of all species of Bodo but that there was a species difference in the extent to which it was developed. Because the ‘cote’ or ‘cordon siderophile’, as Hollande called it, was best developed in B. saltans, he paid more attention to this species and described in some detail the path taken by it through the cell. As noted by Pitelka (1961), it seems probable that Hollande was describing the mitochondrion since electron microscopy shows that this is the only organelle which follows an identical path through the cell.
The enclosure of the DNA-containing kinetoplast in a dilatation of the mito- chondrial tube is a feature uniting Bodo with members of the Trypanosomatidae. However, they differ in the organization of the kinetoplast for whereas in the Trypanosomatidae the kinetoplast is disk shaped with its component fibrils arranged antero-posteriorly, in Bodo it is spherical with its fibrils forming a reticulum. Although in all kinetoplastid flagellates the basal bodies of the flagella lie very close to the surface of the kinetoplast capsule, no physical connection between the two has been found. Such a connection has been sought because in many trypanosomatids the two structures are linked in morphogenesis and appear connected after cell rupture (Simpson, 1968). The present study suggests that in the case of Bodo caudatus no direct connection is possible because the two structures are separated by a thick electron dense pad. The observations made by Simpson (1968) on Leish- mania tarentolae led him to suggest that the kinetoplast capsule is attached to the basal body by an EDTA-sensitive cytoplasmic cement. There seems no reason why this explanation cannot be extended to other kinetoplastid flagellates.
Electron dense bodies similar to those found in the matrix of the mitochondrion have also been described from Crithidia fasciculata (Brooker, 1971) and in Tetra- hymena pyriformis Levy and Elliott (1968) found that they become more numerous when the ciliates are starved. Although their significance is unknown, it is inter- esting that these bodies resemble the altered kinetoplast DNA of trypanosomatid flagellates which have been exposed to acriflavine (Kusel, Moore and Weber, 1967; Hill and Anderson, 1969). Although dyskinetoplastic Bodo caudatus was obtained
100 B. E. BROOKER
by Robertson (1929) using acriflavine, the electron miscroscopy of the kinetoplast after this treatment has not been studied.
The endocytoplasmic bacteria of Bodo saltans appear to be a constant feature of this species. Profiles showing constrictions across the equator of some bacteria support the assumption that they multiply to keep pace with division of the flagellate. Gram stained preparations of B. saltans provide little useful information since it is difficult to distinguish endocytoplasmic bacteria from food organisms. However, since the walls of Gram negative bacteria are very thin (Kellenberger and Ryter, 1958) compared with those of Gram positive bacteria (Glauert, 1962) the cytoplasmic bacteria of Bodo are judged to be Gram negative.
Bodies thought to be bacteria constantly occur in or on many species of protozoa as shown by the review of Kirby (1941). In most cases their identity and function is unknown. Since, in the case of Bodo saltans, there is no evidence to suggest that the flagellate benefits or is harmed by the association it is impossible at this stage to decide whether the bacteria are parasitic or symbiotic. However, in another kinetoplastid flagellate, namely Crithidia oncopelti, cytoplasmic bacteria or ‘polar bodies’ have been described by Newton and Horne (1957) which appear to provide the flagellate with lysine (Gill and Vogel, 1962).
Although there now appears to be general acceptance of the theory which holds that trypanosomes originated from the intestinal flagellates of insects (Hoare, 1948; Baker, 1963), the origins of the family still seem unclear. The presence in all trypanosomatid flagellates of a barren basal body in addition to that which produces the single motile flagellum (Rudzinska and Vickerman, 1969) may indicate origin from a biflagellate ancestor. Since at some stage this is likely to have been a free living flagellate, possible descent from Bodo or a Bodo-like organism is worth a brief consideration. Although the kinetoplast-mitochondrion is an obvious character uniting both groups of flagellates, it appears that an organelle comparable to the cytopharynx of Bodo has also been retained in a more or less modified form by many trypanosomatids (Brooker, 1971). In trypanosomatids, it is a deep (cytopharynx) or shallow (cytostome) invagination of the cell membrane associated, as in Bodo, with a number of microtubules. In both cases this organelle is endocytotic but whereas in Bodo it is primarily concerned with the ingestion of bacteria, in those trypanosomatids which have been examined it appears to be pinocytotic (Steinert and Novikoff, 1960; Preston, 1969; Brooker, 1971). Qualitative and quantitative differences in the nature of food ingested by Bodo and trypanosomatids may go some way to explain observed differences both in cell shape and the spatial arrangement of some organelles. Thus, the cytopharynx of Bodo, unlike that of trypanosomatids, is situated some distance from the flagellar pocket in order to facilitate prey capture and virtually the entire posterior half of the cell is devoted to the accommodation of food vacuoles. It is proposed therefore that the adoption of a parasitic mode of life by a Bodo-like flagellate and the subsequent abandonment of bacterophagic nutrition could have produced changes in body form which, together with other physiological adaptations led to the emergence of an ancestral trypanosomatid. Although such changes would have resulted in the retention of the cytopharynx, it
STRUCTURE OF BODO SALTANS AND B, CAUDATUS 101 is assumed that the loss of one flagellum and development of the relatively sparse
microtubular system occurred at some later stage.
ACKNOWLEDGEMENTS
I gratefully acknowledge the technical assistance given by Mr. C. G. Ogden who was responsible for many of the electron micrographs used in this paper.
KEY TO ABBREVIATIONS USED IN THE PLATES
ax axoneme G Golgi apparatus
b bacterium ger granular endoplasmic reticulum
bb basal body ifc interflagellar connective
be buccal cavity k kinetoplast
cb cytoplasmic bacterium m mitochondrion
cbl circumbuccal lappets mast mastigonemes
cv contractile vacuole mt microtubule
cyt cytopharynx n nucleus
f flagellum pr paraxial rod
fp flagellar pocket rt rootlet
fv food vacuole ves vesicle REFERENCES
ALEXEIEFF, A. 1912. Sur la revision du genre Bodo Ehrbg. Arch. Protistenk. 26 : 413-419.
Anverson, W. A. & Erris, R. A. 1965. Ultrastructure of Trypanosoma lewisi: flagellum, microtubules and the kinetoplast. J. Protozool. 12 : 483-499.
Baker, J. R. 1963. Speculations on the evolution of the family Trypanosomatidae Doflein, 1got. Expl. Parasit. 13 : 219-233.
Brooker, B. E. 1965. Mastigonemes in a bodonid flagellate. Expl. Cell Res. 37 : 300-305
1971. The fine structure of Crithidia fasciculata with special reference to the organelles
involved in the ingestion and digestion of protein. Z. Zellforsch. 116 : 532-563.
Git, J. W. & VocEL, H. J. 1962. Lysine synthesis and phylogeny: biochemical evidence for a bacterial-type endosymbiote in the protozoon Herpetomonas (Strigomonas) oncopelti. Biochim. biophys. Acta 56 : 200-201.
GravErtT, A.M. 1962. Fine structure of bacteria. By. med. Bull. 18 : 245-250.
Hi1t, G. C. & ANDERSon, W. A. 1969. Effects of acriflavine on the mitochondria and kineto- plast of Crithidia fasciculata. Correlation of fine structural changes with decreased mito- chondrial enzyme activity. J. Cell Biol. 41 : 547-561.
Hoare, C. A. 1948. The relationship of the haemoflagellates. In Proceedings of Fourth International Congress of Tropical Medicine. 2 : 1110-1116.
HotianpeE, A. 1936. Sur la cytologie d’un flagellé du genre Bodo. C.r. Séanc. Biol. 123:
651-653.
1942. Etude cytologique et biologique de quelques flagellés libres. Archs. Zool. exp. gén.
83 : 1-268.
Hotwitt,M.E.J. 1966. Physical aspects of flagellar movement. Physiol. Rev. 46 : 696-785.
Joyon, L. & Lom, J. 1966. Sur l'ultrastructure de Costia necatrix Leclercq (Zoofiagellé); place systématique de ce Protiste. C.v. Acad. Sci. (Paris) 262 : 660-663.
1969. Etude cytologique, systématique et pathologique d'Ichtyobodo necator (Henneguy, 1883) Pinto, 1928 (Zooflagelle). J. Protozool. 16 : 703-710.
KELLENBERGER, E. & RyTeR, A. 1958. Cell wall and cytoplasmic membrane of Escherichia coh. J. biophys. biochem. Cytol. 4 : 323-326.
102 B. E. BROOKER
Kirspy, H.W. 1941. In Protozoa in Biological Research. Eds. Calkins, G. N. and Summers, F. N. New York.
KusEL, J. P., Moore, K. E. & WEBER, M.M. 1967. The ultrastructure of Crithidia fasciculata and morphological changes induced by growth in acriflavin. J. Protozool. 14 : 283-206.
Levy, M. R. & Extiotr, A. M. 1968. Biochemical and ultrastructural changes in Tetva- hymena pyriformis during starvation. J. Protozool. 15 : 208-222.
Micnot, J. P. 1965. Etude ultrastructurale de Cyathomonas tvuncata From. (Flagellé Cryptomonadine). J. Microscopie 4 : 239-252.
Newton, B. A. & Horne, R. W. 1957. Intracellular structures in Stvigomonas oncopelti 1. Cytoplasmic structures containing ribonucleoprotein. Expl. Cell. Res. 13 : 563-574.
Piretka, D. R. 1961. Observations on the kinetoplast-mitochondrion and the cytostome of Bodo. Expl. Cell Res. 25 : 87-93.
Preston, T. M. 1969. The form and function of the cytostome-cytopharynx of the culture forms of the elasmobranch haemoflagellate Tyypanosoma vaiae Laveran and Mesnil. J. Protozool. 16 : 320-333.
Pyne, C. K. 1960. L’ultrastructure de l’appareil basal des flagelles chez Cryptobia helicis (Flagellé, Bobonidae). C.v. Acad. Sci. (Paris) 250 : 1912-1914.
Ropertson, M. 1929. The action of acriflavine upon Bodo caudatus. Parasitology 21 : 375- 410.
Ruopzinska, M. A. & VICKERMAN, K. 1969. The fine structure. In Infectious blood diseases of man and animals. New York 1 : 217-306.
ScuustErR, F. L. 1968. The gullet and trichocysts of Cyathomonas truncata. Expl. Cell Res 49 : 277-284.
Stmpson, L. 1968. Behaviour of the kinetoplast in Leishmania tarentolae on cell rupture. J. Protozool. 15 : 132-136.
Sinton, J. A. 1912. Some observations on the morphology and biology of Prowazekia uvinarvia (Bodo urinarius, Hassall).