Full text of "Bulletin of the British Museum (Natural History)"

BULLETIN OF

THE BRITISH MUSEUM

(NATURAL HISTORY)

2 2 HOV i963
A

ZOOLOGY

VOL. 10

1963

BRITISH MUSEUM (NATURAL HISTORY)

LONDON : 1963

DATES OF PUBLICATION OF THE PARTS

No. i 3 January 1963

No. 2 ...... 14 March 1963

No. 3 29 March 1963

No. 4 ...... 29 March 1963

No. 5 28 May 1963

No. 6 31 May 1963

No. 7 9 July 1963

No. 8 16 July 1963

No. 9 ...... 20 May 1963

PRINTED IN GREAT BRITAIN
BY THOMAS DE LA RUE &
COMPANY LIMITED LONDON

ZOOLOGY VOLUME 10

PAGE

No. I. Ethiopian mites of the genus Androlaelaps Berlese s. lat. (Acari :

Mesostigmata). By W. M. TILL i

No. 2. The Francolins, a study in speciation. By B. P. HALL 105

No. 3. Taxonomy and biology of the genus Lebetus (Teleostei-Gobioidea) .

By P. J. MILLER (PI. i) 205

No. 4. The freshwater gastropod molluscs of Western Aden Protectorate.

By C. A. WRIGHT (Pis. 1-2) 257

No. 5. Observations on the chaetotaxy of the legs in free-living Gamasina

(Acari : Mesostigmata). By G. OWEN EVANS 275

No. 6. A revision of the recent round herrings (Pisces : Dussumieriidae) .

By P. J. P. WHITEHEAD 305

No. 7. The polychaete fauna of South Africa. Part 8 : new species and

records from grab samples and dredgings. By J. H. DAY 381

No. 8. The freshwater gastropod mollusca of Angola. By C. A. WRIGHT

(Pis. 1-16) 447

No. 9. New marine nematodes from off the coast of South Africa. By

WILLIAM G. INGLIS 529

Index Volume 10 553

ETHIOPIAN MITES OF THE
GENUS ANDROLAELAPS

BERLESE s.lat.
(ACARI : MESOSTIGMATA)

W. M. TILL

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. i

LONDON : 1963

ETHIOPIAN MITES OF THE GENUS
ANDROLAELAPS BERLESE s. lat.
(ACARI : MESOSTIGMATA)

W. M. TILL

British Museum (Natural History)

Pp. 1-104 >' I 95 Text-Figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. i

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. i of the Zoological series.

Trustees of the British Museum 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued January, 1963 Price Thirty Jive shillings

ETHIOPIAN MITES OF THE GENUS

ANDROLAELAPS BERLESE s. lat.

(ACARI: MESOSTIGMATA).

By W. M. TILL

British Museum (Natural History], London

CONTENTS

INTRODUCTION
EXTERNAL MORPHOLOGY
CLASSIFICATION
HOST-PARASITE LIST
REFERENCES

INTRODUCTION

The earliest species of the genus Androlaelaps s. lat. (including Haemolaelaps} to be
recorded from the Ethiopian region are those described by Berlese (1911-1918)
and Hirst (1916). Radford (1939-1944) added several more species from this region
and Zumpt & Patterson (1950, 1951) provided the first keys to the Ethiopian species
of the genera Androlaelaps and Haemolaelaps. Since then many new species have
been described by Zumpt and his co-workers, by Keegan (1956) and by Lavoipierre
(1955, 1956). With the increased number of species, the existing keys have been
found to be inadequate as they were based largely on superficial characters which
are subject to intraspecific variation. It was also suspected that many of the
new species might be synonyms of older, imperfectly described species, the types
of which had never been thoroughly re-examined. The aim of the present
investigation is therefore to give redescriptions of the known species, based wherever
possible on type material, to try to evaluate the morphological characters observed
in the different species, and to prepare a key to the females. Several new species
have been included and a host-parasite list is given.

This investigation has been made possible by grants from the British Museum
(Natural History) and from the Wellcome Foundation, London. It is based chiefly
on material in the collections of the British Museum (B.M.N.H.) and the South
African Institute for Medical Research, Johannesburg (S.A.I.M.R.), the latter
collection having been placed at my disposal by Dr. F. Zumpt. Material has also
been received from Dr. P. L. G. Benoit, Musee Royal de 1'Afrique Centrale,
Tervuren, Belgium (M.R.A.C.), Dr. R. Taufflieb, Institut de Recherches Scientifiques
au Congo, Brazzaville (I.R.S.C.) and Miss J. B. Walker, East African Veterinary
Research Organization, Kenya. Types and other specimens have been lent by
Miss M. A. Johnson, Liverpool School of Tropical Medicine, Dr. M. M. J. Lavoipierre,

ZOOL. 10, I I

4 W. M. TILL

George Williams Hooper Foundation, California, Dr. C. D. Radford, Manchester,
and by the Smithsonian Institute, United States National Museum, Washington.

The work was carried out in the British Museum (Natural History) and I am
grateful to the Trustees for providing me with the necessary facilities. My thanks
are due to Dr. G. O. Evans for his guidance and many helpful suggestions, and to
Dr. J. G. Sheals and Mr. D. Macfarlane for advice and assistance in various ways.

EXTERNAL MORPHOLOGY

Females

GNATHOSOMA (figs. 1-5) : The gnathosoma is attached subterminally to the anterior
end of the body. It consists of a cylindrical basis capituli (gnathosomal base)
enclosing the pharynx and chelicerae and bearing antero-laterally a pair of seg-
mented pedipalps. The dorsal wall of the basis capituli extends anteriorly as a
membranous flap known as the tectum. In the genus Androlaelaps this is a trans-
parent structure more or less rounded anteriorly and without ornamentation.
On the ventral surface is a median longitudinal groove, the deutosternum, which
normally bears six rows of anteriorly directed teeth. The number of teeth in each
row varies from one to about six. Anteriorly the ventral wall is produced to form
the hypostome and its processes, the horn-shaped corniculi (external malae) and the
fimbriated hypostomal processes (internal malae). In most Androlaelaps species the
corniculi are prominent and well sclerotized (figs, i & 3), but in A. sangsteri and
A. heliosciuri they are reduced in size, weakly sclerotized and generally incon-
spicuous (fig. 4).

The hypostomal processes show a variety of modifications and are sometimes
very difficult to see. In most Androlaelaps species they are similar to those of
A. murinus (fig. i), the ventral processes having the form of a pair of triangular
flaps, the outer margins of which are fringed. The depth of the fringe varies in
different species and is relatively great in A. tachyoryctes (fig. 2). The hypostomal
processes are unusually large and pilose in A. georychi (fig. 5), whilst in A. walker ae
both the ventral and dorsal processes are modified to form prominent brush-like
structures (fig. 3). In A. heliosciuri the inner margins of the processes are fringed

(fig- 4)-
The ventral surface of the gnathosoma bears four pairs of setae, the anterior

rostral (g.s. i) and the posterior external and internal rostral setae (g.s. 2 and g.s. 3)
on the hypostome, and the capitular setae (g.s. 4) on the basis capituli. The lengths

FIGS. 1-5. Types of gnathosoma (ventral view) in females of the genus Androlaelaps
s. lat.

Fig. i. Androlaelaps murinus (Berlese)
Fig. 2. Androlaelaps tachyoryctes (Radford)
Fig. 3. Androlaelaps walkerae sp nov.
Fig. 4. Androlaelaps heliosciuri sp. nov.
Fig. 5. Androlaelaps georychi sp. nov.

b.c., basis capituli; deut., deutosternum; ext. m., external mala; g.s., gnathosomal
seta; hyp., hypostome; int. m., internal mala; la., labrum.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS RERLESE 5. lat.

6 W. M. TILL

of these setae are relatively constant within the genus. The anterior rostral and
capitular setae are usually subequal in length, except in the scapularis-group, and the
posterior internal rostral seta is 2-3 times as long as the external seta. Lying
dorsal to the hypostome and ventral to the chelicerae is a pointed, ciliated structure,
the labrum.

The chelicerae occupy the dorsal region of the gnathosoma. Each terminates
in a fixed and a movable digit, both of which are normally dentate and well
sclerotized. The fixed digit is usually shorter than the movable one, less strongly
dentate, and bears a transparent, seta-like structure, the pilus dentilis, the form
of which varies in different species. The structure of the chelicerae is peculiar in
the scapularis-group, in that the movable digit is edentate, except for the terminal
hook and at the most one rather blunt subterminal tooth, and bears an additional
cutting process provided with large, outwardly directed teeth. The fixed digit is
reduced, weakly sclerotized, and bears a flagellar pilus dentilis (fig. 170). In the
majority of the Androlaelaps species the pilus dentilis is inflated distal to the point
of insertion (subgenus 1 ' schnolaelaps sensu Fonseca), while in a few, for example
A . hystrici, A . phoenicidi, A . casalis, A . walkerae, it is slightly inflated near its point
of insertion, tapering progressively to a point (subgenus Atricholaelaps sensu
Fonseca) .

On the ventral surface of the chelicera, at the base of the movable digit, is a
semicircle of arthrodial filaments. In most of the Androlaelaps species these are
subequal in length or only slightly variable. In A. heliosciuri (fig. 81) two of
the filaments are considerably longer than the others, and in A. murinus
(fig. 109), and A. zumpti (fig. 195) there is one filament which is strikingly
longer and thicker than the others. The arthrodial membrane between the chelae
appears to form an extension which is usually a small, rounded, transparent flap,
but which, in A . zumpti, is greatly elongated, extending to the tips of the chelae.

The pedipalps are 5-segmented; the first free segment, or trochanter, bears
2 setae, the femur bears 5, the genu 6 and the tibia 14 setae. The tarsus, in addition
to a number of simple setae, bears a two-tined seta-like structure at its base.

IDIOSOMA: The idiosoma is usually ovoid in shape with no sharp demarcation
between the dorsal and ventral surfaces. Almost the entire dorsum is covered by a
single scutum which usually has recticular markings and in some species, e.g. those
of the scapularis-group, may also have a granular appearance. The chaetotaxy of
the dorsal shield follows the basic pattern described by Costa (1961, fig. i) for the
genus Haemolaelaps. Most of the species have the pattern comprising 39 pairs of
setae, although in some species, for example those of the mesopicos-group, one or
more of the marginal pairs may be situated on the adjacent integument and seta
PX$ may be lost. A. longipes has an additional pair of setae (pxi) on the posterior
part of the plate (fig. 93), whilst A. centrocarpus and A. hirsti have extra pairs of
setae (ax and PXI} on both the anterior and posterior parts of the shield. Many
species show some degree of hypertrichy, a variable number of unpaired, accessory
setae developing on the median part of the plate between the "/" series. In
A . later ae and A . villosissimus the dorsal shield has a dense covering of setae which
completely obscures the basic pattern, An intermediate condition occurs in

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lot. 7

A . walker ae in which the basic pattern is partially obscured by the addition of both
unpaired and paired setae. The setae vary considerably in length and in the
scapularis- and w^so^'cos-groups (excluding A. wilkini] and in A.galagus those on
the central part of the shield are reduced to microsetae.

The ventral surface of the idiosoma bears a sternal, a genital and an anal shield,
all of which may show some degree of intraspecific variation. The sternal shield is
usually broader than long, reticulate in most species but with a granular appearance
in mites of the scapularis-group, and bears two pairs of pores and three pairs of
setae. In most Androlaelaps species the sternal setae are more or less subequal in
length, the second and third pairs being very slightly longer than the first pair. In
A. arvicanthis, A. graingeri, A. marshalli and A. theseus the first pair is relatively
short, being |-| as long as the second pair. Metasternal setae are always present.
The presternal area also has a reticulate pattern and bears the tritosternum, a
bipartite structure of uniform appearance throughout the genus. It consists of a
short, undivided, basal portion and two ciliated arms, the laciniae.

The genital shield consists of a relatively more strongly sclerotized posterior
portion, varying in shape from the slender tapering form seen in A . capensis (fig. 20)
and A. tateronis (fig. 165) to the greatly expanded, flask-shaped form of A.
rhodesiensis (fig. 128), and a more weakly sclerotized anterior portion which forms a
flap covering the genital orifice. The shield usually has a linear or reticulate
pattern at least on the posterior portion. This pattern may be very weakly
developed, as in the scapularis-group, or it may be very clearly defined and extend
on to the anterior portion, as in A. dasymys (fig. 51). In some mites the shield
also has a granular appearance and in all the species it bears only one pair of genital
setae.

The anal shield is ovate or triangular, with the apex directed posteriorly, and
encloses the anal orifice. It bears a pair of paranal setae and a single postanal seta.
The metapodal plates are small sclerotized structures lying posterior to the fourth
pair of coxae. The peritreme extends from the stigma, between coxae III and IV,
to a point over coxa I, and the peritremal plate extends posterior to the stigma for
a distance equal to approximately twice the diameter of the stigma.

The setae arising from the integument of the idiosoma may be simple or barbed
and are variable in number, some species being very sparsely and others relatively
densely haired, as for instance A . taterae and A . georychi. The marginal setae are
defined as those lying between the dorsal shield and an arbitrary line commencing
at the posterior tip of the anal shield and continuing anteriorly on both sides and
equidistant from the lateral margin of the body.

The names of the idiosomal shields have been abbreviated and their measure-
ments taken as follows:

DL = median length of dorsal shield

DW = greatest width of dorsal shield

SL = median length of sternal shield

SW = width of sternal shield at level of second pair of setae

GW X = greatest width of genital shield

GW 2 = width of genital shield at level of genital setae

8 W. M. TILL

AL = median length of anal shield
AW = greatest width of anal shield.

LEGS : The legs consist of six recognizable segments, namely, the coxa, trochanter,
femur, genu, tibia and tarsus, each tarsus terminating in a caruncle with claws.
The segments are usually smooth in outline, but in A. callosus some of them have
prominent, heavily chitinized swellings (fig. 18). Leg II is usually stouter than the
other legs and this feature is most pronounced in the Androlaelaps-group. The
chaetotaxy of the legs follows a basic pattern which is subject to slight variations.
The number of setae per segment found in the majority of the Androlaelaps species,
and which I have taken to be the normal number for this genus, is given in the
following table :

Leg Coxa Trochanter Femur Genu Tibia Tarsus

I 2 6 13 13 13

II ... 2 5 ii ii 10 16

III ... 2 5 6 9 8 16

IV ... i 5 6 10 10 16

The number of tarsal setae listed does not include the two fine setae at the tip of
the segment.

Variations have been noted in the number of setae on the first and third tibiae
and on the third and fourth genua. Tibia I has 14 setae in A. rhodesiensis and
A. walker ae; tibia III has 9 setae in these two species and also in the species of
the mesopicos-group, in A. centrocarpus , A. hirsti, A. longipes, A. phoeniculi, A. oliffi,
A. tateronis, A. cricetomydis, A. congoensis, A. taterae and A. villosissimus. The last
three species have ten setae on genu III and A. spatuliformis has nine setae on
genu IV. The setae are variable in form and may be smooth or barbed, hair-like,
spine-like, pointed or blunt. In the mites of the Androlaelaps-group one of the
ventral setae on femur II is modified to form a stout spur-like process.

Males

The gnathosoma of the male is similar in structure to that of the female. The
chelicera, however, has its movable digit modified to form a spermadactyl which
serves as a copulatory organ. This may be relatively short and straight, as in
A . casalis (fig. 29) , A . hystrici (fig. 88) and others, or relatively long and coiled as in
A. taterae (fig. 164), A. patersoni (fig. 124) and A. spreo (fig. 145). The fixed digit
bears a pilus dentilis, sometimes very difficult to detect, which resembles the pilus
dentilis of the corresponding female.

The dorsal shield has the same chaetotactic pattern as in the female, although the
setae are often relatively longer.

The ventral shields in most Androlaelaps species are fused to form a single
holoventral plate which shows varying degrees of expansion behind the fourth pair
of legs. The shield bears four pairs of sternal, one pair of genital, and the usual

FIGS. 6-9. Androlaelaps casalis (Berlese). Venter (Fig. 6) and dorsum (Fig. 7) of
deutonymph; dorsum (Fig. 8) and venter (Fig. 9) of protonymph.

ETHIOPIAN MITES OF THE GENUS A NDROLA ELA PS BERLESE s. lat.

TO W. M. TILL

three anal setae. The number of setae in the preanal region varies between two
or three pairs, in the species with a narrow plate, for example the mesopicos-group,
and many, in those species with an expanded plate, for instance A. murinus and
A . later ae. The usual number is five pairs.

The anal shield is separated from the sternito-ventral shield in A. centrocarpus ,
A . hirsti and A . longipes (species with 40 or 41 pairs of setae on the dorsal shield)
and also in A, tauffliebi. In all species the genital orifice is presternal.

The chaetotaxy of the legs follows the same pattern as in the corresponding
females although certain setae, for example in the mesopicos-group, may be modified
to form thickened spines or spurs.

Immature Stages

According to Bregetova (1956) Haemolaelaps females are able to lay eggs with
developed larvae or they may give birth to larvae or even to protonymphs. As
far as the Ethiopian species are concerned only proto- and deutonymphs have been
recorded up to now.

The following description of the nymphal stages is based on specimens of A . casalis
from deep poultry litter in England. The immature stages of the other species will
not be described as so very little material is available.

PROTONYMPH (figs. 8-9) : The gnathosoma resembles that of the female except that
the palps have not yet acquired their full complement of setae. The trochanter bears
one ventral seta, and the femur, genu and tibia bear 4, 5 and 12 setae respectively.

The dorsum has a large anterior podonotal shield bearing eleven pairs of setae,
a smaller, posterior, pygidial shield bearing eight pairs of setae, and three pairs of
small intermediate platelets or mesonotal scutellae. The sternal shield, which
bears three pairs of setae and two pairs of pores, does not extend beyond the middle
of coxae IV. The peritremes are short and do not reach beyond coxae III. There
is no genital shield or genital orifice.

The legs have not yet acquired their full number of setae, those present being
listed in the following table:

Leg Coxa Trochanter Femur Genu Tibia Tarsus

I ... 2 4 10 8 8

II ... 2 4 8 6 7 15

III ... 2 4 5 6 7 15

IV . i 4 4 6 7 15

DEUTONYMPH (figs. 6-7) : The deutonymph is larger than the protonymph and the
palps and legs have the same number of setae as in the adult stages. The podonotal,
mesonotal and pygidial shields have fused to form a single dorsal scutum which
has a lateral incision on either side, behind seta 57. The number of setae on the
shield is the same as in the adult. On the ventral surface the sternal shield extends
to the level of the posterior border of coxa IV. It bears four pairs of setae (the
sternals and metasternals) and three pairs of pores. The peritremes are long,
reaching coxae I. Metapodal plates are present, and also several small areas of
sclerotization on the median part of the venter,

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. n

CLASSIFICATION

It has long been recognized that the genera Androlaelaps and Haemolaelaps are
very closely related to one another as well as to the genus Hypoaspis Canestrini.
Both Androlaelaps and Haemolaelaps were originally described by Berlese as sub-
genera of Laelaps Koch. Subsequently Haemolaelaps (Berlese, 1916) and then also
Androlaelaps (Vitzthum, 1943; Zumpt, 1950; Zumpt & Patterson, 1950) were
placed as subgenera of Hypoaspis, and more recently both have been given generic
status (Strandtmann, 1949; Zumpt & Patterson, 1951; Keegan, 1956; and others).

There is still some confusion, however, regarding the delimitation of the genus
Haemolaelaps. The genera Atricholaelaps Ewing and Ischnolaelaps Fonseca were
synonymized with Haemolaelaps by Strandtmann (1949), a decision which has
been accepted by most other workers on this group. Fonseca (1959), however,
recommended that Haemolaelaps be kept as a monotypic genus for H. marsupialis
Berlese, the remaining species being allocated to the genus Atricholaelaps with its
subgenus Ischnolaelaps. He based these divisions on the form of the pilus dentilis,
a feature which he himself admits is sometimes difficult to interpret. Eubrachylae-
laps Ewing (= Cyclolaelaps Ewing) and Zygolaelaps Tipton were added to the
synonymy by Zumpt and his co-workers (1951, 1958). Cavilaelaps Fonseca was
included in the synonymy by Zumpt & Patterson (1951), but was not listed by
Zumpt & Till (1961). Its status is still doubtful as I have not been able to examine
the type species. The genus Turkiella Zumpt & Till was synonymized with
Androlaelaps by Keegan (1956).

The genera Androlaelaps and Haemolaelaps resemble one another very closely in
their dorsal and ventral sclerotization and chaetotaxy, in their leg chaetotaxy and
in the structure of the chelicerae. In both genera the sternal shield of the female
is broader than long, the pilus dentilis is inflated or long and slender, and the digits
of the male chelicerae are weakly sclerotized and edentate, the fixed digit being
shorter than the movable one. The only differences between Androlaelaps and
Haemolaelaps are that in females of the former genus leg II is relatively more stout
and one of the ventral setae on femur II is thickened to form a spur-like structure.
These features, however, are not clear-cut. There is a gradation both in the thickness
of the second leg and in the nature of its setae from the Haemolaelaps to the
Androlaelaps species. Furthermore, males of the mesopicos-group of Haemolaelaps
species have a stout spur-like seta in the position corresponding to the Androlaelaps
spur, whilst the females have only a simple seta.

The genus Hypoaspis differs from Androlaelaps and Haemolaelaps in that the
length of the sternal shield of the female usually equals or exceeds the breadth ; the
pilus dentilis is short, slender, scarcely visible, and the digits of the male chelicera
are well sclerotized, equal in length, with dentate internal margins. The genus
Gaeolaelaps 1 Tragardh, with Laelaps aculeifer Canestrini as type species, appears to
bear the same relationship to Hypoaspis as Androlaelaps does to Haemolaelaps.

It seems possible that these mites may all have arisen from a common free-living

1 The genus Gaeolaelaps was mentioned in Berlese's note book but was never described. Berlese
referred to the genus Geolaelaps in 1924 (Redia 15 : 254), but without designating a type. Tragardh
(1952, Ark, Zool. 4 ser, 2 : 66) states that Berlese (in MS) had designated L, aculeifer as the type.

12 W. M. TILL

Hypoaspis-\ike ancestor, giving rise on the one hand to species with simple setae
on leg II (Hypoaspis} and on the other to species having an enlarged second leg with
thickened setae (Gaeolaelaps} . Modified, elongate, edentate male chelae could
have developed in the mites of both branches which became adapted to a parasitic
mode of life, giving rise to species of the Haemolaelaps and Androlaelaps types
respectively.

In deciding upon the status of these four genera there seemed to be three
possibilities :

1. To give all four generic status.

2. On the basis of leg setae to combine Haemolaelaps with Hypoaspis and
Gaeolaelaps with Androlaelaps.

3. On the basis of male chelae, female sternal plate and pilus dentilis, to
combine Haemolaelaps with Androlaelaps and Gaeolaelaps with Hypoaspis.

Of these three propositions the third seems to be the most appropriate and
Haemolaelaps has consequently been synonymized with Androlaelaps. In this
connection it is interesting to note that two species described recently (H. travisi
Delfinado 1961 and H. vietnamensis Grochovskaya and Nguyen-Xuan-Hoe 1961)
as having the Androlaelaps type of leg setation have been allocated to the genus
Haemolaelaps by their respective authors.

Bregetova (1956), in her diagnosis of the genus Androlaelaps, describes the
chelicerae as being similar in structure in both sexes, the male having a spermadactyl
added parallel to the movable digit. This is the type of chelicera found in males
of the genus Hypoaspis. However, Berlese's (1887) drawing of A. hermaphrodita,
the type species of Androlaelaps, shows that the male chelae are elongate and
edentate. The Androlaelaps species described from the Ethiopian region are clearly
co-generic with A. hermaphrodita, whereas those listed by Bregetova (.4. sardous
Berlese, A. karawaiewi Berlese and A. pavlovskii Bregetova) are of the Hypoaspis
type.

In the key which I have given to the females of the genus Androlaelaps s. lot.,
couplets 2 to 5 refer to the species bearing the modified spur-like seta (Androlaelaps-
group), whilst the remaining couplets refer to species previously assigned to the
genus Haemolaelaps (Haemolaelaps-group).

On the basis of morphological features, many of the species which I have included
in the Haemolaelaps-group could be arranged in well-defined sub-groups. A.
cryptomius, A. scapularis and A. tauffliebi, three species parasitic on mole-rats,
form a distinct group having the movable cheliceral digit in two parts, microsetae
on the dorsal shield, and " granular " dorsal and ventral idiosomal shields. The
status of these mites may have to be reconsidered and the group regarded either as a
subgenus of Androlaelaps or even as a distinct genus.

A. taterae and A. villosissimus differ from all other Androlaelaps species in their
dorsal chaetotaxy. A. centrocarpus, A. hirsti and A. longipes form a fairly well-
defined group having 40-41 pairs of dorsal setae, nine setae on tibia III and a separate
anal shield in the male.

The mesopicos-group, comprising four species associated with birds, namely,

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 13

A. haydocki, A. mesopicos, A. steyni and A. wilkini, is characterized by the sinuous
posterior setae, the position of the first pair of sternal setae on the presternal area
and by the presence of nine setae on tibia III.

The zulu-glasgowi-group comprises a number of apparently very closely allied
species (A. dasymys s. lat., A. glasgowi, A. pachyptilae, A. patersoni, A. spreo, A.
suncus and A. zulu) whose status is not quite clear, and it is possible that some of
them are no more than subspecies or varieties.

A. glasgowi represents a cosmopolitan species complex which has been recorded
from a wide variety of mammalian and bird hosts (Strandtmann, 1949 ; Bregetova,
1956). According to Bregetova, this species has only 38 pairs of setae on the dorsal
shield, setae 53 being absent. This has been confirmed in specimens from Japan
and Astrakhan in the collection of the British Museum (Natural History) . American
specimens which I have been able to examine, namely, the type specimen of A.
glasgowi (lent by the Smithsonian Institute, Washington) and several specimens
from the collection of the South African Institute for Medical Research, have 39 pairs
of setae. Mites from Israel (Costa, 1961) and from Southern Africa also have
39 pairs of dorsal setae as well as a number of asymmetrically arranged setae on the
posterior part of the dorsal shield. There are also slight morphological differences,
chiefly in relative measurements, between mites from different host species
(Strandtmann, 1949).

A. zulu very closely resembles A. glasgowi, differing from it in having a larger
genital shield and longer setae on the dorsal shield. In both species the pilus
dentilis is strongly inflated proximally, the terminal portion being slender and
recurved. Specimens previously identified by Zumpt & Till (1956) as forms of
A. glasgowi in which the genital shield is bordered by three pairs of setae would
appear to belong to A . zulu.

The remaining members of this group differ from A. glasgowi and A. zulu, and
from one another, in the form of the pilus dentilis, the relative lengths of the
arthrodial filaments at the base of the movable digit of the chelicera and the dorsal
setae, and the relative measurements of parts of the body.

A. dasymys s. lat. is considered to include A. davisi and A. labuschagnei which
appear to differ from A. dasymys s. sir. only in size, degree of sclerotization and
relative setal lengths. A. dasymys was based on a single damaged specimen which
lacks most of the dorsal setae and of which the pilus dentilis is not clearly visible.
By courtesy of Dr. P. L. G. Benoit (Musee Royal de 1'Afrique Centrale) three mites
from Rhabdomys pumilio have been examined which appear to be identical with
A. dasymys and which are in rather better condition. These mites are relatively
large and robust, the dorsal shield measuring 1025-1064^ by 700-750^.

The type series of A . davisi consists of rather smaller, less heavily sclerotized
mites, in which the dorsal shield measures 822-888{jL by 516-569^. Examination
of further material from the type host of A . davisi shows the dorsal chaetotaxy of
this species to be variable in that accessory setae may be present or absent. Mites
from different host species show a variation in the relative lengths of the dorsal
setae. The type specimens of A . labuschagnei, which unfortunately are not in very
good condition, appear to be intermediate between A . davisi and A . dasymys s. str.

I 4 W. M. TILL

A. dasymys s. lat. has been collected from a variety of hosts, but appears to be
restricted to the Muridae.

The remaining species in the genus Androlaelaps are more difficult to arrange in
groups. Whatever feature is used for the purpose of separation, the resulting
groups are heterogeneous with respect to other features.

Family LAELAPTIDAE
Genus Androlaelaps Berlese

Androlaelaps Berlese, 1903, Zool. Anz. 27 : 14; Zumpt, 1950, Parasitology 40 : 299; Zumpt &

Patterson, /. ent. Soc. S. Afr. 1951, 14 : 69.

Haemolaelaps Berlese, 1910, Redia 6 : 216 & 1916, Redia 12 : 170; Strandtmann, ig^g.J.Parasit.

35 : 3 2 5I Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 70; Bregetova, 1956, Acad. Sci.,

U.S.S.R. 61 : 84; Fonseca, 1959, Mem. Inst. ButantanS. Paulo (1957/58) 28 : 46 (syn. nov.).

Atricholaelaps Ewing, 1929, Manual of External Parasites : 186.

Eubrachylaelaps Ewing, 1929, Manual of External Parasites : 186.

1 schnolaelaps Fonseca, 1936, Mem. Inst. Butantan S. Paulo 10 : 19.

Cyclolaelaps Ewing, 1933, Proc. U.S. Nat. Mus. 82 : 5.

Turkiella Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 215.

Zygolaelaps Tipton, 1957, / Parasit. 43 : 367.

Type species: Laelaps (Iphis) hermaphrodita Berlese, 1887

KEY TO THE FEMALES

1. Femur II with a stout, blunt or thorn-like seta on its ventral surface (Androlaelaps-
group) . . 2

- Femur II with only simple setae on its ventral surface (Haemolaelaps-group) . . 6

2. Dorsal shield with seta ax present. Genital shield small, slightly widened behind genital
setae, then tapering posteriorly (figs. 173-174) .... A. theseus Zumpt

- Dorsal shield with seta ax absent. Genital shield large, flask-shaped, broadly rounded
posteriorly ............. 3

3. Dorsal shield with setae px2 and px$ present ....... 4

- Dorsal shield with setae px2 and px$ absent ....... 5

4. Tarsus II terminating in pointed setae only. Genital shield very closely approached to
anal shield. Genital setae relatively short, less than half GW 2 (fig. 188).

A. zuluensis Zumpt

Tarsus II with three stout, blunt, terminal spines. Genital shield widely separated from
anal shield. Genital setae relatively long, approximately equal to GW 2 (fig. 98).

A. marshalli Berlese

5. Integument of idiosoma with about 21 pairs of relatively short, fine setae (figs. 70-71).

A. graingeri Zumpt & Patterson

- Integument of idiosoma with about 37 pairs of long, coarse setae (fig. 10).

A. arvicanthis Radford

6. Hypostomal processes forming two pairs of prominent brushes, one ventral and one dorsal
(fig. 3) . . . . . . . . . . .A. walkerae sp. nov.

Hypostomal processes fimbriated or pilose, but not forming brushes ... 7

7. Dorsal shield with an overall covering of setae which completely obscures the normal
chaetotactic pattern (fig. 161) . . . . . .A. villosissimus (Berlese)

A. taterae (Zpt. & Patt.)

Dorsal shield with chaetotactic pattern composed of not more than 41 pairs of setae,
excluding the accessory setae on the postero-median part of the shield ... 8

8. Corniculi weakly sclerotized (fig. 4) or reduced, inconspicuous .... 9

- Corniculi well sclerotized, normally developed, conspicuous (figs. 1-3, 5) . . 10

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lai. 15

g. First pair of sternal setae inserted on anterior margin of sternal shield. Distance between
genital and anal shields more than twice distance from anterior margin of anal shield to
anterior margin of anus. Dorsal shield almost parallel-sided (figs. 131-132).

A. sangsteri (Radford)

First pair of sternal setae inserted on presternal area. Distance between genital and anal
shields about half the distance from anterior margin of anal shield to anterior margin of
anus. Dorsal shield ovoid (figs. 79-80) .... A. heliosciuri sp. nov.

10. Dorsal shield with seta /4 at least 5 times as long as seta J r (fig. 55).

A. galagus (Lavoipierre)

- Dorsal shield with setae /i and /4 subequal . . . . . . . n

11. Setae Z$ whip-like and sinuous (fig. 102). From birds (mesopicos-group) . . 12

- Setae Z$ normal. From mammals and birds . . . . . . . 15

12. Setae 24 and 22 very minute, their lengths about J the distance between their bases, or less.
Sternal shield granular ........... 13

- Setae 14 and 22 longer, their lengths at least the distance between their bases. Sternal
shield reticulate (fig. 182-183) ....... A. wilkini (Till)

13. Anterior seta of coxa III thick, spine-like (fig. 148) ... A. steyni (Till)

- Anterior seta of coxa III normal, hair-like . . . . . . . . 14

14. Marginal setae of posterior half of dorsal shield simple (fig. 102) A mesopicos (Radford)
Two pairs of marginal setae (82 and 53) of posterior half of dorsal shield whip-like, sinuous
(fig. 78) A. haydocki (Till)

15. Some leg segments with pronounced chitinized swellings (fig. 18) . A. callosus (Berlese)

Legs without swellings ........... iG

16. Dorsal shield with seta px i developed (fig. 83) ....... 17

Dorsal shield with seta pxi absent ......... 19

17. Seta ax present (figs. 31, 83) .......... 18

- Seta ax absent (fig. 93) ....... A. longipes (Bregetova)

1 8. Posterior dorsal shield without or with a few accessory setae restricted to the region /3 to
/5 (fig. 83) . .... .A. hirsti (Keegan)

Posterior dorsal shield with numerous accessory setae extending anterior to seta/i (fig. 31).

A. centrocarpus (Berlese)

19. Tibia III with 9 setae ........... 20

Tibia III with 8 setae ........... 24

20. Genital shield greatly expanded, greatest width at least 1-6 times width at level of genital
setae .............. 21

- Genital shield not, or only slightly expanded, greatest width not exceeding 1-3 times width
at level of genital setae . . . . . . . . . . .23

21. Genu III with 10 setae ....... A. congoensis sp. nov.

Genu III with 9 setae ........... 22

22. Lengths of setae 24 and z-z at least twice the distance between their bases (fig>. 129). Tibia 1
with 14 setae ........ A. rhodesiensis (Zpt. & Patt.)

Lengths of setae 44 and z-z less than half the distance between their bases (fig. 43). Tibia I
with 13 setae ........ A. cricetomydis sp. nov.

23. Tarsus II terminating in pointed setae only. Lengths of setae 24 and 22 approximately J,
or less, the distance between their bases. Dorsal shield with 37 pairs of setae (fig. 126).

A. phoeniculi (Zumpt & Till)

Tarsus II with at least one of the terminal setae stout and blunt. Lengths of setae ^4 and
Z2. approximately equal to the distance between their bases. Dorsal shield with 39 pairs
of setae (figs. 115, 166) ....... A. tateronis (Radford)

A. oliffl (Zumpt & Patterson)

24. Tarsus II with 3 stout, blunt terminal spines (fig. 157) ..... 25
Tarsus II terminating in pointed setae only ..... 26

25. Sternal shield granular. Caruncle of tarsus IV normal, its length being i times the greatest
width of the tarsus . A. tachyoryctes (Radford)

iC W. M. TILL

- Sternal shield faintly reticulated. Caruncle of tarsus IV elongated, its length being at
least twice the greatest width of the tarsus. .... A. capensis (Hirst)

26. Movable digit of chelicera in two parts, one part being dentate; fixed digit of chelicera
weakly sclerotized, membranous in appearance (figs. 135, 170). Setae 24 and 22 very
minute, their lengths not exceeding J the distance between their bases (fig. 134). (scapularis-
group) . 27

- Movable digit of chelicera normal; fixed digit normally developed, well sclerotized.
Setae 14 and z^ relatively longer, their lengths nearly half the distance between their bases,
or longer ............. 29

27. Genital shield short, widely separated from anal shield (fig. 45) A. cryptomius (Radford)

- Genital shield large, flask-shaped, extending almost to anal shield .... 28

28. Width of sternal shield at most 2 \ times its length. Arthrodial filaments at base of movable
digit of chelicera subequal in length (figs. 133, 135) . . A. scapularis (Berlese)

- Width of sternal shield 3 times its length. Four or 5 of the arthrodial filaments at the base
of the movable digit of the chelicera conspicuously longer and stouter than the others
(figs. 168, 170) A. tauffliebi sp. nov.

29. Genu IV with 9 setae A. spatuliformis (Lavoipierre)

Genu IV with 10 setae ........... 30

30. Integument of venter, behind coxae IV, relatively densely haired, about 15 pairs, or more,
of truly vental setae (figs. 59, 107) ......... 31

Integument of venter relatively sparsely haired, about 10 pairs, or less, of truly ventral
setae .............. 32

31. Chelicera with arthrodial filaments subequal in length. Pilus dentilis moderately and
more or less evenly inflated, elongate (fig. 61). Hypostomal processes very large and hairy
(fig. 5) . ..... A. georychi sp. nov.

- Chelicera with one very long arthrodial filament. Pilus dentilis strongly inflated basally,
terminal portion slender, curved (fig. 109). Hypostomal processes normal (fig. i).

A. murinus (Berlese)

32. Setae on dorsal shield very long and thick; seta Ji reaches almost to base of 74; lengths
of setae 24 and 22 at least twice the distance between their bases (fig. 63).

A. ghanensis sp. nov.

Setae on dorsal shield relatively shorter and more slender; seta /i does not reach base of
/3; lengths of setae 24 and 22 less than if the distance between their bases . . 33

33. Pilus dentilis slender (figs. 27, 91) ......... 34

Pilus dentilis inflated (figs. 187, 195) ......... 35

34. Genital shield widest anterior to first pair of flanking setae, then tapering posteriorly.
Distance between genital and anal shields more than twice distance from anterior margin
of anal shield to anus (fig. 89) A. hystrici (Zumpt & Till)

Genital shield widest at level of second pair of flanking setae, broadly rounded posteriorly.
Distance between genital and anal shields not exceeding distance from anterior margin of
anal shield to anus (fig. 25) A. casalis (Berlese)

35. Anal shield i times longer than broad (fig. 193). . . A. zutnpti sp. nov.

- Anal shield about as long as broad .... zulu-glasgowi complex

a. Parasites of birds ........... b

Parasites of mammals .......... d

b. Tarsus IV with length to greatest width not greater than 5:1. Seta /i extends to a
point approximately level with base of seta /2 (fig. 118).

A. pachyptilae (Zumpt & Till)

Tarsus IV with length to greatest width about 7:1. Seta /i extends well beyond
base of /2, at least half way to base of /3 . . . . . . . c

c. Pilus dentilis spatulate, terminating in a short, slightly bent, sharp point (fig. 122)

A. patersoni (Zumpt & Till)

Pilus dentilis with a larger proximal and a smaller distal inflation, the latter terminating
in a very fine point (fig. 143) .... A. spreo (Zumpt & Till)

ETHIOPIAN MITES OF THE GENUS ANDBOLAELAPS BERLESE s. la/. 17

d. Chelae 50-65^ long; 2 or 3 of the arthrodial filaments at the base of the movable
digit distinctly longer than the others. Pilus dentilis as in fig. 53 A. dasymys

(Radford)

- Chelae 30-40(0. long; arthrodial filaments at base of movable digit usually subequal in
length e

e. Pilus dentilis with distal portion inflated (fig. 153). Tarsus IV with length to greatest
width not more than 5:1 . . . . . A. suncus sp. nov.
Pilus dentilis with distal portion slender, usually curved (figs. 67, 187). Tarsus IV
with length to greatest width at least 6:1. . . . . . . f

FIGS. lo-ii. Androlaelaps arvicanthis Radford, female. Venter (Fig. 10) ; dorsum (Fig. n).

/. Setae of dorsal shield relatively short, lengths of 24 and Z2 at most slightly exceed the
distance between their bases. Genital shield bordered by two pairs of setae (figs. 65-
66) A. glasgowi (Ewing)

- Setae of dorsal shield relatively long, lengths of 34 and zz at least i times the distance
between their bases. Genital shield bordered by three pairs of setae (figs. 185-186).

A. zulu (Berlesc)

Androlaelaps arvicanthis Radford

Androlaelaps arvicanthis Radford, 1944, Parasitology 35 : 162, fig. 4.

Turkiella arvicanthis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 225.

Androlaelaps equatoriensis Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 230, fig. 15. (synnov.).

ZOOL. IO, I

i8 W. M. TILL

FEMALE (figs. 10-13) : Deutosternum with 6 rows of teeth, 4 to 5 in each row.
Chelae 84(0, long, both digits dentate; pilus dentilis slender, arthrodial filaments at
base of movable digit subequal in length.

Dorsal shield relatively broad, with an overall reticular pattern, well sculptured,
especially anteriorly. DL 1235-1 264^; DW= 950-1045^; DL/DW= 1-2-1-3.
The shield bears only 37 pairs of setae, px2 and px% being absent. Seta ri distinctly
longer than seta si ; lengths of setae 24 and Z2 about i| times the distance between
their bases; seta Z$ about ij times as long as seta 75.

Sternal shield with a distinct reticular pattern as well as a fine granulation,
anterior and posterior margins very slightly concave. SL= 114-1 24|A; SW=257(x;
SL/SW=o-4-o-5. First pair of sternal setae relatively short, not reaching posterior
margin of sternal shield ; second and third pairs and metasternal setae considerably
longer.

Genital shield flask-shaped, with a reticular and granular pattern. GW X =
257-284^; GW 2 2igy.; GW 1 /GW 2 =i-2-i-3. Genital setae relatively long,
approximately equal to GW 2 , and extending to bases of second pair of flanking setae.

Anal shield with linear and granular markings, anterior margin straight or convex,
anus situated at less than its length from the anterior margin. AL=i8o(j,; AW=
200-205^; AL/AW=o-9. Paranal setae level with posterior half of anus, about
4/5 as long as postanal seta.

Integument of idiosoma with about 37 pairs of long ventral and marginal setae,
many of which are barbed. Metapodal plates broadly oval, length to width about

4 =3-
Chaetotaxy of legs normal, anterior and posterior setae on coxa I subequal in

length. Inner ventral seta on tibia II slightly shorter and stouter than outer seta;
inner ventral seta on genu II slightly shorter and stouter than outer seta, but
longer than in A. marshalli. Tarsus II bears three stout, blunt, terminal setae.
Length of tarsus IV about 7 times its width at the base.

MALE (figs. 14-15): Deutosternum as in female; chelicera as illustrated.
Chaetotaxy of dorsal shield as in female. DL= 1064-1092^; DW= 779-884^1;
DL/DW=i-2-i-4. Holoventral shield reticulate, expanded behind coxae IV,
bearing 28-29 setae, 15-16 being situated in the preanal region. Chaetotaxy of
legs as in female, but inner ventral seta on tibia II stouter, resembling the condition
in A. marshalli females.

HOSTS AND LOCALITIES. Arvicanthis niloticus (Desmarest), Bunyoro, Uganda
(types in B.M.N.H.).

Tatera valida (Bocage), Bunyoro, Uganda (S.A.I.M.R.).

Taiera species, Kerio Valley, Kenya and W. Nile district, Uganda (S.A.I.M.R.).
Rattus natalensis (Smith), an Elephant Shrew (single record) and an unidentified
rodent, Kerio Valley, Kenya (S.A.I.M.R.).
Unidentified rodent, Yei, Sudan (paratype of A. equatoriensis Keegan in B.M.N.H.).

Androlaelaps callosus (Berlese)

Hypoaspis (Haemolaelaps) callosits Berlese, 1916, Redia 12 : 171.

Cyclolaelaps lophuromius Radford, 1939, Parasitology 31 : 243, fig. i (syn. nov.}.

ETHIOPIAN MITES OF THE GENUS AN DRO LA EL A PS BERLESE s. la/. 19

Hypoaspis (Haemolaelaps) lophuromius, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 70.
Haemolaelaps lophuromius, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa, 10 : 238.
Haemolaelaps sudanicus Zumpt & Till, 1954, ^ ev - Ecuat. Ent. Parasit. 2 : 213, fig. 3; Kee^an,
1956, /. Egypt, publ. tilth. Ass. 31 : 237, fig. 22 (syn. nov.).

FEMALE (figs. 16-19): Deutosternum with 6 rows of 2 to 4 fine teeth. Chelae
40-46^1 long; pilus dentilis with basal 4/5 inflated, apex sharply recurved according

FIGS. 12-15. Androlaelaps arvicanthis Radford. Chelicera (Fig. 12) and ventral view of
femur, genu and tibia of leg II (Fig. 13) of female. Chelicera (Fig. 14) and venter
(Fig. 15) of male.

to Keegan (1956) but indistinct in the specimens examined; one of the arthrodial
filaments at the base of the movable digit longer than the others.

Dorsal shield reticulate, relatively broad. DL=6oo-620(j.; DW= 445-486^;
DL/DW=i'3-i'4. It bears 39 pairs of setae and one accessory seta at the level of
/4. Anterior and marginal setae longer and stouter than those on central part
of shield and provided with fine barbs. Lengths of setae 14 and Z2 approximately
half the distance between their bases; seta Z$ about i| times as long as seta /5_

W. M. TILL

Sternal shield with a reticular pattern which is most pronounced antero-laterally
and indistinct on the central part of the shield. SL= 83-93^; SW= 134-145^;
SL/SW=o-6-o-7. Sternal setae stout, first pair the same length as the metasternal
setae, second and third pairs slightly longer.

Genital shield granular and also with a reticulate pattern which is well marked

FIGS. 16-19. Androlaelaps callosus (Berlese), female. Venter (Fig. 16); dorsum
(Fig. 17) ; femur, genu, tibia and tarsus of leg IV (Fig. 18) ; chelicera (Fig. 19).

posteriorly but less distinct anteriorly. GW X = 130^1; GW 2 = 118-127^. Length of
genital setae less than half GW 2 .

Anal shield approximately as wide as long, the length and width varying between
103^ and ii4(x. Anus situated at approximately its length from the anterior
margin of the shield; paranal setae level with middle of anus, slightly shorter than
postanal seta.

Integument of idiosoma bears 25 pairs of ventral and marginal setae, the latter
being barbed. Metapodal plates ovoid, their length to width about 2:1.

Chaetotaxy of legs normal ; anterior seta on coxa I about 4/5 as long as posterior
seta. Length of tarsus IV 4! times its width at the base. This species is charac-

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. zi

terized by the strongly chitinized swellings on several of the leg segments, namely,
femur, tibia and tarsus of leg I, femur II, tarsus III, tibia and tarsus of leg IV.

MALE: Not known.

HOSTS AND LOCALITIES: Lophuromys flavopimctatus Thomas (=Lophuromys zend)
from Kenya (Berlese, 1916; Keegan, 1956), from Kampala, Uganda (syntype of
C. lophuromius in B.M.N.H.), and from Torit, Sudan (type series of H. sudanicus in
S.A.I.M.R.).

Unidentified rodent from Uganda (B.M.N.H.).

Keegan (1956) records this species from Tachyoryctes sp. and a mixed collection of
Otomys and Arvicanthis species in the Rift Valley Province, Kenya.

Androlaelaps capensis (Hirst)

Haemolaelaps capensis Hirst, 1916, /. zool. Res. i : 79, figs, 12-14.

Hypoaspis (Haemolaelaps} capensis, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 72.
Haemolaelaps capensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa, 10 : 246.
Haemolaelaps spinitarsus Berlese, 1918, Redia 13 : 123; Keegan, 1956, Trans. Amer. micr. Soc.

75 : 314, figs. 11-12; Zumpt & Till, 1958, /. ent. Soc. S. Afr. 21 : 267.
Haemolaelaps mystromys Radford i942a, Parasitology 34 : 300, figs. 3-4 (syn. nov.).

FEMALE (figs. 20-22) : Deutosternum with 6 rows of 3-4 small teeth. Chelae
56-64^ long; pilus dentilis slender; arthrodial filaments subequal in length.

Dorsal shield ovoid, reticulate. DL=656-827[z; DW= 394-500^; DL/DW=
1-6-1-7. It bears 39 pairs of setae of moderate length, one or two unpaired
accessory setae sometimes being present. Lengths of 24 and Z2 about 4/5 the distance
between their bases; anterior and marginal setae a little longer; seta Z$ at least
i \ times as long as /5.

Sternal shield reticulate, merging anteriorly with presternal area, pattern weakly
developed posteriorly. SL= 105-133^; SW= 142-150^; SL/SW= 0-7-0-9.

Genital shield apparently smooth, rather slender, and may be slightly widened
behind the genital setae. GW^ 102-128^; GW 2 = 94-106^; GW 1 /GW 2 =i-i-i-3.
Length of genital setae about 2/3 GW 2 .

Anal shield pear-shaped, 1-3-1-5 times longer than broad. AL= 114-134^;
AW=76-ioo[z. Anus situated at approximately its length from the anterior
margin. Paranal setae level with middle of anus, about 4/5 as long as postanal seta.

Integument of idiosoma bears 23 pairs of ventral and marginal setae, some of the
latter having fine barbs. Metapodal plates elongate, length to width about 4:1.

Chaetotaxy of legs normal. Anterior seta on coxa I about 4/5 as long as posterior
seta. Tarsus II has 3 stout, blunt, terminal spines and 3 stout ventral spines.
Caruncle of tarsus IV relatively long, its length being 2f times the greatest width
of the tarsus. Length of tarsus IV about 8 times its width at the base.

MALE (figs. 23-24): Chelicera as figured, with a slender pilus dentilis. Dorsal
shield reticulate, chaetotaxy as in female. DL=570[j.; DW=323[x; DL/DW=i-8.
Holoventral shield reticulate, with 23 setae. Integument bears about 23 pairs of
marginal and ventral setae, Chaetotaxy of legs as in female, but femur II with a
stout spur,

W. M. TILL

FIGS. 20-24. Androlaelaps capensis (Hirst). Venter (Fig. 20), dorsum (Fig. 21) and
chelicera (Fig. 22) of female, Venter (Fig. 23) and chelicera (Fig. 24) of male,

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 23

HOSTS AND LOCALITIES. Cryptomys hottentotus (Lesson), Grahamstown, Cape

Province (type series in B.M.N.H.) and several localities in Natal and the Transvaal

(S.A.I.M.R.).

Cryptomys holosericeus (Wagner), Kruger National Park, Transvaal (S.A.I.M.R.).

Georychus capensis (Pallas), Knysna, Cape Province (S.A.I.M.R.).

Mystromys albicaudatus (Smith), Albany, Cape Province (types of H. mystromys in

B.M.N.H.).

Androlaelaps casalis (Berlese)

Iphis casalis Berlese, 1887, Acari, Myr. Scorp. Ital. Fasc. 38 N.8, figs. 3-5.

Haemolaelaps casalis, Bregetova, 1956, Acad. Sci. U.S.S.R. 61 : 89 & 99, figs. 143-4.

Hypoaspis oculatus Oudemans, 1915, Arch. Naturgesch. (A) 81 : 134.

Haemolaelaps molestus Oudemans, 1929, Ent. Ber. 8 (169) : 13; Willmann, 1939, Abh. Nat.

Ver. Bremen 31 : 173, figs. 5-7.

Atricholaelaps megaventralis Strandtmann, 1947, Proc. ent. Soc. Wash. 49 : 112.
Hypoaspis freemani Hughes, 1948, Mites associated with stored food products: 129, figs. 173-179;

Strandtmann, 1956, /. Kansas ent. Soc. 29 : 138.
Haemolaelaps haemorrhagicus Asanuma, 1952, Misc. Rep. Res. Inst. Nat. Res. No. 25 : 87, fig. i.

(For further references see Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites:
34)-

FEMALE (figs. 25-27) : Deutosternum with 6 rows each of 5 fine teeth. Chelae
34-40^ long ; pilus dentilis slender ; arthrodial filaments subequal in length.

Dorsal shield oval, reticulate. DL=62o-7iojj.; DW= 376-429^; DL/DW
1-5-1-7. It bears 39 pairs of symmetrically arranged setae, as well as a variable
number of unpaired accessory setae between the / series. Setae 34 and Z2 are
longer than half the distance between their bases; the longer posterior setae are
barbed; seta Z$ is ii-i|- times as long as /5.

Sternal shield reticulate; SL= 86-96^; SW= 124-140^; SL/SW=o-7-o-8. First
pair of sternal setae about as long as the metasternal setae, but a little coarser;
second and third pairs of sternal setae slightly longer than first pair.

Genital shield expanded and broadly rounded posteriorly, with a well-defined
reticulate pattern which extends at least as far as the anterior flap of the shield.
GW 1 =i43-i66[jL; GW 2 = 96-105^; GW 1 /GW 2 =i-4-i-6. Genital setae short, less
than half GW 2 .

Anal shield reticulate, approximately as long as broad. AL=9O-H4fx; AW=
96-ii4fji. Anus situated at about its length from the anterior margin. Paranal
setae inserted on a level between middle and posterior end of anus, and approximately
the same length as the postanal seta.

Integument of idiosoma bears about 21 pairs of ventral and marginal setae.
Metapodal plates elongate, length 4-7 times the width.

Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than anterior
seta. Three ventral and 3 apical setae on tarsus II are thickened, but do not form
blunt spines. Setae generally rather fine and of moderate length. Length of
tarsus IV about 6 times width at base.

MALE (figs. 28-29) : Chelicerae as figured, the fixed digit bearing a slender pilus
dentilis. Chaetotaxy of dorsal shield as in the female. DL= 466-485^; DW=

W. M. TILL

FIGS. 25-29. Androlaelaps casalis (Berlese). Venter (Fig. 25), dorsum (Fig. 26) and
chelicera (Fig. 27) of female. Venter (Fig. 28) and chelicera (Fig. 29) of male.

ETHIOPIAN MITES OE THE GENUS ANDROLAELAPS BERLESE s. hit. 25

284-322(0.; DL/DW 1-5-1-6. Holoventral shield reticulate, expanded behind
coxae IV, bearing 23 setae. Chaetotaxy of legs as in female.

HOSTS AND LOCALITIES : Cosmopolitan, from a wide variety of bird and mammalian
hosts, and also found in straw, hay, detritus, etc. Recorded from the Ethiopian
region from Rattus rattus (Linnaeus) and Xerus inauris (Zimmermann), Transvaal,
and from Sciurus carolinensis Gmelin, Cape Province (S.A.I.M.R.).

Androlaelaps centrocarpus (Berlese)

Haemolaelaps centrocarpus Berlese, 1911, Redia 7 1432; Costa, 1961, Bull. Brit. Mus. (nat.

Hist.) Zool. 8 : 27, figs. 39~44-
Haemolaelaps mauritanicus Hirst, 1925, Bull. Com. Etud. Hist. Sci. : 98, fig. 3 (syn. nov.).

FEMALE (figs. 30-32) : Deutosternum with 6 rows of 2-3 teeth. Chelae 6o[o, long;
pilus dentilis slightly inflated, tapering towards the tip ; arthrodial filaments subequal
in length.

Dorsal shield reticulate, widest at level of setae r6, then tapering gradually
posteriorly. DL= 836-911(0.; DW= 485-523(0.; DL/DW= 1-7-1 -8. The shield bears
additional paired setae in the ax and pxi positions, as well as numerous unpaired
setae on the postero-median part and extending anteriorly as far as setae 25. The
setae are long and subequal in length, except for /5, which are shorter. Lengths
of setae 24 and 22 approximately i| times the distance between their bases; seta
Z5 at least i| times as long as seta /5. Some marginal setae have fine barbs.

Sternal shield reticulate. SL= 105-130(0.; SW= 180-190(0.; SL/SW=o-6-O7.
Metasternal setae relatively short, approximately half as long as the sternal setae.

Genital shield granular, with a weakly defined reticular pattern posteriorly;
slightly widened behind the genital setae, then tapering abruptly. GW X = 143-162(0. ;
GW 2 = 130-143(0.; GW 1 /GW 2 =i-i. Genital setae fairly long, about 4/5 GW 2 ,
reaching to base of second pair of flanking setae.

Anal shield pear-shaped, anterior margin straight. AL= 120-145(1; AW=
110-124(0.; AL/AW=i-i-i-3. Anus situated at less than its length from the
anterior margin. Paranal setae level approximately with middle of anus, very
slightly longer than postanal seta.

Integument of idiosoma bears numerous long ventral and marginal setae, many
of the latter being barbed. Metapodal plates elongate, length to width about 6 : i.

Chaetotaxy of legs normal except that tibia III bears 9 setae. Posterior seta of
coxa I thicker than anterior seta, but approximately the same length. Femur I
has 2 short, stout, spine-like setae on its dorsal surface, and trochanter I has a stout,
spine-like dorsal seta. Leg II has a ventral spine-like seta on the trochanter,
femur and genu, and 2 on the tibia. Tarsus II has 3 spine-like ventral setae, and
3 stout, blunt, apical setae. The upper dorsal seta on trochanter IV is considerably
thicker than the other setae on this segment. Length of tarsus IV nearly 10 times
width at base.

MALE (figs. 33-34) : Chelicerae as figured, with a pilus dentilis resembling that of
the female. Dorsal shield as in female. And shield separated from sternito-
ventral shield, the latter strongly reticulate and bearing 7 pairs of setae, 2 pairs
being placed on the area posterior to coxae IV. Chaetotaxy of legs as in female.

W. M. TILL

FIGS. 30-34. Androlaelaps centrocarpus (Berlese). Venter (Fig. 30), dorsum (Fig. 31)
and chelicera (Fig. 32) of female. Venter (Fig. 33) and chelicera (Fig. 34) of male.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lat. 27

HOSTS AND LOCALITIES. Berlese recorded this species from " Mures " in British
Somaliland. The above description is based on females from Taterillm species,
Kerio Valley, Kenya (S.A.I.M.R.) and on males from Gerbillus gerbillus Olivier,
Israel (received by courtesy of Dr. M. Costa).

FIGS. 35-37. Androlaelaps congoensis sp. nov., female. Venter (Fig. 35); dorsum

(Fig. 36); chelicera (Fig. 37).

NOTE : The specimens described above have been compared with drawings of the
type made by Dr. G. O. Evans. Haemolaelaps mauritanicus Hirst, from Gerbillus
pyramidum Geoffrey, Mauritania, is most probably a synonym of A. centrocarpus.
It is slightly smaller, its setae are relatively finer, and the ventral setation is more
symmetrical than in A. centrocarpus.

W. M. TILL

Androlaelaps congoensis sp. nov.

FEMALE (figs. 35-37): Deutosternum with 6 rows of 3-5 very fine teeth and an
ill-defined 7th row with 4 minute teeth. Chelae 48^ long, pilus dentilis small,
slightly inflated, terminating in a sharp point ; arthrodial filaments short, subequal
in length.

FIGS. 38-41. Androlaelaps congoensis sp. nov., venter (Fig. 38) and chelicera (Fig. 39)
of male.
Androlaelaps cricetomydis sp. nov.. venter (Fig. 40) and chelicera (Fig. 41) of male.

Dorsal shield reticulate. 01^=703-722^; DW= 456-504^; DL/DW=i-4~i'6.
The shield bears the usual 39 pairs of setae as well as 11-13 unpaired accessory
setae between the / series. Lengths of setae 24 and Z2 approximately equal to the
distance between their bases ; seta Z5 twice as long as seta 75. Setae all apparently
simple except seta Z$ which has very fine barbs.

Sternal shield reticulate, merging with presternal area, SL (from level of first

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BEKLESE 5. la/. 29

sternal setae) = 105(0,; SW 114-120(0.; SL/SW^o-g. Sternal and metasternal
setae subequal in length.

Genital shield patterned posteriorly, greatly expanded. GW X = 205-219(0.;
GW 2 = 1 14-133(0.; GW 1 /GW 2 =i-6-l'8. Genital setae relatively short, slightly
more than half GW 2 .

Anal shield broader than long. AL= 109(0.; AW= 12410.; AL/AW=o-9. Anus
situated at approximately its length from the anterior margin of the shield. Paranal
setae near posterior margin of anus, extending beyond base of postanal seta and
about 3/5 as long as postanal seta.

Integument of idiosoma bears 21 pairs of ventral and marginal setae, all apparently
simple. Metapodal plates elongate, length about 4 times the breadth.

Genu III bears 10 setae and tibia III has 9. Anterior seta on coxa I is about 2/3 as
long as posterior seta. Posterior ventral seta on femur IV extremely long, about
3/5 the total length of the femur. Tarsus IV long and slender, its length about
7 times its width at the base.

MALE (figs. 38-39) : Deutosternum with 4 rows of very fine teeth, the 5th and
6th rows ill defined. Fixed digit of chelicera bears a short pilus dentilis, as in the
female; spermadactyl about Sojo. long.

Dorsal shield 560(0, long, 330(0, wide. Holoventral shield bears 23 setae. Chaeto-
taxy of legs as in female.

HOSTS AND LOCALITIES: Holotype ($), allotype (<^) and 7 female paratypes
(1933.11.14.309-318 part) collected by F. W. ]. Cox from the nest of Spermestes
cucullatus Swainson at Mulungu, Belgian Congo, 5 November, 1919. This species
appears to be very common in birds' nests and has also been found in the nests of
a serin finch, weaver, oriole, sparrow, bunting and wagtail, all from Zomba,
Nyasaland. The specimens are all in the collection of the British Museum (Natural
History) .

Androlaelaps cricetomydis sp. nov.

FEMALE (figs. 42-44) : Deutosternum bears 6 rows of 3-4 small teeth. Chelae
about 45(0. long; pilus dentilis inflated; arthrodial filaments subequal in length.

Dorsal shield reticulate, widest between the posterior setae Si and 82. DL=
850-883(0.; DW= 560-599(0.; DL/DW=i-5. The shield bears 39 pairs of setae,
accessory setae being absent ; setae on central part of shield relatively short, lengths
of 54 and Z2 being less than half the distance between their bases. Marginal setae
longer, coarser and barbed; Si more than twice as long as /i, Z$ about i% times as
long as /5.

Sternal shield reticulate, heavily sclerotized. SL= 128-138(0.; SW= 190-195(0.;
SL/SW=o-7.

Genital shield greatly expanded behind the genital setae, posterior margin straight
or slightly concave. The reticulate pattern extends anterior to the genital setae
where it becomes less well defined. GW X = 238-247(0.; GW 2 = 128-138^; GW X /GW 2
= 1-8-1-9. Genital setae relatively short, less than half GW 2 .

Anal shield slightly broader than long. AL= 133-144(0.; AW= 152(0.. Anus

3 o W. M. TILL

situated at less than its length from the anterior margin. Paranal setae level with
posterior margin of anus, about as long as postanal seta.

Integument of idiosoma bears about 23 pairs of marginal and ventral setae. The
3 pairs of ventral setae which flank the genital shield are apparently simple, the
remaining ventral and marginal setae are barbed. Metapodal plates approximately
twice as long as broad.

42 43

FIGS. 42-44. Androlaelaps cricetomydis sp. nov. Venter (Fig. 42), dorsum (Fig. 43)

and chelicera (Fig. 44) of female.

Chaetotaxy of legs normal, except that tibia III bears 9 instead of 8 setae.
Tarsus II bears 3 stout, ventral setae; terminal setae not particularly stout.
Anterior and posterior setae on coxa I approximately equal in length. Length of
tarsus IV about 6|- times its width at the base.

MALE (figs. 40-41) : Deutosternum bears 6 rows of 2-5 teeth ; pilus dentilis inflated
basally. Dorsal shield 650^ long, 437^ wide ; chaetotaxy similar to that of female.
Holoventral shield reticulate, expanded behind coxae IV and bearing 23 setae,
5 pairs lying in the region between coxae IV and the anus. Chaetotaxy of legs
similar to that of the female.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 31

HOST AND LOCALITY: The mites, received from Mr. H. J. Disney, were found in
the food store of Cricetomys gambianus Waterhouse, Malengo Highlands, Tanganyika,
30 August, 1960. Holotype female (1962.6.12.1), allotype male (1962.6.12.2)
and ten female paratypes (1962.6.12.3-7) in the collection of the British Museum
(Natural History).

FIGS. 45-47. Androlaelaps cryptomius (Radford), female. Venter (Fig. 45); dorsum

(Fig. 46); chelicera (Fig. 47).

Androlaelaps cryptomius (Radford)

Ischnolaelaps cryptomius Radford, 1939, Parasitology 31 : 248, fig. 5.

Hypoaspis (Haemolaelaps) cryptomius, Zumpt & Patterson, 1951, J. ent. Soc. S. Afr. 14 : 71.
Haemolaelaps cryptomius, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 241.
Liponyssus lawrencei Zumpt & Patterson, 1951,7. ent. Soc. S. Afr. 14 : 89, fig. 10; Zumpt &
Till, 1953, S. Afr. J . Med. Sci. 18 : 8 (syn. nov.).

FEMALE (figs. 45-47) : Deutosternum provided with 6 rows of teeth, 2 large and
2 small teeth in the first row, i large tooth in the 2nd to 5th rows, and 2 small teeth
in the sixth row. Anterior rostral seta about i| times as long as capitular seta.
Chelae small, about 24^ long. Fixed digit weakly sclerotized, bearing a long,
slender pilus dentilis; movable digit in two parts, one part edentate except for
terminal hook, the other part weakly sclerotized but with large teeth. Arthrodial
filaments not distinct in the specimens examined.

W. M. TILL

Dorsal shield widest at the level of setae ?5, with a reticulate and granular pattern.
DL= 508-530^; DW= 350-385^; DL/DW=i-4-i-5. The shield bears 39 pairs of
setae. Those on the central part are short and fine, the lengths of setae i<\ and Z2
being less than 1/5 the distance between their bases. The anterior and marginal
setae are longer, and seta Z$ is approximately 2.\ times as long as seta /5-

Sternal shield granular, with some reticulation near the lateral margins ; presternal
area reticulate. SL=7o~75[j.; SW= 130-140^ ; SL/SW=o-5-o-6. First pair of

FIGS. 48-50. Androlaelaps cryptomius (Radford), venter of male (Fig. 48).
Androlaelaps dasymys (Radford), venter (Fig. 49) and chelicera (Fig. 50) of male.

sternal setae 4/5 as long as second and third pairs : metasternal setae about half as
long as first pair of setae.

Genital shield relatively short, granular, very slightly widened behind the genital
setae, with a weakly defined reticular pattern. GW X = 96-104^; GW 2 = 84-88^.
Genital setae relatively long, about 2/3 GW 2 .

Anal shield approximately as long as broad, length and width varying between
94(x and 104^. Anus situated very close to anterior margin. Paranal setae level
approximately with middle of anus, slightly longer than postanal seta, but not
reaching the base of the latter.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s.lat.

Integument of idiosoma bears about 24 pairs of ventral and marginal setae.
Metapodal plates elongate, their length to width about 8 : i.

Chaetotaxy of legs normal. Posterior seta on coxa I very slightly longer and
thicker than anterior seta. The leg setae are all simple and are not modified to form
thickened or blunt spines. Length of tarsus IV 5-6 times its width at the base.

FIGS. 51-53. Androlaelaps dasymys (Radford), female (from Mystromys albicaudatus)
Venter (Fig. 51); dorsum (Fig. 52); chelicera (Fig. 53*2). Pilus dentilis of specimens
from Tatera afra (Fig. 536) and Otomys sp. (Fig. 53c).

MALE (fig. 48) : Structure of the chelicerae not very distinct in the only specimen
available, but there appears to be a flagellar pilus dentilis as in the female. Dorsal
shield similar to that of the female, but setae on central part relatively longer.
DL=475pi.; DW = 284^; DL/DW=i-7. Holoventral shield expanded behind
coxae IV and bears 21 setae. The shield has an overall reticulate pattern, and also
a granular appearance especially anterior to coxae IV. Chaetotaxy of legs as in
female.

HOSTS AND LOCALITIES: Georychus capensis (Pallas) from Wynberg, Cape Province

ZOOL. IO, I 3

34 W. M. TILL

(types in Dr. Radford's collection and M.R.A.C.), and from Knysna and Drostdy,
Cape Province (S.A.I.M.R.).

Androlaelaps dasymys (Radford) (s. lat.)

[schnolaelaps dasymys Radford, 1939, Parasitology 31 : 245, fig. 2.

Hypoaspis (Haemolaelaps) dasymys, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 73.

Haemolaelaps dasymys, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 247.

Hypoaspis (Haemolaelaps) labuschagnei Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 85,

fig. 7 (syn. nov.).
Haemolaelaps davisi Zumpt & Till, 1956, Z. Parasitenk 17 : 287, figs. 5-8 (syn. nov.).

FEMALE (figs. 51-53) : Deutosternum with 6 rows of 3-4 teeth, one tooth in each
row being larger than the others in some specimens. Chelae 50-65^ long; pilus
dentilis as figured, its terminal portion appearing either swollen or ribbon-like,
depending on the orientation; two or three of the arthrodial filaments distinctly
longer than the others.

Dorsal shield reticulate, with 39 pairs of setae; unpaired accessory setae may be
present (2-6) or absent. Lengths of setae 24 and Z2 at least 4/5 the distance between
their bases, up to 1-4 times this distance in specimens from Tatera afra, Rattus
rattus, Parotomys species, and some specimens from Otomys species. Marginal
setae coarse, barbed; seta Z$ 1-6-2-5 times as long as seta /5. DL= 822-1 064^;
DW=5i6-750|^; DL/DW=i-4-i-7.

Sternal shield reticulate, sharply demarcated from presternal area. SL=
123-163^.; SW=i54-200[x; SL/SW=o-8-i-o. Metasternal setae a little shorter
than first pair of sternal setae, second and third pairs of sternal setae slightly longer
than first pair.

Genital shield with a distinct reticulate pattern extending almost to the anterior
extremity. GWj= 170-209^; GW 2 =ii4-i65[x; GW 1 /GW 2 =i-3-i-5. Length of
genital seta slightly more than half GW 2 .

Anal shield reticulate anteriorly and laterally, approximately as long as broad.
AL=i27-200[z; AW=i27-i95[x. Anus situated at a little less than its length from
the anterior margin ; paranal setae level with middle of anus, 4/5 as long as postanal
seta and extending beyond its base.

Integument of idiosoma bears 18-28 pairs of ventral and marginal setae, the
latter being longer, coarser, and provided with fine barbs. Metapodal plates vary
in shape from roughly elliptical to almost round, their length 1-4-2-5 times their
width.

Chaetotaxy of legs normal. Posterior seta on coxa I very slightly longer and
thicker than anterior seta. Tarsus II bears three stout ventral setae but no blunt
spines ; trochanter IV bears a stout seta on its dorsal surface. Length of tarsus IV
about 7 times its width at the base.

MALE (figs. 49-50) : Chelicera as figured ; dorsal shield similar to that of female.
DL=684-788[x; DW 418-420^; DL/DW= 1-6-1-9. Holoventral shield reticulate,
expanded behind coxae IV, bearing 23 setae. Chaetotaxy of legs as in female.

HOSTS AND LOCALITIES: Dasymys incomtus (Sundevall), Kakumiro, Uganda
(syntype of /. dasymys in B.M.N.H.).

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal.

FIGS. 54-58. Androlaelaps galagus (Lavoipierre). Venter (Fig. 54), dorsum (Fig. 55)
and chelicera (Fig. 56) of female. Venter (Fig. 57) and chelicera (Fig. 58) of male.

36 W. M. TILL

Mystromys albicaudatus (Smith) from Basutoland (type series of H. davisi, the

figured specimens) and from Aliwal North, Cape Province (S.A.I.M.R.).

Rattus rattus (Linnaeus) from the Transvaal (type series of H. labuschagnei in

S.A.I.M.R.).

Rattus natalensis (Smith) from Natal, Cape Province, Transvaal, Orange Free State

(S.A.I.M.R.).

Rattus chrysophilus (De Winton) and Rattus namaquensis (Smith) (Zumpt & Till,

1961; material not re-examined).

Otomys irroratus (Brants) and Otomys species from several localities in South Africa

(S.A.I.M.R.).

Otomys saundersiae Roberts (Zumpt & Till, 1961; material not re-examined).

Otomys sloggetti (Thomas), Basutoland (S.A.I.M.R.).

Parotomys brantsi (Smith), Port Nolloth, Cape Province (S.A.I.M.R.).

Parotomys littledalei Thomas, van Rhynsdorp, Cape Province (S.A.I.M.R.).

Tatera afra (Gray), Basutoland and Transvaal (S.A.I.M.R.).

Rhabdomys pumilio (Sparrman), Ngorongoro, Tanganyika (M.R.A.C.).

Lophuromys sikapusi (Temminck), Cameroons (Taufflieb & Mouchet, 1959).

Androlaelaps galagus (Lavoipierre)

Haemolaelaps galagus Lavoipierre, 1955, Ann. trap. Med. Parasit. 49 : 304, figs. 5-6.

FEMALE (figs. 54-56) : Deutosternum with 6 rows each of 2 small teeth. Chelae
54-56^1 long; pilus dentilis inflated, elongate, bent near tip and terminating in a
fine point; arthrodial filaments subequal in length.

Dorsal shield reticulate, bearing 39 pairs of setae. Those on the central part of
the shield are very short and fine ; they are extremely minute and difficult to detect
in the type specimen examined (by courtesy of Dr. M. M. J. Lavoipierre), but are
more conspicuous in the specimens from Cryptomys sp., on which the drawings are
based. Some of the setae are extremely long and barbed, namely, /4, 3-5 and
54-5; seta /4 is at least 5 times as long as /i. DL= 910-960^; DW= 605-709^;
DL/DW=i-3-i-5.

Sternal shield with a reticulate pattern which is most marked antero-laterally.
SL=i45(j,; SW= 205-2 14^; SL/SW=o-6-o-7- First sternal seta i| times as long
as metasternal seta, second and third sternal setae a little longer than first.

Genital shield sculptured posteriorly, greatly expanded behind the genital setae.
GW 1= 247-271 {j.; GW 2 =i52-i70[z; GW 1 /GW 2 = 1-5-17.

Anal shield approximately as wide as long. AL=i43-i62(x; AW= 152-170^.
Anus situated at approximately its length from the anterior margin. Parana!
setae inserted on a level between middle and posterior end of anus, about f as long
as postanal seta.

Integument of idiosoma bears 26 pairs of marginal and ventral setae, many of
which are barbed. Metapodal plates variable in shape, length to width about 2:1.

Chaetotaxy of legs normal ; anterior and posterior setae on coxa I approximately
the same length. Trochanter IV has 2 thick spine-like setae on its anterior surface,
and femur IV has a stout ventral seta and a stout dorsal spine. Antero-dorsal seta

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 37

FIGS. 59-61, Androlaelaps

georychi sp. nov., female.
(Fig. 60); chelicera (Fig. 61).

Venter (Fig.

59) ; dorsum

on this segment barbed, its length exceeding the width of the femur. Length of
tarsus IV about 6 times its width at the base.

MALE (figs. 57-58) : Chelicera with an inflated pilus dentilis which tapers distally.
Dorsal shield as in female, but setae on central part relatively longer, and 5 accessory
setae present between /3 and //j.. DL=694[j.; DW=456[z; DL/DW=i-5- Holo-
ventral shield reticulate, expanded behind coxae IV, bearing 20-21 setae. Chaeto-
taxy of legs as in female; one ventral seta on femur, genu, tibia and tarsus of leg II
stout and spine-like.

HOSTS AND LOCALITIES: Galago alleni Waterhouse, Barombikang Forest, British
Cameroons (type).

38 W. M. TILL

Cricetomys sp., burrow, Kivu, Belgian Congo (M.R.A.C.).

Cricetomys gambianus Waterhouse, Yaounde, Cameroons (Taufflieb & Mouchet,

1959)-

Androlaelaps georychi sp. nov.

FEMALE (figs. 59-61) : Deutosternum with 6 rows of 3-4 very fine teeth, a larger
median tooth in rows 2 to 4. Hypostomal processes very hairy and large, extending
to anterior margin of palp femur. Chelae about 58 [i long; pilus dentilis elongate
(about 40(ji long), moderately inflated; arthrodial filaments subequal in length.

Dorsal shield reticulate, ovoid in shape, widest at level of seta sj. DL= 964-988^ ;
DW=6i8[x; DL/DW=i-6. The shield bears the usual 39 pairs of setae, which are
more or less subequal in length, and 5-6 accessory setae in the postero-median
region. Lengths of setae 24 and Z2 about i-| times the distance between their
bases; seta ^5 i| times as long as seta 75.

Sternal shield reticulate; SL= 143-152^; SW= 185-200^; SL/SW= 0-7-0-8.
Genital shield with a reticulate pattern which extends over the anterior flap;
greatest width of shield 1-4-1-5 times width at level of genital setae (GW 2 ). Length
of genital setae about 4/5 GW 2 .

Anal shield approximately as broad as long. AL= 143-1 62 [z; AW =148-162^.
Anus situated at less than its length from the anterior margin of the shield. Paranal
setae near posterior margin of anus, extending to a point level with the posterior
tip of the anal shield, their length about f that of the postanal seta.

Integument of idiosoma bears numerous ventral and marginal setae which all
appear to be barbed, except those adjacent to the genital shield. Metapodal plates
short, ovoid, length to width not greater than 2:1.

Chaetotaxy of legs normal. Length of tarsus IV about 8 times its greatest width.
Caruncle IV elongated, its length more than twice the width of tarsus IV and about
i^ times the length of caruncle III.

HOST AND LOCALITY: Four females from Georychus capensis (Pallas), Knysna,
Cape Province, 24 April, 1954. Holotype and one paratype in the collection of the
S.A.I.M.R., Johannesburg, two paratypes (1962.6.12.8-9) in the collection of the
British Museum (Natural History).

Androlaelaps ghanensis sp. nov.

FEMALE (figs. 62-64) : Deutosternum with 6 rows of 2-4 teeth. Chelae about
62 [x long; pilus dentilis moderately inflated; arthrodial filaments subequal in length.

Dorsal shield reticulate, 817-903^ long, 570-580^ wide at the broadest point,
DL/DW=i-4-i-6. The shield bears 39 pairs of setae and one unpaired accessory
seta at the level of setae 74- All the setae except ii and ri are long and thick.
Lengths of setae 24 and Z2 about twice the distance between their bases; seta Z$
about 1 1 times as long as seta 75 ; seta Ji reaches almost to the base of seta 74-

Sternal shield reticulate, approximately as long as broad; SL=i7i-i8o(x;

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 39

SW=i8o[x. Metasternal setae about 2/3 as long as first pair of sternal setae, second
and third pairs of sternal setae slightly longer.

Genital shield widened near its posterior end and with a distinct pattern extending
to its anterior margin. GW X = 143-152^ ; GW 2 =io5p.; GW 1 /GW 2 =i-4. Genital
setae about as long as first pair of sternal setae, slightly shorter than GW 2 .

Venter (Fig. 62) ; dorsum

FIGS. 6264. Androlaelaps ghanensis sp. nov., female.

(Fig. 63); chelicera (Fig. 64).

Anal shield as long as broad or slightly longer. AL= 157-171^;
AL/AW=i-o-i-i. Anus situated at approximately its length from the anterior
margin of the shield. Paranal setae level with posterior margin of anus, about
3/5 as long as postanal seta.

Integument of idiosoma bears about 27 pairs of ventral and marginal setae; the
marginal and outermost ventral setae are all barbed. Metapodal plates oval in
shape, length to width about 5 : 2.

4 o W. M. TILL

Chaetotaxy of legs normal. Trochanter IV bears two very stout spines, one on
the lower ventral and one on the upper dorsal surface. Femur IV bears a stout
sword-like basal spine. Length of tarsus IV about 6 times the width at the base.

HOST AND LOCALITY: Four females from a dormouse (Claviglis spec.) in Ghana,
i January, 1911. Holotype (1962.6.12.10) and three paratypes (1962.6.12.11-13) in
the collection of the British Museum (Natural History).

Androlaelaps glasgowi (Ewing)

Laelaps glasgowi Ewing, 1925, Proc. ent. Soc. Wash. 27 : 6.

Haemolaelaps glasgowi, Strandtmann, 1949, /. Parasit. 35 : 343 ; figs. ; Zumpt & Till, 1956,
Z. Parasitenk. 17 : 282, figs. 1-2.

Laelaps californicus Ewing, 1925, Proc. ent. Soc. Wash. 27 : 5.
Laelaps virginianus Ewing, 1925, Proc. ent. Soc. Wash. 27 : 6.

Hypoaspis microti Oudemans, 1926; Ent. Ber. 7 : 101; Bregetova, 1952, Zool. Zh. 31 : 867.
Haemolaelaps mohrae Oudemans, 1928, Ent. Ber. 7 : 374.

Haemolaelaps cricetophilus Vitzthum, 1930, Zool. Jb. Jena, Abt. Syst. 60 : 417, figs.
Laelaps stegemani Hefley, 1935, /. Kansas ent. Soc. 8 : 22.

1 ' schnolaelaps rhabdomys Radford, 1939, Parasitology 31 : 249, fig. 6 (syn. nov.).
Haemolaelaps scalopi Keegan, 1946, Trans. Amer. Micr. Soc. 65 : 71.
Atricholaelaps sigmodoni Strandtmann, 1946, J. Parasit. 32 : 164.
Atricholaelaps strandtmanni Fox, 1947, Ann. ent. Soc. Amer. 40: 580, figs.
Hypoaspis (Haemolaelaps) eos Zumpt & Patterson, 1951, J. ent. Soc. S. Afr. 14 : 79, fig. 2.
(For further references see Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites : 37)

FEMALE (figs. 65-67) : Deutosternum with 6 rows of 3-5 teeth. Chelae about
36(0, long; pilus dentilis strongly inflated basally, distal portion slender, curved or
recurved; arthrodial filaments usually subequal in length, slightly variable in
populations from Otomys irroratus.

Dorsal shield reticulate, with a " double " margin. DL= 613-653^; DW=
394-436^; DL/DW= 1-5-1-6. The shield bears 39 pairs of setae and 2-3 unpaired
accessory setae between setae /2 and 74- Lengths of setae 24 and 22 slightly exceed
the distance between their bases; seta Z$ about twice as long as 75-

Sternal shield reticulate and granular. SL 96-113^; SW= 130-155^; SL/SW
0-75. Genital shield slightly widened behind genital setae, its reticular pattern
extending over the anterior flap. GW 1 =i23-i4O[x; GW 2 =96-io5{ J t; GW 1 /GW 2 =
1-2-1-4. Genital setae relatively short, about half GW 2 .

Anal shield approximately as long as broad. AL= 90-109^; AW= 100-105^.
Anus situated at less than its length from the anterior margin ; paranal setae level
with middle of anus, about 2/3 as long as postanal seta.

Integument of idiosoma bears 22 pairs of ventral and marginal setae, the latter
being a little coarser than the former and provided with fine barbs. Two pairs of
ventral setae flank the genital shield. Metapodal plates 3 or 4 times longer than
broad.

Chaetotaxy of legs normal ; posterior seta on coxa I slightly thicker than anterior
seta, but subequal in length. Length of tarsus IV at least 6 times its greatest
width.

MALE (figs. 68-69) : Chelicera as illustrated, pilus dentilis inflated basally. Dorsal
shield reticulate, chaetotaxy as in female. DL=428[x; DW=257[x; DL/DW=i-7.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE ,. //.

FIGS. 65-69. Androlaelaps glasgowi (Ewing). Venter (Fig. 65), dorsum (Fig. 66) and
chelicera (Fig. 67) of female. Venter (Fig. 68^ and chelicera (Fig. 69) of male.

42 W. M. TILL

Holoventral shield reticulate, expanded behind coxae IV and bearing 23 setae.
Integument with about 14 pairs of ventral and marginal setae. Chaetotaxy of legs
as in female.

HOSTS AND LOCALITIES: Rhabdomys pumilio (Sparrman) and Otomys irroratus
(Brants) from several localities in Southern Africa (S.A.I.M.R. ; B.M.N.H; types of
/. rhabdomys in Dr. C. D. Radford's collection).

FIGS. 70-73. Androlaelaps graingeri Zumpt & Patterson, female. Venter (Fig. 70);
dorsum (Fig. 71); chelicera (Fig. 72); femur, genu, tibia of leg II (Fig. 73).

Androlaelaps graingeri Zumpt & Patterson

Androlaelaps graingeri Zumpt & Patterson, 1952, /. ent. Soc. S. Afr. 15 : 162, fig. 2.
Turkiella graingeri, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 224.

FEMALE (figs. 70-73): Gnathosoma ventral in position in all the specimens
examined, only the free segments of the palps extending beyond the anterior margin
of the idiosoma. Deutosternum with 6 rows of 3-4 teeth, only one tooth visible
in the 5th row. Chelae 74-76^ long, both digits dentate, movable digit with a
pronounced terminal hook which arches over the fixed digit ; pilus dentilis slender,
arthrodial filaments subequal in length.

Dorsal shield subcircular, with a " double " margin, granular and weakly reticulate.
DL=iooo-io50[x; DW= 825-900^; DL/DW=i-2. Chaetotaxy of dorsal shield
similar to that of A . arvicanthis, i.e. setae px2 and px$ missing. Seta ri considerably
longer than seta ii ; seta Z$ about twice as long as seta /5 ; setae 4-*5> 22, /i-/4

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat.

rubbed off in all the specimens examined, but are probably subequal with 23, as
they are in the male. Setae generally finer than in A. arvicanthis and apparently
without barbs.

Sternal shield with an overall granulation and very weak reticular pattern.
SL=8o-96[ji; SW= 197-228^; SL/SW=o-4-o-5. First sternal seta relatively
short, 2nd and 3rd sternal and the metasternal setae considerably longer.

FIGS. 74-76. Androlaelaps graingeri Zumpt & Patterson, venter of male (Fig. 74).
Androlaelaps haydocki (Till), male. Venter (Fig. 75) ; dorsal view of femur and genu of
leg I (Fig. 76).

Genital shield broadly rounded posteriorly but not greatly expanded behind the
genital setae, granular, with a weak reticular pattern. GW X = 2 15-236^; GW 2 =
i88-200{x; GW 1 /GW 2 =i-i 1-2. Genital setae relatively long, slightly less than
GW 2 , reaching base of second pair of flanking setae.

Anal shield reticulate, with a pair of lateral, granular patches; approximately
as long as broad. AL= 162-184^; AW= 158-175^1. Paranal setae near posterior
margin of anus, about 4/5 as long as postanal seta.

W. M. TILL

Integument of idiosoma with 21-22 pairs of fine setae. Metapodal plates
subcircular.

Chaetotaxy of legs normal. Anterior and posterior setae of coxa I approximately
equal in length. Ventral setae on genu and tibia II similar to those in A . arvicanthis.
Length of tarsus IV about 6| times the width at the base.

77 78

FIGS. 77-78. Androlaelaps haydocki (Till), female. Venter (Fig. 77); dorsum (Fig. 78).

MALE (fig. 74) : Gnathosoma as in female; spermadactyl about 115^ long. Dorsal
shield with chaetotactic pattern similar to that of female; DL=846[x; DW=646[x;
DL/DW=i - 3. Holoventral shield granular with a weak reticular pattern, not
expanded behind 4th pair of coxae, bearing 21 setae, of which 4 pairs lie in the preanal
region. Metapodal plates oval, length about 2| times width. Chaetotaxy of legs
as in female.

HOST AND LOCALITY: Unidentified rodent, Kerio Valley, Kenya (type series,
S.A.I.M.R.).

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 45

Androlaelaps haydocki (Till)

Haemolaelaps haydocki Till, 1959, /. ent. Soc. S. Afr. 22 : 426, figs.

Androlaelaps haydocki is very closely allied to A. mesopicos (Radford) and may
prove to be only a subspecies or variety.

FEMALE (figs. 77-78): The female differs from A. mesopicos in that the dorsal
shield bears 38 pairs of setae, and the marginal setae 57, S2, and 53 are long and

FIGS. 79-81. Androlaelaps heliosciuri sp. nov., female. Venter (Fig. 79); dorsum

(Fig. 80); chelicera (Fig. 81).

sinuous. The genital shield is slightly more expanded posteriorly (GW 1 /GW 2 =
1-3-1-5), and the anal shield is slightly more elongate (AL/AW=i-2-i-4).

MALE (figs 75-76) : The male differs from A . mesopicos in having a long dorsal
seta on genu I.

HOSTS AND LOCALITIES : Thripias namaquus (Lichtenstein) , Luanshya (type series)
and Muliashi Area, Northern Rhodesia; Debeete, Bechuanaland ; Mabelikwa, N.

46 W. M. TILL

Transvaal ; Mapalane, Mozambique (S.A.I.M.R. and B.M.N.H.).
Campethera abingoni (Smith), Luanshya, N. Rhodesia (B.M.N.H.).

Androlaelaps heliosciuri sp. nov.

FEMALE (figs. 79-81) : Deutosternum with 6 rows of 2-3 small teeth; corniculi very
weakly sclerotized. Chelae about 4O(x long, edentate except for terminal hook on
movable digit; pilus dentilis slender; two of the arthrodial filaments considerably
longer and thicker than the others.

Dorsal shield reticulate, widest in posterior half at level of seta 52. DL 618-
637^; DW=385-4O4(u.; DL/DW=i-6. All 39 pairs of setae are present, but rz,
r\ and r6 are off the shield. Lengths of setae 24 and 22 not greater than half the
distance between their bases; seta Z$ about 3 times as long as seta /5.

Sternal shield granular, with a faint reticular pattern as well. SL=76-89[j.;
SW=i28-i33[A; SL/SW=o-6-o-7. First pair of sternal setae situated off the
plate, slightly shorter than 2nd and 3rd pairs ; metasternal setae as long as 2nd and
3rd pairs of sternal setae.

Genital shield reticulate, the pattern extending over the anterior flap. It is
closely approached to the anal shield and widest at the level of the 2nd pair of
flanking setae. GW 1= 142-152^; GW 2 =9o-95^; GW 1 /GW 2 = 1-5-17. Genital
setae relatively short, less than half GW 2 .

Anal shield broader than long, anterior margin straight. AL=ioi-iiO[ji;
AW 1 20-123 [j.; AL/ AW 0-8-0-9. Paranal setae situated between middle and
posterior end of anus, about 2/3 as long as postanal seta.

Integument of idiosoma bears about 34 pairs of ventral and marginal setae,
the former being fine and simple, the latter coarser and barbed. Metapodal plates
long and slender, length to width at least 6 : i.

Chaetotaxy of legs normal. Femur IV bears a stout antero-dorsal seta. Length
of tarsus IV about 4 times the width at the base.

HOST AND LOCALITY: Eight females from Heliosciurus gambianus, (Ogilby),
Amani, Tanganyika, 3 February, 1957. Holotype and three paratypes in the collec-
tion of the South African Institute for Medical Research, four paratypes (1962.6.12.
14-17) in the collection of the British Museum (Natural History).

Androlaelaps hirsti (Keegan)

Haemolaelaps hirsti Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 242, figs. 26-30; Costa, 1961,
Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 14, figs. 17-20.

FEMALE (figs. 82-84) : Deutosternum with 6 rows each of 4 small teeth, except
the last row which has 3 teeth. Chelae about 80 [x long ; pilus dentilis inflated basally,
tapering distally; arthrodial filaments subequal in length.

Dorsal shield ovoid, with a " double " margin, widest about the middle, tapering
posteriorly, granular in appearance with faint reticulations. DL= 1092-1188^;
DW= 703-817^; DL/DW=i-4-i-6. The shield bears 41 pairs of setae, one pair
being present in the ax and one pair in the pxi position. One of the specimens

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 47

examined has an unpaired accessory seta between setae /4 ; the remaining specimens
have no accessory setae. The setae on the central part of the shield are relatively
long, the lengths of 24 and Z2 exceeding the distance between their bases. The
posterior terminal setae Z$ are approximately the same length as setae 14 and at
least 2\ times as long as setae 75.

FIGS. 82-84. Androlaelaps hirsti (Keegan), female. Venter (Fig. 82); dorsum (Fig. 83);

chelicera (Fig. 84).

Sternal shield granular with a faint reticulate pattern, presternal area reticulate
and slightly granular. SL= 162-180^; SW= 225-257^; SL/SW=o-6-o-8. Sternal
setae long, first pair reaching posterior margin of shield, third pair reaching middle
of coxae IV ; metasternal seta about half as long as second sternal seta.

Genital shield granular, not expanded posteriorly; GW 2 = 162-180^. Genital
setae relatively long, reaching base of second pair of flanking setae, their length
nearly equal to GW 2 .

Anal shield as long as, or slightly longer than, broad. AL= 175-190^; AW=

48 W. M. TILL

i62-i75(x; AL/AW=i-o-i-2. Anus less than its length from the anterior margin
of the shield. Paranal setae inserted on a level between middle and posterior margin
of anus, extending beyond base of postanal seta, and slightly longer than this seta.
Integument of idiosoma with numerous long marginal and ventral setae, many
of the former being barbed. Metapodal plates oval, length about twice breadth.

FIGS. 85-88. Androlaelaps hirsti (Keegan), venter (Fig. 85) and chelicera (Fig. 86) of
male.
Androlaelaps hystrici (Zumpt & Till), venter (Fig. 87) and chelicera (Fig. 88) of male.

Tibia III has 9 setae instead of the usual 8. Ventral setae on coxa I nearly equal
in length but posterior seta much thicker. Two short, stout, spur-like setae on
femur I ; one stout ventral seta on genu II, two on tibia II, 3 stout ventral and 3 stout
apical setae on tarsus II, two of the apical ones being relatively blunt. Length of
tarsus IV is 7-8 times its width at the base.

MALE (figs. 85-86) : Fixed digit of chelicera with a long slender pilus dentilis.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat.

Deutosternum with 6 rows of 2 or 3 teeth. Dorsal shield as in the female, but the
setae are relatively longer. Anal shield separated from the sternito-ventral shield
which bears 7 pairs of setae, 2 pairs being situated on the part posterior to coxae IV.
Chaetotaxy of legs as in the female.

HOSTS AND LOCALITIES : Unidentified rat from Yubo, Equatoria, Sudan (Paratype
in B.M.N.H.). The above description is based on specimens from Gerbillus
Pyramidum Geoffroy from Israel, received by courtesy of Dr. M. Costa.

' /'/ T\ '

FIGS. 89-91. Androlaelaps hystrici (Zumpt & Till), female.

(Fig. 90); chelicera (Fig. 91).

Venter (Fig. 89) ; dorsum

Androlaelaps hystrici (Zumpt & Till)

Haemolaelaps hystrici Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 235, figs. 12-16;
Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 239, fig. 23.

FEMALE (figs. 89-91): Deutosternum with 6 rows each of 3 teeth. Chelae 34^
long ; pilus dentilis slender ; arthrodial filaments subequal in length. Anterior rostral
seta at least i| times as long as capitular seta.

Dorsal shield ovoid, posterior margin almost straight. DL= 600-662^; DW=

ZOOL. IO, I 4

5 o W. M. TILL

290-362^; DL/DW =1-8-2-3. Only 36 pairs of setae are present on the shield,
r^-rb being inserted on the adjacent integument. Setae simple, of moderate length ;
lengths of 24 and Z2 about f the distance between their bases ; seta Z$ at least 1 1 times
as long as seta 75.

Sternal shield reticulate. SL= 93-1 14(1; SW= 103-134^; SL/SW=o-8-i-o. Setae
relatively short, first seta reaching base of second.

Genital shield slightly granular, with a few indistinct traces of a reticulate pattern,
very slightly widened behind genital setae, then tapering. GW-,^ 93-1 03^;
GW 2 = 83-93^. Length of genital setae about half GW 2 .

Anal shield pear-shaped. AL= 93-134^; AW =93-103^; AL/AW= 1-0-1-3.
Anus situated at approximately its length from the anterior margin. Paranal setae
inserted on a level between middle and posterior end of anus and approximately as
long as the postanal seta.

Integument of idiosoma bears about 22 pairs of ventral and marginal setae, of
which two pairs border the genital shield. Metapodal plates extremely narrow,
length to width about 10 : i.

Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than
anterior seta. Legs relatively slender; length of tarsus IV 8 times the width at
the base.

MALE (figs. 87-88) : Fixed digit of chelicera with a slender pilus dentilis. Dorsal
shield ovoid, 5oo[x long and 263^ wide ; DL/DW=i-9- Chaetotaxy as in the female,
except that r^-r6 are inserted on the shield. Holoventral shield narrow, bearing
19 setae. Metapodal plates broader than in the female. Chaetotaxy of legs as
in female, but some of the setae relatively stouter, for example, femur II has two
broad, spine-like ventral setae, tibia II has one very broad, spear-like ventral seta,
tarsus II has two broad, spine-like, ventral setae, but the apical setae are slender.

HOSTS AND LOCALITIES: Hystrix cristata Linnaeus from Karamoja district, Uganda
(type series in S.A.I.M.R. and B.M.N.H.) and from Njoro, Rift Valley Province,
Kenya (Keegan, 1956).

Keegan (1956) records this species from Arvicanthis niloticus (Desmarest) and
Rattus rattus (Linnaeus) in Egypt, and also an additional specimen, which should
probably be referred to this species, from Crocidura manni Peters, Kano, Nigeria.

Androlaelaps longipes (Bregetova)

Haemolaelaps longipes Bregetova, 1952, Zoo/. Zh. 31 : 867, figs.; Costa, 1961, Bull. Brit. Mus.

(nat. Hist.) Zoo/. 8 : 21, figs. 28-32.

Haemolaelaps namrui Radford, 1954, Fieldiana, Zoo/. 34 : 310, figs. 56-57 (syn. nov.).
Haemolaelaps aegyptius Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 246, figs. 37-38.

FEMALE (figs. 92-94) : Deutosternum with 6 rows of 3-4 small teeth. Chelae
about 55jj, long; pilus dentilis slightly inflated basally, distal 2/3 slender; arthrodial
filaments subequal in length.

Dorsal shield ovoid, with a reticulate pattern. ~DL=yo^[i ; DW=475[x ; DL/DW=
1-5. The shield bears 40 pairs of setae, one pair being present in the PXI position
on the posterior half of the shield. There is also an unpaired seta near the mid-line

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 51

between setae /3 and /4. The setae on the central part of the shield are relatively
long, the lengths of 14 and Z2 being approximately equal to the distance between
their bases. The anterior and marginal setae are longer, some of the longest setae
having fine barbs. Seta Z5 is more than twice as long as seta 75.

Sternal shield reticulate. SL=g^; SW=i47{ji; SL/SW=o7. Sternal setae
long, the first pair reaching almost to the posterior margin of the shield, second and

FIGS. 92-94. Androlaelaps longipes (Bregetova), female.

(Fig. 93); chelicera (Fig. 94).

Venter (Fig. 92) ; dorsum

third pairs slightly longer, the 3rd pair extending well beyond the bases of the
metasternal setae. Metasternal seta less than half as long as 3rd sternal seta.

Genital shield with a reticulate pattern, slightly widened behind genital setae,
then tapering posteriorly. GW^ 125(1,; GW 2 =i20fi.. Genital setae relatively
long, about 9/10 GW 2 , extending to base of second pair of flanking setae.

Anal shield slightly longer than wide, anterior margin straight. Length (to base
of postanal seta) is 103(1,, greatest width is 98^. Anus approximately equidistant
between anterior margin of shield and postanal seta. Paranal setae situated between
middle and posterior margin of anus, about the same length as the postanal seta.

52 W. M. TILL

Integument of idiosoma bears about 37 pairs of apparently simple ventral and
marginal setae, of which 2 pairs flank the genital shield. Metapodal plates elongate,
7 times as long as broad.

Chaetotaxy of legs normal except that tibia III has 9 setae. Posterior seta on
coxa I is 1 1 times as long and thick as anterior seta. Stout ventral setae are present
on the femur, genu, tibia and tarsus of leg II, three of the terminal setae on tarsus II
are stout and blunt. The dorsal seta on trochanter IV is much thicker than the

FIGS. 95-97. Androlaelaps longipes (Bregetova), venter (Fig. 95) and chelicera (Fig. 96)
of male.
Androlaelaps marshalli Berlese, venter of male (Fig. 97).

other setae on this segment. Length of tarsus IV about 8 times the width at the
base.

MALE (figs. 95-96): Fixed digit of chelicera (in A. longipes from the U.S.S.R.)
bears a large, tapering pilus dentilis. Dorsal shield similar to that of female.
Holo ventral shield divided into a sternito- ventral shield and a separate anal shield.
The sternito-ventral shield is moderately expanded behind the 4th pair of coxae
and bears 7-8 pairs of setae, 2 or 3 pairs being situated on the part posterior to

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s.lat. 53

coxae IV. The ventral shield is more reduced in the Russian specimen than in the
type specimen of A. namrui, on which the description and drawings are based.
Integument of idiosoma bears about 31 pairs of setae. Chaetotaxy of legs as in
the female.

HOSTS AND LOCALITIES: Meriones rex Yerbury & Thomas, Ta'izz, Yemen (types
of A. namrui in Dr. C. D. Radford's collection).

FIGS. 98-100. Androlaelaps marshalli Berlese, female. Venter (Fig. 98); dorsum
(Fig. 99) ; ventral view of femur, genu and tibia of leg II (Fig. 100).

Bregetova (1952) recorded A. longipes from Meriones species in the U.S.S.R. (material
in B.M.N.H.).

Keegan's specimens (A . aegyptius) were collected from a variety of hosts at several
localities in Egypt, and Costa (1961) has found this species chiefly on Meriones
species in Israel.

Androlaelaps marshalli Berlese

Laelaps (Androlaelaps) marshalli Berlese, 1911, Redia 7 : 429.

Androlaelaps marshalli, Keegan, 1956, /. Egypt, publ. Hthl Ass. 31 : 227, fig. 13; Costa, 1961,
Bull. Brit. Mus. (nat. Hist.) Zool. 8 : 45, figs. 75~79-

54 W. M. TILL

Hypoaspis (Androlaelaps) africanus Zumpt, 1950, Parasitology 40 : 30, fig. 2.

Turkiella africanus, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 216.

Hypoaspis (Androlaelaps) africanoides Zumpt & Patterson, 1950, S. Afr. J. med. Sci, 15 : 72,

figs. 2-3.

Androlaelaps oudemansi Radford, 1944, Parasitology 35 : 161, fig. 2 (syn. nov.).
Turkiella oudemansi, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 224.

FEMALE (figs. 98-100) : Deutosternum with 6 rows of teeth, 3 to 6 small, finely
pointed teeth in each row. Chelae similar to those of A. arvicanthis, 70-90^ long;
pilus dentilis slender; arthrodial filaments subequal in length.

Dorsal shield with a " double " margin and an overall reticular pattern. DL=
863-1050(1,; DW= 675-763^; DL/DW= 1-2-1-5. The shield bears 39 pairs of setae
as well as additional unpaired setae between /3 and 74- Many of the setae are
barbed. Lengths of setae 24 and Z2 slightly exceed the distance between their
bases; seta Z$ only slightly longer than 75.

Sternal shield with a well-defined reticular pattern and with concave anterior and
posterior margins. SL= 88-123^; SW=i8o-i93(x; SL/SW= 0-5-0-6. First sternal
seta reaches second pore, second and third sternal and metasternal setae distinctly
longer.

Genital shield broadly rounded posteriorly, with an overall recticular pattern.
GW 1 =228-250(A; GW 2 =i5o-i75[z; GW 1 /GW 2 =i-4-i-6. Genital setae relatively
long, approximately equal to GW 2 , and extending almost to the base of the second
pair of flanking setae.

Anal shield with a reticular pattern ; anus at slightly less than its length from the
anterior margin; paranal setae situated near posterior half of anus, only slightly
shorter than postanal seta. AL=i23-i66[i,; AW 150-175^; AL/AW=o-8-i-o.

Integument of idiosoma bears 30-55 pairs of long ventral and marginal setae,
most of which are barbed. Metapodal plates oval, length about twice the breadth.

Chaetotaxy of legs normal, except that the type of A . oudemansi and some of the
other specimens assigned to this species have 9 setae on tibia III instead of the usual
eight. Anterior and posterior setae on coxa I approximately equal in length.
Thorn-like seta on femur II relatively large; inner ventral seta on genu II short and
spine-like. Inner ventral seta on tibia II a little shorter and considerably stouter
than the outer seta in the types of A. oudemansi and A. africanoides, whereas in
specimens previously identified as A. africanus the differences between these two
setae are less striking, as in A. arvicanthis. Tarsus II bears 3 stout, blunt, terminal
setae. Length of tarsus IV about 8 times its width at the base.

MALE (fig. 97) : Deutosternum as in female ; chelicera similar to that of A . arvi-
canthis. Dorsal shield similar to that of the female. DL= 890-980^; DW=
570-6o8(j,; DL/DW=i-6. Holoventral shield strongly reticulate, expanded behind
fourth pair of coxae, and bears 23-31 setae. The number of setae between coxae IV
and the anus is 10-19, but there may be more on one side than on the other.
Chaetotaxy of legs as in the female.

HOSTS AND LOCALITIES: Recorded in Southern Africa from the following
hosts (Zumpt & Till, 1961) :

Tatera afra (Gray), Gerbillus paeba Smith, Desmodillus auricularis (Smith), Rattus
paedulcus (Sundevall), Rattus natalensis (Smith), Rattus chrysophilus (De Winton),

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 55

Rattus namaquensis (Smith), Rhabdomys pumilio (Sparrman), Lemniscomys griselda
(Thomas), Saccostomus campestris Peters, Steatomys pratensis Peters, Pedetes capensis
(Forster), Cryptomys hottentotus (Lesson) and Oryctolagus cuniculus (Linnaeus).

Specimens from the following East African hosts have been examined:
Arvicanthis niloticus (Desmarest), Bunyoro, Uganda (type of A. oudemansi in
B.M.N.H.).

FIGS. 101-103. Androlaelaps mesopicos (Radford), female. Venter (Fig. 101); dorsum
(Fig. 102) ; tip of tarsus IV, ventral view (Fig. 103).

Tatera valida (Bocage), Bunyoro, Uganda (S.A.I.M.R.).
Tatera nigricauda Peters, Lali Hills, Kenya (S.A.I.M.R.).
Tatera species, W. Nile District, Uganda (S.A.I.M.R.).
Taterillus emini (Thomas), W. Nile District, Uganda (S.A.I.M.R.).
Tachyoryctes splendens (Riippell), Muguga, Kenya (S.A.I. M.R.).
Unidentified rodent, Kerio Valley, Kenya (S.A.I.M.R.).

5 6 W. M. TILL

A. marshalli has also been recorded from an unidentified rodent in the Sudan
and from gerbils in Egypt (Keegan, 1956) and from gerbils in Israel (Costa, 1961).

Two series of mites which have been provisionally assigned to A. marshalli have
been obtained from carnivores, namely, Suricata suricatta (Schreber) from Queens-
town, Cape Province (S.A.I.M.R.), and Ictonyx striatus (Perry) from Muguga, Kenya
(S.A.I.M.R.).

The drawings of A . marshalli are based on paratypes of A . africanoides from Tatera
afra (Gray), S. Rhodesia.

Androlaelaps mesopicos (Kadlord)

Haemolaelaps mesopicos Radford, 1942, Parasitology 34 : 192, figs. 14-19; Till, 1959, /. ent. Soc.

S. Afr. 22 : 423, figs.
Hypoaspis (Haemolaelaps) mesopicos, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 72.

A series of mites from Mesopicos griseocephalus , Natal, has been assigned to this
species (Till, 1959). Although differing slightly from the types, especially in the
male sex, these mites probably represent no more than a form or variety of A.
mesopicos.

FEMALE (figs. 101-103) : Deutosternum provided with 6 rows of 2-4 small teeth.
Chelae about 40^ long; pilus dentilis small, inflated; arthrodial filaments subequal
in length.

Dorsal shield oval in shape, truncate posteriorly, with a reticulate pattern.
DL= 700-735 [x; DW= 420-438^; DL/DW= 1-6-1-7. Not more than 37 pairs of
setae can be distinguished on the shield, seta 53 being absent and seta r6 arising
from the integument adjacent to the shield. Seta p%3 is present in the type but
absent in the specimens from Natal. The marginal setae are variable in size, the
longest of the S setae being more than 3 times as long as the / setae. Seta Z$ is
sinuous and whip-like and about 20 times as long as seta /5. The setae on the
median part of the shield (i2-i$, J series, z series, Zi-Z4, s2-s6) are very minute,
subequal, the lengths of setae 24 and Z2 being less than a quarter the distance
between their bases.

Sternal shield granular, presternal area weakly reticulate. SL=84-io5(x;
SW= 152-158^; SL/SW=o-6-o-7. First pair of sternal setae slightly shorter than
second and third pairs, and lying on the presternal area. Metasternal setae as
long as the 2nd and 3rd sternal setae.

Genital shield slender, weakly reticulate posteriorly; GWj 96-144^; GW 2 =
80-96^; GW!/GW 2 = 1-1-1-3. Genital setae relatively short, less than half GW 2 .

Anal shield pear-shaped with a reticular pattern antero-laterally, relatively
shorter and broader in the specimens from Natal. AL= 123-150^; AW= 96-105^;
AL/AW=i-2-i-6. Paranal setae level approximately with middle of anus, not
more than 2/3 as long as postanal seta.

Integument of idiosoma bears about 23 pairs of marginal and ventral setae, which
all appear to be simple, except for the long, sinuous, posterior terminal pair.
Metapodal plates slender, slightly curved.

Tibia HI bears 9 setae and tarsi II, III and IV each bear two stout, blunt,

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 57

terminal spines. The anterior and posterior setae on coxa I are approximately
equal in length and thickness. Length of tarsus IV about 4 times the width at
the base.

MALE (figs. 104-106): Fixed digit of chelicera with a small setiform structure
near its base. Dorsal shield reticulate, and also granular in the specimen from
Natal ; shape and chaetotaxy as in the female, the marginal setae, however, being

104 105

FIGS. 104-106. Androlaelaps mesopicos (Radford), male. Venter of type specimen from
Uganda (Fig. 104); venter (Fig. 105) and chelicera (Fig. 106) of specimen from Natal.

considerably longer. In the type, setae 82 and 5$, and in the Natal specimen 84
also, are long and sinuous. DL=6i8-646[ji; DW=35o-36o[i.; DL/DW=i -6-1-7.
Holoventral shield bears 19 setae. Sternal region similar in shape to sternal shield
of female, with a granular appearance and faint reticular markings ; reticular pattern
more distinct on remaining part of shield. Posterior to coxae IV the shield is narrow
with irregular margins.

Integument of idiosoma bears about 16 pairs of marginal and ventral setae, the

58 W. M. TILL

posterior pair being long and sinuous as in the female. In the specimen from Natal
the longest marginal setae are approximately twice as long as in the type specimen.
Metapodal plates as in the female.

The legs bear the same number of setae as in the female, but there are modifications
in the form of some of the setae. On femur II, the lower medio-ventral seta is
modified to form a thickened spur. Tarsi II, III and IV each bear one stout,
terminal spine, and tarsus II has, in addition, two blunt ventral spurs.

107

FIGS. 107-109. Androlaelaps murinus (Berlese), female. Venter (Fig. 107); dorsum

(Fig. 1 08); chelicera (Fig. 109).

HOSTS AND LOCALITIES: Mesopicos griseocephalus ruwenzori Sharpe, Kabale,
Uganda (types in B.M.N.H.).

Mesopicos g. griseocephalus (Boddaert), Pietermaritzburg, Natal (S.A.I.M.R. and
B.M.N.H.).

Androlaelaps murinus (Berlese)

Laelaps (Haemolaelaps] murinus Berlese, 1911, Redia 7 : 432.

Haemolaelaps murinus, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 233, figs. 16-17.

Ischnolaelaps arvicanthis Radford, 1939, Parasitology 31 : 246, fig. 3.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat.

FEMALE (figs. 107-109) : Deutosternum provided with 6 rows of 3-4 teeth. Chelae
52-54(1 long, movable digit with a pronounced terminal hook which arches over the
fixed digit ; pilus dentilis strongly inflated basally, terminal portion slender, curved ;
one of the arthrodial filaments strikingly longer than the others.

Dorsal shield ovoid; DL= 860-912(1; DW= 595-648(1; DL/DW=i-4-i-5. The
shield bears 39 pairs of setae and 2-3 unpaired accessory setae. Lengths of setae

110

112

FIGS. 110-113. Androlaelaps murinus (Berlese), venter (Fig. no) and chelicera
(Fig. in) of male.

Androlaelaps oliffi (Zumpt & Patterson), venter (Fig. 112) and chelicera (Fig. 113) of
male.

24 and Z2 a little less than the distance between their bases; seta Z$ about twice
as long as seta 75.

Sternal shield strongly reticulate; SL= 120-140(1; SW= 166-184(1; SL/SW=
07-0-8. Metasternal setae shorter than first pair of sternal setae.

Genital shield slightly widened posteriorly and having a reticulate pattern which
extends over the anterior flap. GW^ 184-210(1; GW 2 = 128-140(1; GW l /GW 2 =
1-4-1-5. Length of genital setae about half GW 2 .

Anal shield approximately as long as broad; AL= 145-150(1; AW= 145-16051.

60 W. M. TILL

Anus a little less than its length from the anterior margin; paranal setae inserted
slightly behind middle of anus, a little shorter than the postanal seta.

Integument of idiosoma bears about 60 pairs of setae of which about 15 pairs are
truly ventral in position. The marginal setae tend to be stouter than the ventral
ones and have fine barbs.

Chaetotaxy of legs normal. Anterior seta on coxa I 3/4 as long as posterior seta ;
dorsal seta on trochanter IV stout and spine-like ; two of the dorsal setae on femur IV
spine-like. Length of tarsus IV about 7 times its width at the base.

MALE (figs, iio-iu): The chelicerae have the structure indicated in the figure.
Holoventral shield reticulate, expanded behind coxae IV, bearing 31-32 setae.
Integument of idiosoma bears numerous setae. Chaetotaxy of dorsal shield and
legs as in the female. DL=665[x; DW=437[x; DL/DW 1-5.

HOSTS AND LOCALITIES: Arvicanthis niloticus (Desmarest) from Uganda (type of
/. arvicanthis in B.M.N.H.) and from Kenya (Keegan, 1956).
Otomys species, Uganda (B.M.N.H.).

Otomys irroratus (Brants) and Rattus natalensis (Smith) from Southern Rhodesia
(S.A.I.M.R.).

Oenomys hypoxanthus (Pucheran) from the Belgian Congo (B.M.N.H.) and from the
Cameroons (I.R.S.C.).

Dasymys incomtus (Sundevall) from Kampala, Uganda (S.A.I.M.R.) and from the
Rift Valley Province, Kenya (Keegan, 1956).

Rhabdomys pumilio (Sparrman), Lophuromys flavopunctatus Thomas and Lemnis-
comys striatus (Linnaeus) from the Rift Valley Province, Kenya (Keegan, 1956).

Androlaelaps oliffi (Zumpt & Patterson)

Hypoaspis (Haemolaelaps) oliffi Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 81, fig. 3.
Haemolaelaps oliffi, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 243.

FEMALE: (figs. 114-116) : Deutosternum with 6 rows of 2-4 teeth. Chela 38-40^
long; pilus dentilis slightly inflated; arthrodial filaments subequal in length.

Dorsal shield ovoid, reticulate. DL= 620-644^; DW=368-4iO(x; DL/DW=
I '5~ 1 '7- The shield bears the usual 39 pairs of setae; lengths of setae 24 and Z2
slightly exceed the distance between their bases; seta Z$ is if times as long as
seta 75.

Sternal shield reticulate; SL=96-io5fx; SW= 130-1 35^; SL/SW=o-7-o-8.
Metasternal setae slightly shorter than first pair of sternal setae.

Genital shield granular and with a reticulate pattern which extends over the
anterior flap. The shield is not expanded behind the genital setae, GWj and GW 2
both varying between 127^ and i35[x. Length of genital setae about 2/3 GW 2 .

Anal shield approximately as long as broad, or slightly longer; AL= 123^1 ;
AW=ii4-n8(jL. Anus situated at about i| times its length from the anterior
margin. Paranal setae level approximately with middle of anus, almost reaching
base of postanal seta. Postanal seta about the same length as the paranal setae,
but slightly thicker.

Integument of idiosoma bears 15 pairs of setae, of which 3 pairs border the

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 61

genital shield; some marginal setae are barbed. Metapodal plates slender, length to
width about 7:1.

Chaetotaxy of legs normal except that tibia III bears 9 setae. Posterior seta
on coxa I nearly twice as long and thick as anterior seta. Anterior seta on coxa II
and both setae on coxa III rather stout. Three of the apical setae on tarsus II
are stout and one of these is blunt. Length of tarsus IV about 5| times the width
at the base.

I.' V! I'./ M
'U

FIGS. 114-116. Androlaelaps oliffi (Zumpt & Patterson), female,
dorsum (Fig. 115); chelicera (Fig. 116).

115

Venter (Fig. 114);

MALE (figs. 112-113): Chelicera as figured; dorsal shield reticulate, chaetotaxy
as in the female ; DL = 428-437/4 ; DW = 276/4 ; DL/DW = 1-6. Holoventral
shield reticulate, bearing 23 setae ; these, excluding the anals, are longer and stouter
than those of the integument. Integument bears about 16 pairs of ventral and
marginal setae ; metapodal plates elongate. Chaetotaxy of legs as in the female.

HOSTS AND LOCALITIES : Widely distributed in Southern Africa, mainly on Tatera
afra (Gray) and Gerbillus paeba Smith, and occasionally found on Saccostomus
campestris Peters and Desmodillus auricularis (Smith) (S.A.I.M.R., B.M.N.H. and
M.R.A.C.). Found also in Tanganyika on Tatera afra (Gray) (B.M.N.H.).

62 W. M. TILL

Androlaelaps pachyptilae (Zumpt & Till)

Haemolaelaps pachyptilae Zumpt & Till, 1956, Z. Parasitenk. 17 : 285, fig. 3.

FEMALE (figs. 117-119): Deutosternum with 6 rows each of 3-4 small teeth.
Chelae 40[x long; pilus dentilis inflated basally, distal half slender and curved;
arthrodial filaments subequal in length.

Dorsal shield ovoid, reticulate, widest between setae rj and 57. DL= 604-691^;

117

FIGS. 117-119. Androlaelaps pachyptilae (Zumpt & Till), female,
dorsum (Fig. 118); chelicera (Fig. 119).

118

Venter (Fig. 117);

DW= 385-446^; DL/DW 1-5-1-6. The shield bears 39 pairs of setae. Those on
the central part of the shield are relatively short, the lengths of setae 24 and zz
being about half the distance between their bases. The anterior and marginal
setae are considerably longer; seta Z$ is barbed and is about 2\ times as long as
seta /5.

Sternal shield with a reticulate pattern, which is more pronounced antero-
laterally, the central part of the shield having a granular appearance. Presternal

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 63

area strongly reticulate. SL=79-ioo[x; SW= 140-158^; SL/SW=o-5-07. First
sternal seta nearly i| times as long as metasternal seta.

Genital shield short, slightly widened behind genital setae and then tapering
posteriorly; a reticulate pattern is present posteriorly but is not very sharply
defined. GW 1 =ioo-ii4 f x; GW 2 =g2-g6yi; GWJGWg^ 1-1-1-2. Genital setae the
same length as the ventral setae which flank the shield, about half GW 2 .

Anal shield slightly wider than long. AL= 105-1 14^; AW= 114-123^; AL/AW=
0-9. Anus situated at less than its length from the anterior margin. Paranal
setae level with middle of anus, not more than 2/3 as long as postanal seta. Length
of postanal seta almost equal to its distance from the anterior margin of the shield.

Integument of idiosoma bears 16 pairs of ventral and marginal setae, two pairs
bordering the genital shield and a third pair lying between the genital and anal
shields. The marginal setae are mostly barbed, the ventral ones simple. Metapodal
plates elongate, length to width about 4:1.

Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer than anterior
seta. Length of tarsus IV about 5 times the width at the base.

MALE: Not known.

HOST AND LOCALITY: Pachyptila desolata (Gmelin) from Heard Island (type
series in S.A.I. M.R.).

Androlaelaps pater soni (Zumpt & Till)

Haemolaelaps patersoni Zumpt & Till, 1956, Z. Parasitenk. 17 : 286, fig. 4.

FEMALE (figs. 120-122) : Deutosternum with 6 rows of 3-6 very fine teeth. Chelae
42-46(0, long; pilus dentilis spatulate, terminating in a short, fine, slightly bent
point ; arthrodial filaments subequal in length.

Dorsal shield bears 39 pairs of setae and a variable number (5-12) of unpaired
accessory setae which may extend anterior to setae 25. Lengths of setae 24 and 22
approximately equal to the distance between their bases; seta Z$ at least twice as
long as seta 75. DL= 683-75 3^; DW =420-481^; DL/DW-i-5-1-6.

Sternal shield reticulate; SL= 100-114^; SW= 130-149^; SL/SW= 0-7-0-8.
Metasternal setae a little shorter than first pair of sternal setae.

Genital shield moderately expanded posteriorly, with a reticulate pattern ex-
tending over the anterior flap. GW X = 135-154^; GW 2 = 96-114^; GWJGW^
I '3~ 1 '5- Length of genital setae at least half GW 2 .

Anal shield with an overall reticulate pattern; AL= 114-131^; AW 127-131^;
AL/AW=o-9~i-o. Anus situated at approximately its length from the anterior
margin of the shield. Paranal setae level with middle of anus, about 4/5 as long as
postanal seta.

Integument bears about 30 pairs of setae. Outer metapodal plates elongate,
length to width 4-5 : i. Chaetotaxy of legs normal. Anterior and posterior
setae on coxa I subequal in length and thickness. Length of tarsus IV about 7 times
the width at the base.

MALE (figs. 123-124) : The chelicerae have the structure shown in the figure, the
fixed digit bearing a pilus dentilis similar to that of the female.

6 4

W. M. TILL

124

122

123

FIGS. 120-124. A ndrolaelaps patersoni (Zumpt & Till). Venter (Fig. 120), dorsum (Fig.
121) and chelicera (Fig. 122) of female. Venter (Fig. 123) and chelicera (Fig. 124) of
male.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 65

Chaetotaxy of dorsal shield as in the female, but the setae are relatively longer.
DL= 569-586^; DW=324-359(jt; DL/DW=i5. Holoventral shield reticulate,
expanded behind coxae IV and bearing 23 setae. The integument bears about
15 pairs of setae. Chaetotaxy of legs as in the female.

HOST AND LOCALITY: Gyps coprotheres (Forster) from Rustenburg, Transvaal
(type and other material, S.A.I.M.R.).

125

126

FIGS. 125-127. Androlaelaps phoeniculi (Zumpt & Till), female. Venter (Fig. 125);
dorsum (Fig. 126); chelicera (Fig. 127).

Androlaelaps phoeniculi (Zumpt & Till)

Haemolaelaps phoeniculi Zumpt & Till, 1954; R ev - Ecuat. Ent. Parasit. 2 : 209, figs. 1-2.

FEMALE (figs. 125-127) : Deutosternum with 6 rows of 1-3 teeth. Chelae 30-32(0,
long; pilus dentilis slender; arthrodial filaments subequal in length.

Dorsal shield reticulate, truncate posteriorly, bearing 37 pairs of setae, two pairs
being absent from the anterior part of the shield. Lengths of setae 24 and Z2 do
not exceed | the distance between their bases ; the setae at the anterior end of the
shield are longer, ii being at least twice as long as 24. Posterior terminal seta Z5
barbed, 5-6 times as long as seta /5. DL=5i7-579[j.; DW= 290-340^; DL/DW=
i -6-1-8. "

ZOOL. IO, I

66 W. M. TILL

Sternal shield similar to that of the mesopicos-group, but relatively broader.
SL=62~72[A; SW= 134-156(0.; SL/SW= 0-4-0-5. Anterior and lateral parts of the
shield granular and reticulate, remaining part of shield apparently smooth ; presternal
area reticulate. First pair of sternal setae and metasternal setae subequal in
length, second and third pairs slightly longer.

Genital shield extends almost to the anal shield and is apparently without any
distinct pattern. It is slightly widened behind the genital setae, with its lateral
margins subparallel and its posterior border almost straight. GW 1 =ioo-ii4(j.;
GW z =88-g6[L', GW 1 /GW 2 =i-o-i-3. Genital setae as long as first pair of sternal
setae, about 1/3 GW 2 .

Anal shield distinctly longer than broad, anterior margin straight or slightly
convex. AL=93-ii4[jt.; AW=72-83[z; AL/AW= 1-1-1-4. Anus situated at less
than half its length from the anterior margin of the shield. Postanal seta about
1 1 times as long as paranal setae, the latter being level approximately with the middle
of the anus.

Integument of idiosoma bears about 29 pairs of ventral and marginal setae; the
ventral setae are short, fine and simple, the marginal ones coarser and mostly
barbed. Metapodal plates elongate, slightly curved, length to width about 8 : i.

Tibia III with 9 setae ; anterior seta on coxa I at most i| times as long as posterior
seta. The leg setae are simple and there are no stout, blunt spurs. Legs short and
stumpy; length of tarsus IV about 4 times the width at the base.

MALE (not figured) : Length 530^, width 340^. Holoventral shield narrow behind
coxae IV, with a constriction anterior to the anal region. It bears 15 setae, the
first pair lying on the presternal area as in the female. Metapodal plates drop-
shaped.

HOST AND LOCALITIES: Phoeniculus purpureus Miller from Potchefstroom,
Transvaal (type series), from Tsessebe, Bechuanaland, and from Northern Rhodesia
(S.A.I.M.R.).

Androlaelaps rhodesiensis (Zumpt & Patterson)

Hypoaspis (Haemolaelaps} rhodesiensis Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 82,

fig. 4.
Haemolaelaps rhodesiensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 243.

FEMALE (figs. 128-130) : Deutosternum provided with 6 rows of 3-5 small teeth.
Chelae 62-70^ long ; pilus dentilis slightly inflated, ribbon-like ; arthrodial filaments
subequal in length.

Dorsal shield broadly oval, reticulate, widest about level of seta Si. DL=
825-912^; DW= 586-700^; DL/DW=i-3-i-4. The shield bears the usual 39 pairs
of setae, as well as two accessory setae at the level of 74. The setae are all very
long, some having fine barbs; lengths of setae 24 and Z2 exceed twice the distance
between their bases; seta /5 is about 4/5 as long as seta Z$.

Sternal shield reticulate; SL=i 14-130^; SW= 170-193^; SL/SW= 0-6-0-7.
Metasternal seta about as long as first sternal seta, second and third sternal setae
slightly longer than the first.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 67

Genital shield with a reticulate pattern extending to the anterior flap, greatly
expanded behind the fourth pair of coxae. GWj= 254-300^; GW 2 = 130-140^;
GWj/GW 2 = 2-0-2-2. Genital setae relatively short, their length about 2/3 GW 2 .

Anal shield triangular, broader than long. AL= 150-160^; AW= 170-184^;
AL/AW= 0-8-0-9. Anus slightly more than its length from the anterior margin.
Paranal setae inserted slightly behind middle of anus, 2/3 to 4/5 as long as the
postanal seta.

128

FIGS. 128-130.

Androlaelaps rhodesiensis (Zumpt & Patterson), female. Venter
(Fig. 128); dorsum (Fig. 129); chelicera (Fig. 130).

Integument of idiosoma bears 15 pairs of long ventral and marginal setae, of which
3 pairs flank the genital shield; most of the setae have fine barbs. Metapodal
plates roughly oval, length about twice the width.

Tibia I has 14 and tibia III has 9 setae. Anterior seta on coxa I about 4/5 as

W. M. TILL

long as posterior seta. Most of the leg setae are barbed; setae on some segments
stouter than on others, but there are no blunt spines. Length of tarsus IV about
7 times the width at the base.

MALE: Not known.

HOSTS AND LOCALITIES: Saccostomus campestris Peters from several localities in
Southern Rhodesia (S.A.I. M.R.).

Rattus rattus (Linnaeus) and Rattus morio (Trouessart) from Brazzaville, French
Congo (Zumpt & Till, 1961).

Steatomys pratensis Peters from Southern Rhodesia (one female only, Zumpt &
Till, 1961).

131 132

FIGS. 131-132. Androlaelaps sangsteri (Radford) female. Venter (Fig. 131); dorsum (Fig. 132).

Androlaelaps sangsteri (Radford)

Liponyssus sangsteri Radford, 1942, Parasitology 34 : 190, fig. n; Zumpt & Till, 1953, 5. Afr.

J. Med. Sci. 18 : 7.
Haemolaelaps sangsteri, Strandtmann & Wharton, 1958, Manual of Mesostigmatid Mites : 44.

FEMALE: (figs. 131-132) : Deutosternum provided with 6 rows of teeth, 4 teeth in
the first row, 2 in the second and one in each of the remaining rows. Corniculi
reduced, comparatively small and inconspicuous. Chelicerae missing in the type,
the only available specimen.

Dorsal shield narrow with lateral margins subparallel, granular in appearance,
with faint reticulations. DL=5i8pi; DW=257pi; DL/DW=2-o. The number of

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 69

setae is reduced, only 34 pairs being distinguishable with certainty. Those on the
central part of the shield are short, the lengths of setae 24 and 22 not exceeding half
the distance between their bases. The anterior and marginal setae are longer;
seta Z5 is at least 4 times as long as seta /5 and has fine barbs.

Sternal shield granular, anterior margin slightly convex, posterior margin straight,
SL=96(ji; SW=i32(x; SL/SW=o7. Sternal setae subequal in length, relatively
short, not reaching bases of following setae. Metasternal setae about as long as
sternal setae.

FIGS. 133-135. Androlaelaps scapularis (Berlese), female. Venter (Fig. 133); dorsum

(Fig. 134); chelicera (Fig. 135).

Genital shield granular, with an indistinct reticulate pattern; widest near its
posterior border, between the first and second pairs of flanking setae ; anterior part
of shield rather sharply pointed. GW 1 =ioopi; GW 2 =72[x; GWJGWg 1-4.
Length of genital setae about half GW 2 .

Anal shield triangular, approximately as long as broad. AL=82[j.; AW=79jji;
Parana! setae level with middle of anus, about 4/5 as long as postanal seta.

Integument of idiosoma bears 26 pairs of ventral and marginal setae, the longer
ones near the posterior end of the body being barbed. Metapodal plates not
visible.

Legs short and stumpy, their chaetotaxy following the typical Androlaelaps

yo W. M. TILL

pattern ; all the setae simple and hair-like. Length of tarsus IV nearly 4 times the
width at the base.

MALE : Not known.

HOST AND LOCALITY : Protoxerus stangeri (Waterhouse) from Bwamba Valley,
Toro, Uganda (type in B.M.N.H.).

Androlaelaps scapularis (Berlese)

Hypoaspis (Haemolaelaps) scapularis Berlese, 1918, Redia 13 : 124; Keegan, i956a, Trans.

Amer. micr. Soc. 75 : 314, figs. 13-14.

1 'schnolaelaps bathyergus Radford, 1939, Parasitology 31 : 250, figs. 7-8 (syn. nov.).
Haemolaelaps eloffi Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 226, figs. 6-9 (syn. nov.}.
Haemolaelaps natalensis Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 230, figs. 10-11

(syn. nov.}.

FEMALE (figs. 133-135) : Deutosternum provided with 6 rows of 2-4 teeth, the
teeth in rows 5 and 6 being very small and inconspicuous ; anterior rostral seta at
least 1 1 times as long as capitular seta. Chelae very small, 18-2 2 fz long; fixed
digit weakly sclerotized, bearing a slender pilus dentilis ; movable digit in two parts,
one part edentate except for the terminal hook and a small subterminal tooth, the
other part dentate but more weakly sclerotized. Arthrodial filaments subequal in
length.

Dorsal shield broadly oval, granular, with a faint reticulate pattern. DL=445-
570(x; DW=28o-390[x; DL/DW=i-3-i-6. The shield bears 39 pairs of setae.
The centrally situated ones are very small, the lengths of 24 and 22 being less than
1/5 the distance between their bases; the anterior and marginal setae are longer;
seta Z5 is at least 2.\ times as long as seta 75.

Sternal shield granular, presternal area reticulate. SL=^i-82[i; SW= 103-133^;
SL/SW= 0-4-07. Sternal setae subequal in length, first pair slightly shorter than
the others ; metasternal setae about half as long as third pair of sternal setae.

Genital shield flask-shaped, moderately widened behind the genital setae, granular
in appearance, with a faint reticulate pattern. GW^ i 03-162^; GW 2 = 75-105 pi;
GWj/GW2= 1-2-1-5. Genital setae relatively short, not more than half GW 2 .

Anal shield triangular, distinctly wider than long. AL=52-94fx; AW=83-ii4(j.;
AL/AW=o-5~o-7. Paranal setae level approximately with middle of anus, or
slightly anterior, about i times as long as postanal seta.

Integument of idiosoma bears about 20 pairs of simple ventral and marginal
setae, of which 3 pairs border the genital shield. Metapodal plates elongate, length
to width about 6 : i.

Chaetotaxy of legs normal. Posterior seta on coxa I is i|-i| times as long as
the anterior seta. Anterior seta on coxa III about twice as long and thick as
posterior seta. Length of tarsus IV about 4 times the width at the base.

MALE (fig. 136) : Fixed digit of chelicera with a slender pilus dentilis as in the
female. Dorsal shield as in the female except that the centrally situated setae are
relatively longer. Holoventral shield expanded behind coxae IV, bearing 23 setae,
5 pairs lying in the region between coxae IV and the anus. Integument of iosoma
with about 13 pairs of setae. Chaetotaxy of legs as in female.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE 5. lot. 71

HOSTS AND LOCALITIES: Bathyergus suillus (Schreber) from Strandtfontein, Cape
Town (types in Dr. C. D. Radford's collection and in M.R.A.C.), and from Tulbagh
District, Western Cape Province (S.A.I.M.R.).

Cryptomys hottentotus (Lesson) from Bloemfontein, Orange Free State, and
Pietermaritzburg, Natal (S.A.I.M.R. and B.M.N.H.).

136

FIGS. 136-138. Androlaelaps scapularis (Berlese), venter of male (Fig. 136).

Androlaelaps spatuliformis (Lavoipierre), venter (Fig. 137) and chelicera (Fig. 138) of
male.

Cryptomys holosericeus (Wagner) from Pietermaritzburg, Natal (S.A.I.M.R.).
NOTE: The specimens from Bathyergus suillus are larger than those from the
Cryptomys species, but there are no reliable features which could be used to separate
these groups. The types of /. bathyergus, H. eloffi and H. natalensis have been
compared with drawings of the type of H. scapularis Berlese made by Dr. F.
Gorirossi (Keegan, 1956) and it is believed that they are all conspecific. The deep
posterior concavity of the sternal shield, indicated by Zumpt & Till (1953) in their
drawing of H. eloffi, appears to be the result of over-clearing.

72 W. M. TILL

Androlaelaps spatuliformis (Lavoipierre)
Haemolaelaps spatuliformis Lavoipierre, 1956, Ann. trop. Med. Parasit. 50 : 293, fig. 2.

FEMALE (figs. 139-140) : Deutosternum provided with at least 4 rows of 2-3 small
teeth, 5th and 6th rows very indistinct. Chelae about 40^ long; pilus dentilis and
arthrodial filaments not clearly visible in the one specimen available for examination.

139

140

FIGS. 139-140.

Androlaelaps spatuliformis (Lavoipierre), female,
dorsum (Fig. 140).

Venter (Fig. 139);

Dorsal shield oval; DL=525(x; DW=36o(j.; DL/DW=i-4- Several of the setae
are missing from the specimen examined, but the chaetotaxy appears to follow the
38- or 39-pair pattern. Seta px$ cannot be distinguished in this specimen. Seta
Z5 is more than twice as long as seta /5.

Sternal shield has a few very indistinct reticulations; presternal area weakly

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 73

reticulate. Sternal and metasternal setae subequal in length; first pair of sternal
setae inserted on presternal area. SL= 94-96^; SW= 140-1 43^; SL/SW=O7.

Genital shield spatulate in shape, very slightly widened behind genital setae,
then tapering gradually; GW 1 =ioo[j.; GW 2 =97;j.. Genital setae relatively short,
about i GW 2 .

Anal shield slightly longer than broad; AL=ioo-ii2[x; AW=9i^.. Paranal setae
inserted near posterior margin of anus, about f as long as postanal seta.

Integument of idiosoma bears about 19 pairs of ventral and marginal setae.
Metapodal plates provided with a pronounced medial protrusion.

Chaetotaxy of legs normal except for genu IV which has only 9 instead of 10 setae.
Posterior seta on coxa I about i| times as long as anterior seta.

MALE (figs. 137-138) : Structure of chelicera indicated in fig. 138 ; pilus dentilis not
visible. Dorsal shield as in female; seta px$ present. DL= 404-409^; DW=266^;
DL/DW=i-5. Holoventral shield expanded behind coxae IV, bearing 21 setae,
excluding the first pair of sternal setae which is situated on the presternal area.
Chaetotaxy of legs as in the female. Anterior and posterior setae of coxae I
approximately equal in length and thickness. Femur II has one stout ventral
spine and tarsus II has two stout ventral spines.

HOST AND LOCALITY : Anomaluris derbianus Gray from Kumba, British Cameroons.
One female paratype received on loan from Dr. M. M. J. Lavoipierre, and two males
from the type host and locality received from the Liverpool School of Tropical
Medicine, by courtesy of Miss M. Johnson.

Androlaelaps spreo (Zumpt & Till)

Haemolaelaps spreo Zumpt & Till, 1956, Z. Parasitenk. 17 : 290, fig. 9.

FEMALE (figs. 141-143): Androlaelaps spreo is very closely allied to A. dasymys,
differing from it mainly in the structure of the chelicerae. The pilus dentilis is
different in shape and the arthrodial filaments are subequal in length. The chelae
are 54~56pL long.

The dorsal shield bears 39 pairs of setae as well as 4-6 accessory setae between the
/ series. All the setae are relatively long, the lengths of setae 24 and Z2 being at
least 1-5 times the distance between their bases. The marginal setae are slightly
longer and thicker than those on the median part of the shield and are barbed.
Seta Z5 is 17-1-8 times as long as seta /5. DL=849~936[x; DW=525~6o4fz;
DL/DW=i-6.

Sternal shield reticulate; SL= 114-140^; SW=i66-i75[A; SL/SW=07~o-8.
Genital shield slightly widened posteriorly, with a pronounced reticular pattern
extending to the anterior margin. GW^ 158-175^; GW 2 = 114-1 23^; GWJ
GW 2 =i-3-i-4. Genital setae relatively long, slightly more than 2/3 GW 2 .

Anal shield approximately as long as broad and with a reticular pattern; AL=
140-162^ ; AW= 140-158(4. Anus situated at slightly more than its length from the
anterior margin of the shield; paranal setae about 4/5 as long as postanal seta.

Integument bears about 33 pairs of ventral and marginal setae, the latter being

W. M. TILL

longer and thicker than the former and provided with barbs. Metapodal plates
oval, 1-5-1-8 times as long as broad.

Chaetotaxy of legs normal. Posterior seta on coxa I slightly longer and thicker
than anterior seta. Tarsus II has three stout ventral bristles but no blunt spines.
Trochanter IV has a stout dorsal bristle. Length of tarsus IV 6|~7| times width
at base.

141

142

FIGS. 141-143. Androlaelaps spreo (Zumpt & Till), female. Venter (Fig. 141); dorsum

(Fig. 142); chelicera (Fig. 143).

MALE (figs. 144-145) : Chelae as illustrated, pilus dentilis resembling that of female.
Dorsal shield bears 8 accessory setae, otherwise chaetotaxy as in female. DL=
675^; DW=447[A; DL/DW=i-5. Holoventral shield reticulate, expanded, bearing
23 setae. Chaetotaxy of legs as in female.

HOSTS AND LOCALITIES: Spreo bicolor (Gmelin), Springs, Transvaal (type series,

S.A.I.M.R.).

Riparia paludicola (Vieillot), Johannesburg, Transvaal (S.A.I.M.R.).

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 75

Androlaelaps steyni (Till)

Haemolaelaps steyni Till, 1959, /. ent. Soc. S. Afr. 22 : 430, figs.

FEMALE (figs. 148-150) : This species is very closely allied to A. mesopicos. Dorsal
shield relatively shorter and broader than in A. mesopicos and bearing 38 pairs of
setae. DL= 578-648^; DW= 368-420^; DL/DW=i-5-i-6.

Sternal shield relatively shorter than in A. mesopicos. SL= 53-79(1 ; SW=

144

147

146

FIGS. 144-147. Androlaelaps spreo (Zumpt & Till), venter (Fig. 144) and chelicera
(Fig. 145) of male.
Androlaelaps steyni (Till), venter (Fig. 146) and chelicera (Fig. 147) of male.

149-166(1; SL/SW=o-3-o-5. Genital shield similar to that of A. mesopicos. Anal
shield triangular, with a straight anterior margin; AL= 96-105(0.; AW 105-114^;
AL/ AW = 0-9-1-0. Postanal seta more than twice as long as paranal setae.

Coxa III has a greatly thickened anterior spine; tarsus IV has only one stout
terminal spine. Length of tarsus IV about 5 times width at base.

MALE (figs. 146-147) : Chelicera as figured, fixed digit bearing a relatively short,
moderately inflated pilus dentilis. Dorsal shield similar to that of female, but

7 6

W. M. TILL

with longer marginal setae. DL= 530-534^; DW=3i5-333[j.; DL/DW= 1-6-1-7.
Coxa III has a stout anterior spine as in the female.

HOSTS AND LOCALITIES: Lybius torquatus (Dumont), Mabelikwa, Transvaal (type
series, S.A.I.M.R. and B.M.N.H.).

Trachyphonus vaillanti Ranzani, near Johannesburg, Transvaal (S.A.I.M.R.).
Lamprotornis nitens (Linnaeus), Rustenburg, Transvaal (S.A.I.M.R.).

FIGS. 148-150. Androlaelaps steyni (Till), female. Venter (Fig. 148) ; dorsum (Fig. 149) ;
ventral view of tip of tarsus IV (Fig. 150).

Androlaelaps suncus sp. nov.

FEMALE (figs. 151-153) : Deutosternum provided with 6 rows of 2-5 teeth. Chelae
about 30[j, long ; movable digit with a pronounced terminal hook which arches over
the tip of the fixed digit ; pilus dentilis inflated, hour-glass-shaped in some positions ;
arthrodial filaments subequal in length.

Dorsal shield oval, reticulate, widest about the level of seta Si. DL= 608-694^;

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 77

DW=385-456{x; DL/DW= 1-5-17. The shield bears the usual 39 pairs of setae
and one accessory seta. Lengths of setae 14 and 22 about 2/3 the distance between
their bases ; seta Z$ nearly twice as long as seta 75. Setae Z$ and 53-55 are barbed,
the others appear to be simple.
Sternal shield reticulate, except in the postero-median third. SL = 90-105^ ;

151 152

FIGS. 151-153. Androlaelaps suncus sp. nov., female. Venter (Fig. 151); dorsum

(Fig. 152); chelicera (Fig. 153).

= 128-148^; SL/SW=o-7. Metasternal setae and first pair of sternal setae
subequal in length, 2nd and 3rd pairs of sternal setae slightly longer.

Genital shield slightly widened behind the genital setae, broadly rounded
posteriorly, with a fairly well-marked, overall, reticular pattern. GW^ 105-120^ ;
GW 2=84-10311; GWj/GW^i-2. Length of genital setae about 3/5 GW 2 .

Anal shield approximately as long as broad; AL 103-120^; AW= 106-115^.
Anus slightly more than its length from the anterior margin. Paranal setae level
approximately with middle of anus, about 5/6 as long as postanal seta.

78 W. M. TILL

Integument of idiosoma bears about 23 pairs of ventral and marginal setae, the
latter being barbed. Metapodal plates elongate, length to width about 4:1.

Chaetotaxy of legs normal. Posterior seta on coxa I is i| times as long as anterior
seta. Length of tarsus IV about 5 times its width at the base.

MALE: Not known.

HOSTS AND LOCALITIES: Suncus varius (Smuts) from Tibedi, Basutoland,
9 October, 1953 (4 females).

Crocidura sp. from Pilgrim's Rest, Transvaal, 14 July 1951 (one female not included
in the type series).

Holotype and one female paratype in the collection of the South African
Institute for Medical Research, two female paratypes (1962.6.12.18-19) in the
collection of the British Museum (Natural History).

Androlaelaps tachyoryctes (Radford)

Haemolaelaps tachyoryctes Radford, 1941, Parasitology 33 : 309, figs. 3-4; Keegan, 1956, /.

Egypt, publ. Hlth. Ass. 31 : 236, figs. 19-21.
Hypoaspis (Haemolaelaps} tachyoryctes, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 71.

FEMALE (figs. 154-157) : Deutosternum provided with 6 rows of 2-4 very small
teeth. Chelae 34-36^ long; pilus dentilis moderately inflated, elongate, tapering to
a point; arthrodial filaments subequal in length.

Dorsal shield granular, with a faint reticulate pattern. DL= 580-627^; DW=
SSS-SS ^ i DL/DW= 1-7-1-8. The shield bears 39 pairs of setae in the type
specimens, but in the other material examined several of the r series are sometimes
off the plate. Most of the setae are relatively short and fine ; setae 24 and Z2 about
half as long as the distance between their bases; seta Z$ at least twice as long as
seta /5 and considerably thicker.

Sternal shield granular, presternal area reticulate; SL=jo-jg\; SW=i45-i48(x;
SL/SW=o-5-o-6. First pair of sternal setae situated on presternal area, as long as
metasternal setae, 2nd and 3rd pairs of sternal setae longer.

Genital shield granular, slender, tapering behind the genital setae; the length of
the latter about 3/4 GW 2 . GW^ 76-86^; GW 2 =8o^.

Anal shield granular, very strongly convex anteriorly. AL=ii4-i22[z; AW=
72-80^; AL/ AW =1-4-1 -6. Anus situated at 1/3 of its length from the anterior
margin. Paranal setae level approximately with middle of anus, about 4/5 as long
as postanal seta and thinner.

Integument of idiosoma bears 23 pairs of ventral and marginal setae, two pairs of
ventral setae bordering the genital shield. Metapodal plates small, nearly triangular
in shape, length to width about 2:1.

Chaetotaxy of legs normal. Setae on coxa I subequal in length. Tarsus II
bears 3 stout ventral setae and 3 stout, blunt, apical setae. Tarsus III has 3, and
tarsus IV has 4 stout, rather blunt, apical setae. Length of tarsus IV slightly less
than 4 times its width at the base.

MALE (figs. 158-159) : Fixed digit of chelicera bears a slightly inflated, tapering
pilus dentilis. Chaetotaxy of dorsal shield as in the female, but all the setae are

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat.

156

157

158

FIGS. 154-159. Androlaelaps tachyoryctes (Radford). Venter (Fig. 154), dorsum (Fig.
155) and chelicera (Fig. 156) of female; ventral view of tarsus II of female (Fig. 157).
Venter (Fig. 158) and chelicera (Fig. 159) of male.

8o W. M. TILL

situated on the shield. DL 475-49450.; DW= 290-304(0.; DL/DW 1-6-1-7.
Holoventral shield greatly expanded behind 4th pair of coxae and bears 23 setae,
5 pairs in the region between coxae IV and the anus. Integument of idiosoma
bears 13 pairs of setae. Chaetotaxy of legs as in female.

HOSTS AND LOCALITIES: Tachyoryctes splendens (Riippell) from several localities
in Kenya (types in B.M.N.H., other material presented by Miss J. B. Walker to
S.A.I.M.R.).

Lemniscomys striatus (Linnaeus), Njoro, Rift Valley Province, Kenya (Keegan,
1956).

Androlaelaps taterae (Zumpt & Patterson)

Hypoaspis (Haemolaelaps) taterae Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 83, fig. 6.
Haemolaelaps taterae, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 244.

FEMALE (figs. 160-162): A. taterae is very closely allied to A. villosissimus,
differing from it chiefly in being more densely haired. This is particularly noticeable
in the region between the genital and anal shields, where there are about 13 setae,
and in the area between the genital shield and the meiapodal plates.

The chelae of A. taterae are 44-48(0, long; pilus dentilis strongly inflated basally,
terminal half slender and sharply bent; arthrodial filaments variable in size, the
longest being about twice the average length.

Dorsal shield faintly reticulate and has a dense overall covering of setae.
DL= 813-900(0,; DW= 480-550(0.; DL/DW= 1-6-1-7.

Sternal shield reticulate, sharply demarcated from the reticulate and granular
presternal area. SL= 128-148(0,; SW= 166-175(0,; SL/SW=o-8-o-9. Metasternal
setae nearly as long as first pair of sternal setae ; the latter are barbed.

Genital shield with an overall reticular and granular pattern, slightly widened
posteriorly. GW X = 150-162(0.; GW 2 = 135-140(0.; GW 1 /GW 2 =i-i-i-2. Length of
genital setae about 3/5 GW 2 .

Anal shield reticulate; AL= 152-162(0,; AW= 124-138(0,; AL/AW= 1-1-1-2. Anus
situated at approximately its length from the anterior margin. Paranal setae level
with middle of anus, about 4/5 as long as postanal seta. Metapodal plates roughly
oval, length about if 2\ times the width.

Genu III bears 10 setae, tibia III bears 9, otherwise the chaetotaxy of the legs
follows the normal pattern for the genus. Posterior seta on coxa I slightly longer
than anterior seta; many leg setae barbed. Length of tarsus IV is 9 times its width
at the base.

MALE (figs. 163-164) : Chelicera as figured, the fixed digit bearing a pilus dentilis
similar to that of the female. Dorsal shield and integument as in female; DL=
599(0.; DW=359(o,; DL/DW=i-7. Holoventral shield reticulate, bearing numerous
setae on the posterior portion.

HOSTS AND LOCALITIES: Widely distributed in Southern Africa on Tatera afra
(Gray). Single specimens have been recorded from Rattus natalensis (Smith),
Rattus chrysophilus (de Winton) and Otomys irroratus (Brants) (S.A.I.M.R.,
B.M.N.H., and M.R.A.C.).

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal.

162

164

163

FIGS. 160164. Androlaelaps taterae (Zumpt & Patterson). Venter (Fig. 160), dorsum
(Fig. 161) and chelicera (Fig. 162) of female. Venter (Fig. 163) and chelicera (Fig. 164)
of male.

ZOOL. 10, i

W. M. TILL

Androlaelaps tateronis (Radford)

I schnolaelaps -tateronis Radford, 1939, Parasitology 31 : 247, fig. 4.

Hypoaspis (Haemolaelaps} tateronis, Zumpt & Patterson, 1951, /. ent. Soc. S. Afr. 14 : 71.
Haemolaelaps tateronis, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 242.
Haemolaelaps radfordi Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 241, figs. 24-25 (syn. nov.}.

FEMALE (figs. 165-167) : Deutosternum provided with 6 rows of 2-4 teeth. Chelae

165

166

FIGS. 165-167.

Androlaelaps tateronis (Radford), female. Venter (Fig. 165); dorsum
(Fig. 166); chelicera (Fig. 167).

about 36^ long ; pilus dentilis inflated basally, terminal portion slender and curved ;
arthrodial filaments subequal in length.

Dorsal shield oval, reticulate. DL= 656-664^; DW= 400-409^; DL/DW=
1-6-1-7. The shield bears 39 pairs of setae of moderate length; lengths of setae 44
and Z2 approximately equal to the distance between their bases; seta Z$, the longest
on the shield, is about twice as long as seta 75.

Sternal shield reticulate, sharply demarcated from granular and reticulate

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. hit. 83

presternal area. SL= 101-116(0.; SW= 136-1 37^; SL/SW^o-y-o-g. Metasternal
seta about as long as first sternal seta, second and third setae slightly longer.

Genital shield narrow, tapering behind genital setae, with a reticulate pattern
which extends to the anterior margin. GW 2 = 106-113^; genital setae extend
beyond bases of first pair of flanking setae, their length about 3/4 GW 2 .

Anal shield pear-shaped. AL= 124-139(0,; AW= 96-106(0. ; AL/AW=i-3. Anus
situated at about its length from the anterior margin; paranal setae level approxi-
mately with middle of anus, about as long as postanal seta.

Integument of idiosoma bears 21 pairs of setae, of which 2 pairs flank the genital
shield, a third pair being placed between the genital and anal shields. Some of the
longer setae have very fine barbs. Metapodal plates slender, length to width
about 6 : i.

Tibia III has 9 setae; the number on tibia II is not clear in all the specimens
examined, but is either 9 or 10. Posterior seta on coxa I twice as long as anterior
seta and considerably thicker. Anterior seta on coxa II and both setae on coxa III
thickened. Three of the apical setae on tarsus II stout and spine-like, one of them
blunt. Length of tarsus IV 6 times the width at the base.

MALE: Not known.

HOSTS AND LOCALITIES: Tatera valida (Bocage), Gulu District, Uganda (syntype
of /. tateronis in B.M.N.H.).

" Rat ", Yei, Sudan (paratype of H. radfordi in B.M.N.H.).
Tatera nigricauda Peters, Lali Hills, Kenya (S.A.I.M.R.).
Unidentified rodents from Uganda (B.M.N.H.) and from the Cameroons (I.R.S.C.).

Androlaelaps tauffliebi sp. nov.

FEMALE (figs. 168-170) : Deutosternum provided with 6 rows of 3-4 teeth; anterior
rostral seta twice as long as capitular seta. Chelae 24(0. long ; movable digit divided
into two parts, a ventral (medial) part which is edentate, except for the terminal hook,
and a dorsal (external) part which is broad and dentate. Fixed digit weakly
sclerotized, bearing a flagellar pilus dentilis. About 4 of the arthrodia] filaments
conspicuously long and stout.

Dorsal shield widest about level of seta ?% with a granular appearance and very
faint indications of a reticular pattern. Laterally it has a distinct double outline.
DL=532-56o[o,; DW= 400-418(0.; DL/DW=i-3-i-4. The shield bears the usual
39 pairs of setae, *3-*'5, 22-2:3, /i-/5 and px2-p%3 being reduced to microsetae.
Seta Z5 five times as long as seta /5. In all the specimens examined the anterior
margin of the shield is tucked under, so that setae si become marginal in position
and ii and n are ventral.

Sternal shield about 3 times as wide as long, deeply concave posteriorly, granular
in appearance. SL= 43-48^; SW= 138-143(0.; SL/SW=o-3O-o-35. Sternal seta III
nearly i| times as long as sternal seta I ; metasternal seta short and fine, about half
as long as sternal seta I.

Genital shield large, broadly rounded posteriorly, granular. GW X = 142-157(0.;
GW 2 =i20-i33(x; GW 1 /GW 2 =i-2-i-3. Genital setae short, about 1/3 GW 2 .

W. M. TILL

170

171

FIGS. 168-172. Androlaelaps taujfiiebi sp. nov. Venter (Fig. 168); dorsum (Fig. 169)
and chelicera (Fig. 170) of female. Venter (Fig. 171) and chelicera (Fig. 172) of male.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 85

Anal shield with some granulation and striations laterally. AL=77~79[x;
AW=9i-98[j.; AL/AW=o-8-o-9. Anus situated at about half its length from the
anterior margin. Paranal setae inserted near anterior third of anus, about the
same length as the postanal seta.

Integument of idiosoma bears about 20 pairs of setae, some of the marginal ones
having fine barbs. Metapodal plates very long and slender, length about 10 times
width.

Chaetotaxy of legs normal, setae simple, without any marked modifications.
Posterior seta on coxa I approximately i| times as long as anterior seta and slightly
thicker. Length of tarsus IV about 6 times its width at the base.

MALE (figs. 171-172) : Deutosternum and gnathosomal setae as in female.
Chelicera provided with an elongate spermadactyl ; minute teeth present at tip of
movable digit; fixed digit reduced, membranous. Total length of chela about

Dorsal shield as in female. DL=4i8[x; DW=295(i; DL/DW=i-4. Sterni-
ventral shield separated from anal shield, granular in appearance, with faint indica-
tions of reticulation, and bearing 8 pairs of setae. Anal shield approximately as
long as broad, otherwise similar to that of female. Metapodal plates relatively
broader than in the female.

Chaetotaxy of legs as in female, except that three setae on leg II are strongly
inflated, namely, the upper, antero- ventral seta on femur II, upper ant ero- ventral
seta on tibia II, and postero-ventral seta on tarsus II.

HOST AND LOCALITY: Seven females and one male from Cryptomys mechowi
Peters, Lac Calundo, Angola, 6 January, 1955 (from the collection of Dr. R. Taufflieb,
I.R.S.C., leg. A. de Barros Machado).

Holotype (1962.2.12.20), allotype (1962.6.12.21) and two female paratypes
(1962.6.12.22-23) in the collection of the British Museum (Natural History) ; two
female paratypes presented to Dr. R. Taufflieb, Institut de Recherches Scientifiques
au Congo ; two female paratypes presented to the Dundo Museum, Angola.

Androlaelaps theseus Zumpt

Hypoaspis (Androlaelaps) theseus Zumpt, 1950, Parasitology 40 : 301, figs. 3-4.

Turkiella theseus, Zumpt & Till, 1953, Ann. Inst. Med. trap. Lisboa 10 : 225.

Androlaelaps theseus, Cooreman, 1954, Ann. Mus. Congo Beige Tervuren, Zool. i : 163, figs. 1-4.

FEMALE (figs. 173-175) : Deutosternum provided with 6 rows each of 4-8 teeth.
Chelae io6-io8[j. long; pilus dentilis slender, arthrodial filaments subequal.

Dorsal shield reticulate and granular. DL= 1350-1570^; DW=iioo-i235fji;
DL/DW=i-2-i-3. The shield bears 40 pairs of setae, an additional pair being
present in the ax position. The setal lengths vary in mites from different hosts.
Seta r-L distinctly longer than seta ii ; setae 14 and zz slightly longer than the distance
between their bases; seta Z$ at least i| times as long as seta 75.

Sternal shield granular and reticular, anterior and posterior margins concave.
SL=i43-i70|ji ; SW=266-2?6y. ; SL/SW=o-5-o-6. First sternal seta reaches almost
to base of third; second and third sternal and metasternal setae at least i| times as
long as first sternal seta,

W. M. TILL

Genital shield granular, slender, very slightly widened behind genital setae,
then tapering posteriorly. GW X = 205-247^; GW 2 = 190-220^; GW 1 /GW 2 =i-i-i-2.
Length of genital setae exceeds the width of the shield at the level of their bases.

Anal shield reticulate and granular. AL= 180-205^; AW= 209-238(1; AL/AW
o-8-i-o. Anus situated at approximately its length from the anterior margin.
Paranal setae level with posterior half of anus, about f as long as postanal seta.

Integument of idiosoma bears about 42 pairs of ventral and marginal setae, many
of which appear to be barbed. Metapodal plates broadly oval, length approximately
ij times breadth.

174

FIGS. 173-175. Androlaelaps theseus Zumpt, female. Venter (Fig. 173); dorsum
(Fig. 174); ventral view of femur, genu and tibia of leg II (Fig. 175).

Chaetotaxy of legs normal. Anterior and posterior setae on coxae I subequal in
length, the posterior one slightly thicker. Spur on second femur relatively small,
its length being about 1/5 the width of the femur. Inner and outer ventral setae
on tibia II subequal in length and thickness ; inner and outer ventral setae on genu II
both simple. Tarsus II has 3 stout, blunt, terminal spines. Length of tarsus IV
about 7 times its width at the base.

MALE (fig. 176-177) : Gnathosoma as in female ; fixed digit of chelicera elongate,
membranous, bearing a slender pilus dentilis.

Dorsal shield resembles that of female; DL= 1283-1302^; DW= 920-940^;
DL/DW=i-4- Holoventral shield strongly reticulate and also granular, not

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lot. 87

expanded behind coxae IV, bearing 23 setae of which 10 (6+4) are situated in the
preanal region behind coxae IV.

Chaetotaxy of legs as in female, but with inner ventral seta on tibia II stouter,
and inner ventral seta on genu II shorter and stouter.

HOSTS AND LOCALITIES: Widely distributed in Southern Africa on Tatera afra
(Gray), and found occasionally on Rattus natalensis (Smith), Rattus chrysophilus

FIGS. 176-178. Androlaelaps theseus Zumpt, venter (Fig. 176) and chelicera (Fig. 177) of male.
Androlaelaps villosissimus (Berlese), venter female (Fig. 178).

(De Winton), Acomys cahirinus (Desmarest), Saccostomus campestris Peters and
Steatomys pratensis Peters (Zumpt & Till, 1961). Recorded from Tatera afra (Gray)
in the Belgian Congo (Cooreman, 1954). Material in S.A.I.M.R. and B.M.N.H.

Androlaelaps villosissimus (Berlese)

Hypoaspis (Haemolaelaps) villosissimus Berlese, 1918, Redia 13 : 128; Zumpt & Patterson, 1951,

/. ent. Soc. S. Afr. 14 : 72, fig. 5.

Haemolaelaps villosissimus, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 244; Keegan,
, Trans, Amer, micr, Soc. 75 : 314, figs. 1-2,

88 W. M. TILL

FEMALE (fig. 178) : Deutosternum provided with 6 rows of 3-4 small teeth.
Chelae 40^ long; pilus dentilis and arthrodial filaments indistinct.

Dorsal shield with only faint indications of a reticular pattern. DL=7i8-744(z;
DW=429~455[x; DL/DW= 1-6-1-7. The shield has a dense overall covering of
setae which completely obscures any pattern. Most of the setae appear to have
fine barbs.

Sternal shield reticulate, sharply demarcated from presternal area. SL=
no-i2O(ji; SW= 153-162^; SL/SW= 0-7-0-8. Metasternal setae as long as first
pair of sternal setae.

Genital shield granular, with an overall reticular pattern. GW^ 140-144^;
GW 2 =n8-i23(ji; GW 1 /GW 2 =i-i-i-2. Length of genital seta about half GW 2 .

Anal shield reticulate anteriorly and antero-laterally and also granular in appear-
ance. AL= 105-1 i4[x; AW=96-ioOfj,; AL/ AW == 1-1-1-2. Anus less than its
length from the anterior margin of the shield. Paranal setae level with middle of
anus, about 4/5 as long as postanal seta.

Integument of idiosoma bears numerous setae, 8 or 9 being situated between
the genital and anal shields. The ventrally placed setae appear to be simple,
whereas the marginal ones are barbed. Metapodal plates oval, length about twice
width.

Genu III bears 10 setae and tibia III bears 9; otherwise the chaetotaxy follows
the normal Androlaelaps pattern. Posterior seta of coxa I slightly longer than
anterior seta; many leg setae barbed. Length of tarsus IV 10 times width at base.

MALE: Not known.

HOSTS AND LOCALITIES : Saccostomus campestris Peters, Zululand (Berlese, 1918)
and Bulawayo, Southern Rhodesia (S.A.I.M.R.). Unidentified rodent from the
Cameroons (I.R.S.C.).

Androlaelaps walker ae sp. nov.

FEMALE (figs. 179-181) : The hypostomal processes form two pairs of prominent
brushes, one ventral in position and the other dorsal. Deutosternum provided
with 6 rows of 2-3 teeth. Chelae about 48^ long; pilus dentilis slender; arthrodial
filaments subequal in length.

Dorsal shield oval, reticulate, showing pronounced hypertrichy which extends
anterior to setae 23 and which is not confined to the median part of the shield.
The basic chaetotactic pattern is thus partially obscured. DL= 760-780^;
DW=475-494|x; DL/DW=i5-i-6.

Sternal shield reticulate, anterior margin not sharply defined. SL=95-iiO(j.;
SW= 133-142(0.; SL/SW= 0-7-0-8. Sternal and metasternal setae subequal in
length.

Genital shield flask-shaped, with a reticular pattern extending to the anterior
flap. GWj= 133-142^; GW 2 =86-io5[j.; GW 1 /GW 2 =i-4-i-5. Genital setae as
long as first pair of sternal setae and approximately equal to GW 2 .

Anal shield arched anteriorly. AL= 142-152^ ; AW= 103-110^; AL/AW=i-4.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 89

Anus situated at approximately its length from the anterior margin of the shield.
Paranal setae level with middle of anus, i| times as long as the barbed postanal
seta.

179

180

FIGS. 179-181. Androlaelaps walkevae sp. nov., female. Venter (Fig. 170); dorsum

(Fig. 1 80); chelicera (Fig. 181).

Integument of idiosoma provided with numerous setae, nearly all of which appear
to be barbed. Metapodal plates oval or irregular in shape, length to width not
more than 3:1.

Chaetotaxy of legs normal except that tibia I has 14 and tibia III has 9 setae.
Tarsus II has 3 stout terminal and 3 stout ventral bristles, and many of the leg

W. M. TILL

setae are barbed. Posterior seta of coxa I slighty longer than anterior seta. Length
of tarsus IV 7 times width at base.

MALE: Not known.

HOST AND LOCALITY: Sixty-three females from Tachyoryctes splendens (Riippell),
Muguga North, Kenya, 6 July, 1954, presented by Miss J. B. Walker, East African
Veterinary Research Organization, to the South African Institute for Medical
Research. Holotype and 38 paratypes in the collection of the S.A. Institute for

FIGS. 182-184. Androlaelaps wilkini (Till), female. Venter (Fig. 182); dorsum (Fig.

183); chelicera (Fig. 184).

Medical Research, 25 paratypes (1962.6.12.23-33) in the collection of the British
Museum (Natural History).

Androlaelaps wilkini (Till)

Haemolaelaps wilkini Till, 1959, J. ent. Soc. S. Afr. 22 : 432, figs.

FEMALE (figs. 182-184) : Deutosternum provided with 6 rows of 3-4 small teeth.
Chelae about 40^ long; pilus dentilis small, inflated; arthrodial filaments subequal
in length.

Dorsal shield oval, reticulate; DL= 709-770^; DW=455-5oo(x; DL/DW=i'5-i-6.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 91

The shield bears 37 pairs of setae, ^4 and r6 being inserted on the adjacent integu-
ment. Marginal setae of the r and 5 series barbed, the S setae being at least twice
as long and thick as those of the / series. Seta Z$ sinuous, 2-^-3 times as long as
seta /5. Setae on the median part of the shield (i series posterior to 13, J series,
Z2-Z3, Zi-Z4, p%2-px3) subequal in length, setae 24 and 22 slightly more than half
as long as the distance between their bases.

Sternal shield and presternal area both reticulate. SL^o-ygyi; SW= 145-158^;
SL/SW=o-5- First pair of sternal setae slightly shorter than 2nd and 3rd pairs and
situated on presternal area. Metasternal setae as long as first pair of sternal setae.

Genital shield relatively slender, very slightly widened behind the genital setae,
lateral margins almost parallel, posterior portion with a reticulate pattern.
GW 1 =i23-i49[x; GW 2 =96-ii4n; GW 1 /GW 2 =i -1-1-3.

Anal shield pear-shaped, with a reticulate pattern anteriorly and antero-laterally.
AL=i23-i40[x; AW=ii4-i27[o.; AL/AW=i-i-i-2. Paranal setae on a level
between middle and posterior margin of anus, their length not more than 2/3 that
of the postanal seta.

Integument of idiosoma bears 24 pairs of ventral and marginal setae, the longer
ones being provided with fine barbs; posterior terminal pair long and whip-like.
Metapodal plates slender, slightly curved.

Tibia III has 9 setae; anterior and posterior setae on coxa I uniform in length and
thickness. Tarsus II bears two blunt terminal spines; tarsi III and IV also have a
pair of terminal spine-like bristles, but these are more slender and pointed than the
spines on tarsus II. Length of tarsus IV 5 times width at base.

MALE: Not known.

HOSTS AND LOCALITIES: Myrmecocichla formicivora (Vieillot), Debeete, Bechuana-
land (type series, S.A.I.M.R. and B.M.N.H.).
Merops apiaster Linnaeus, Johannesburg, Transvaal (S.A.I.M.R.).

Androlaelaps zulu (Berlese)

Hypoaspis (Haemolaelaps) inops var. zulu Berlese, 1918, Redia 13 : 125.
Haemolaelaps inops var. zulu, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 234, fig. 18.

FEMALE (figs. 185-187) : Deutosternum provided with 6 rows each of 4 narrow
pointed teeth. Chelae 40(j. long ; pilus dentilis strongly inflated basally, distal third
or fourth slender, with recurved apex; arthrodial filaments very short, subequal in
length.

Dorsal shield reticulate, bearing 39 pairs of setae and 1-3 unpaired accessory setae
between the / series. Setae on median part of shield relatively long, lengths of
setae 14 and 22 about i times the distance between their bases. Marginal setae
slightly longer and thicker, with fine barbs. Seta Z5 at least i times as long as
seta 75. DL=627-76ofji; DW= 400-494^; DL/DW=i-5-i-6.

Sternal shield reticulate, sharply demarcated from reticulate presternal area.
SL=ioo-io5(x; SW=i28-i52{x; SL/SW=o7-o-9. Metasternal setae about as long
as first pair of sternal setae.

92 W. M. TILL

Genital shield reticulate, moderately expanded behind genital setae. GW X =
143-170^; GW 2 = 105-1 14^; GW 1 /GW 2 =i-3-i-5. Genital setae relatively short,
slightly more than half GW 2 .

Anal shield reticulate antero-laterally. AL=g6-i2O[i.; AW= 103-127^; AL/AW
=0-9-1-0. Anus situated less than its length from the anterior margin of the
shield. Paranal setae between middle and posterior end of anus, reaching beyond
base of, and about 4/5 as long as, postanal seta.

185

FIGS. 185-187. Androlaelaps zulu (Berlese), female. Venter

(Fig. 1 86); chelicera (Fig. 187).

(Fig. 185); dorsum

Integument of idiosoma bears about 22 pairs of marginal and ventral setae, of
which 3 pairs border the genital shield. Marginal setae coarser than the truly
ventral ones, and provided with fine barbs. Outer metapodal plates variable,
usually cigar-shaped, length 2^-3-J- times the width,

ETHIOPIAN MITES OF THE GENUS A N D R O L A E L . IP S BE RLES E 5. tat. 93

Chaetotaxy of legs normal. Anterior seta on coxa I about 4/5 as long as posterior
seta. Length of tarsus IV at least 6 times width at base.

MALE: Not known.

HOSTS AND LOCALITIES: Lemniscomys griselda (Thomas), Rattus natalensis (Smith)
and Otomys irroratus (Brants) from the Transvaal (S.A.I.M.R.).

188

189

FIGS. 188-190. Androlaelaps zuluen&is Zumpt, female. Venter (Fig. 188); dorsum
(Fig. 189); ventral view of femur, genu and tibia of leg II (Fig. 190).

Tatera afra (Gray) from Zululand (S.A.I. M.R.).

(Mites from the above hosts were previously identified by Zumpt & Till (1956) as

forms of Haemolaelaps glasgowi.)

Keegan (1956) has recorded this species from several rodents in Egypt.

94 W. M. TILL

Androlaelaps zuluensis Zumpt

Hypoaspis (Androlaelaps} zuluensis Zumpt, 1950, Parasitology 40 : 300, fig. i.
Turkiella zuluensis, Zumpt & Till, 1953, Ann. Inst. Med. trop. Lisboa 10 : 218, figs. 2-5.
Androlaelaps zuluensis, Keegan, 1956, /. Egypt, publ. Hlth. Ass. 31 : 229, fig. 14.

FEMALE (figs. 188-190) : Deutosternum provided with 6 rows each of 3-4 very
small teeth. Chelae 46^ long; pilus dentilis slender; arthrodial filaments short and
subequal in length.

Dorsal shield reticulate, oval. DL= 656-675^; DW 403-486^; DL/DW=i-5.
The shield bears 39 pairs of setae as well as 2 or 3 unpaired accessory setae between
the J series. Lengths of setae 24 and Z2 approximately equal to the distance
between their bases; seta Z$ about i| times as long as seta /5. The marginal setae
appear to be barbed.

Sternal shield reticulate, posterior margin almost straight or slightly irregular.
SL=io5[j.; SW=n8-i24[A; SL/SW=O'9. Sternal and metasternal setae all
relatively short, a little more than half the length of the sternal shield.

Genital shield with a pattern of striations on its posterior portion; very long,
extending almost to the anal shield. GW 1 =iyo[i; GW 2 = 118-124^; GWJGWg
= 1-4. Genital setae very short, less than half GW 2 .

Anal shield with a reticulate pattern; AL= 105-109^; AW=io5[j.; AL/AW=i-o.
Anus situated at less than its length from the anterior margin of the shield. Paranal
setae near posterior margin of anus; postanal seta broken in both the specimens
examined.

Integument of idiosoma bears 20-22 pairs of marginal and ventral setae. Meta-
podal plates slender, length 3-6 times the breadth.

Chaetotaxy of legs normal. Anterior seta on coxa I slightly shorter than
posterior seta. Spur on femur II rather slender, thumb-like, its length a little less
than half the width of the femur at its base. Inner ventral seta on genu II short,
spur-like; inner ventral seta on tibia II considerably stouter than outer one; tarsus II
terminates in pointed setae only. Length of tarsus IV about 8 times its width at
the base.

MALE (fig. 191) : Deutosternum with 6 rows of 2-5 small teeth. Spermadactyl
6o(A long, fixed digit bears a slender pilus dentilis.

Dorsal shield similar to that of female; DL=52Opt. (approx.) ; DW=36o[u.; DL/
DW=i-4- Holo ventral shield reticulate, expanded behind coxae IV, bearing 23
setae, 5 pairs of setae being situated in the preanal region behind coxae IV.
Chaetotaxy of legs similar to that of female, except that inner ventral setae on genu II
and tibia II are more slender.

HOSTS AND LOCALITIES : Rattus namaquensis (Smith) from Abercorn Pont, Zululand
(holotype in S.A.I.M.R.), and from Kapps Siding, S. W. Africa (S.A.I. M.R.)
Rattus chrysophilus (De Winton), Southern Africa (Zumpt & Till, 1961).
Petromyscus collinus (Thomas & Hinton), Kamanjab-Ohopoho Road, S.W. Africa
(S.A.I.M.R.).

Galago senegalensis Geoffroy (the figured specimens) from Karamoja, Uganda
(S.A.I.M.R.).

ETHIOPIAN MITES OF THE GENUS AND1WLAELAPS BERLESE s. hit. 95

Keegan (1956) records this species from Arvicanthis niloticus (Desmarest) and Mus
musculus Linnaeus in Egypt.

Androlaelaps zutnpti sp. nov.

FEMALE (figs. 193-195): Deutosternum bears 6 rows of 3-5 very fine teeth.
Chelae about 40^. long; pilus dentilis inflated; one arthrodial filament strikingly
longer than the others ; an extension of the arthrodial membrane between the
chelae is elongate, reaching the tips of the chelae.

191

192

FIGS. 191-192. Androlaelaps zuluensis Zumpt, venter of male (Fig. 191).
Androlaelaps zumpti sp. nov., venter of male (Fig. 192).

Dorsal shield reticulate, widest in posterior half at level of setae Si and 52.
DL= 656-732 [A; DW= 442-494(0. ; DL/DW= 1-3-1-5. The shield bears 39 pairs of
setae ; lengths of setae 24 and 22 a little more than half the distance between their
bases ; length of seta /i is f that of seta Si ; seta Z$ is 2-2 1 times as long as seta 75.

Sternal shield reticulate, sharply demarcated from reticulate and granular
presternal area. SL=86-ii4(j.; SW= 124-133^; SL/SW=o-7-o-9. Metasternal

9 6

W. M. TILL

and first pair of sternal setae subequal in length, 2nd and 3rd pairs of sternal setae
slightly longer.

Genital shield scarcely widened behind genital setae, tapering posteriorly, with a
well-defined reticular pattern extending over the anterior as well as posterior parts.

193

194

FIGS. 193-195. Androlaelaps zumpti sp. nov., female. Venter (Fig. 193); dorsum

(Fig. 194); chelicera (Fig. 195).

GW 1 =95-ii4[ji; GW 2 =95-io5[j.. Genital setae about as long as first pair of sternal
setae and about 2/3 GW 2 .

Anal shield about 133^ long, 90^ wide; AL/AW=i-5. Anus situated at less than
its length from the anterior margin of the shield. Paranal setae near posterior
margin of anus, about 2/3 as long as postanal seta.

Integument of idiosoma bears about 23 pairs of ventral and marginal setae, those

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 97

near the posterior extremity of the body having fine barbs. Metapodal plates
slender, length at least 5 times the width.

Chaetotaxy of legs normal, anterior seta on coxa I about 4/5 as long as posterior
seta. Trochanter IV bears a stout dorsal seta. Length of tarsus IV about 6 times
its width at the base.

MALE (fig. 192) : Deutosternum bears 6 rows of 4-7 very fine teeth. Fixed digit
of chelicera bears a pilus dentilis similar in shape to that of the female.

Dorsal shield 55O[j. long, 352[A wide; chaetotaxy similar to that of female. Holo-
ventral shield strongly reticulate, bearing 23 setae, 5 pairs being placed in the region
between coxae IV and the anus. Chaetotaxy of legs as in the female.

HOSTS AND LOCALITIES: Holotype (female), allotype and one male and three
female paratypes from Rattus (Aethomys) sp., Richmond, Cape Province, 31 March,
1959, two female paratypes from Rattus (Aethomys) sp., Middelburg, Cape Province,
28 March, 1959. Holotype, one male and two female paratypes in the collection
of the South African Institute for Medical Research; allotype (1962.6.12.34) and
two female paratypes (1962.6.12.35-36) in the collection of the British Museum
(Natural History).

NOTE: Several populations of mites have been examined which are very similar
to A . zumpti, except that the setae of the dorsal shield are longer (lengths of setae
24 and Z2 about i^ times the distance between their bases) and 2-4 accessory setae
are present. These mites, all from the collection of the S.A. Institute for Medical
Research, have been taken from the following hosts and localities:
Rattus chrysophilus (De Winton), Matopos, S. Rhodesia.

Rattus namaquensis (Smith), Bulawayo, S. Rhodesia; Abercorn Pont, Zululand;
several localities in S.W. Africa.

Rattus paedulcus (Sundevall), Kaokoveld, S.W. Africa.
Otomys unisulcatus Cuvier, Luckhoff, Orange Free State.
Rhabdomys pumilio (Sparrman), Cape Province and S.W. Africa.
Petromyscus collinus (Thomas & Hinton), Kamanjab-Ohopoho, S.W. Africa.

Specimens from Parotomys littledalei Thomas, van Rhynsdorp, Cape Province, are
intermediate between these and the short-haired form, the lengths of setae 24 and Z2
being slightly less than the distance between their bases. For the present, these
will all be regarded as forms of A. zumpti.

HOST-PARASITE LIST.*

CLASS AVES

Order FALCONIFORMES
FAMILY AEGYPIIDAE

Gyps coprotheres (Forster)

Androlaelaps pater soni (Zumpt & Till)

* The host names are the same as those used by Zumpt (1961).

ZOOL 10, I

98 W. M. TILL

Order CORAGIIFORMES
FAMILY MEROPIDAE

Merops apiaster Linnaeus

Androlaelaps wilkini (Till)
FAMILY PHOENIGULIDAE

Phoeniculus purpureus (Miller)

Androlaelaps phoeniculi (Zumpt & Till)
FAMILY CAPITONIDAE

Lybius torquatus (Dumont)
Androlaelaps steyni (Till)
Trachyphonus vaillanti Ranzani

Androlaelaps steyni (Till)
FAMILY PICIDAE

Campethera abingoni (Smith)
Androlaelaps haydocki (Till)
Thripias namaquus (Lichtenstein)

Androlaelaps haydocki (Till)
Mesopicos griseocephalus (Boddaert)
Androlaelaps mesopicos (Radford)

Order PASSERIFORMES
FAMILY TURDIDAE

Myrmecocichla formicivora (Vieillot)

Androlaelaps wilkini (Till)
FAMILY HIRUNDINIDAE

Riparia paludicola (Vieillot)

Androlaelaps spreo (Zumpt & Till)
FAMILY STURNIDAE

Lamprotornis nitens (Linnaeus)

Androlaelaps steyni (Till)
Spreo bicolor (Gmelin)

Androlaelaps spreo (Zumpt & Till)
FAMILY PLOCEIDAE

Spermestes cucullatus Swainson
Androlaelaps congoensis sp. nov.

CLASS MAMMALIA

Order INSEGTIVORA
FAMILY MACROSCELIDIDAE

Elephant Shrew

Androlaelaps arvicanthis Radford
FAMILY SORICIDAE
Suncus varius (Smuts)

Androlaelaps suncus sp. nov.

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lal. 99

Crocidura spec.

Androlaelaps suncus sp. nov.
Crocidura manni Peters

(?) Androlaelaps hystrici (Zumpt & Till)

Order PRIMATES
FAMILY LORISIDAE

Galago alleni Waterhouse

Androlaelaps galagus (Lavoipierre)
Galago senegalensis Geoffrey
A ndrolaelaps zuluensis Zumpt

Order CARNIVORA
FAMILY MUSTELIDAE
Ictonyx striatus (Perry)

Androlaelaps marshalli Berlese
FAMILY VIVERRIDAE

Suricata suricatta (Schreber)
Androlaelaps marshalli Berlese

Order LAGOMORPHA
FAMILY LEPORIDAE

Oryctolagus cuniculus (Linnaeus)
Androlaelaps marshalli Berlese

Order RODENTIA
FAMILY BATHYERGIDAE

Bathyergus suillus (Schreber)

Androlaelaps scapularis (Berlese)
Georychus capensis (Pallas)

Androlaelaps capensis (Hirst)

Androlaelaps cryptomius (Radford)

Androlaelaps georychi sp. nov.
Cryptomys holosericeus (Wagner)

Androlaelaps capensis (Hirst)

Androlaelaps scapularis (Berlese)
Cryptomys hottentotus (Lesson)

Androlaelaps capensis (Hirst)

Androlaelaps marshalli Berlese

Androlaelaps scapularis (Berlese)
Cryptomys mechowi Peters

Androlaelaps tauffliebi sp. nov,

ioo W. M. TILL

FAMILY HYSTRICIDAE
Hystrix cristata Linnaeus

Androlaelaps hy strict (Zumpt & Till)
FAMILY SGIURIDAE

Sciurus carolinensis Gmelin

Androlaelaps casalis (Berlese)
Heliosciurus gatnbianus (Ogilby)

Androlaelaps heliosciuri sp. nov.
Xerus inauris (Zimmermann)

Androlaelaps casalis (Berlese)
FAMILY ANOMALURIDAE

Anomalurus derbianus (Gray)

Androlaelaps spatuliformis (Lavoipierre)
FAMILY PEDETIDAE

Pedetes capensis (Forster)

Androlaelaps marshalli Berlese
FAMILY MUSCARDINIDAE
Claviglis spec.

Androlaelaps ghanensis sp. nov.
FAMILY RHIZOMYIIDAE
Tachyoryctes spec.

Androlaelaps callosus (Berlese)
Tachyoryctes splendens (Riippell)

Androlaelaps marshalli Berlese

Androlaelaps tachyoryctes (Radford)

Androlaelaps walker ae sp. nov.
FAMILY MURIDAE

Arvicanthis niloticus (Desmarest)

Androlaelaps arvicanthis Radford

Androlaelaps marshalli Berlese

Androlaelaps murinus (Berlese)
Rattus spec.

Androlaelaps zumpti sp. nov.
Rattus chrysophilus (De Winton)

Androlaelaps dasymys (Radford)

Androlaelaps marshalli Berlese

Androlaelaps later ae (Zumpt & Patterson)

Androlaelaps theseus Zumpt

Androlaelaps zuluensis Zumpt

Androlaelaps zumpti sp. nov.
Rattus morio (Trouessart)

Androlaelaps rhodesiensis (Zumpt & Patterson)
Rattus namaquensis (Smith)

Androlaelaps dasymys (Radford)

Androlaelaps marshalli Berlese

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESK s. hit. 101

Androlaelaps zuluensis Zumpt

Androlaelaps zumpti sp. nov.
Rattus natalensis (Smith)

Androlaelaps arvicanthis Radford

Androlaelaps dasymys (Radford)

Androlaelaps marshalli Berlese

Androlaelaps murinus (Berlese)

Androlaelaps later ae (Zumpt & Patterson)

Androlaelaps iheseus Zumpt

Androlaelaps zulu (Berlese)
Rattus paedulcus (Sundevall)

Androlaelaps marshalli Berlese

Androlaelaps zumpti sp. nov.
Rattus rattus (Linnaeus)

Androlaelaps casalis (Berlese)

Androlaelaps dasymys (Radford)

Androlaelaps rhodesiensis (Zumpt & Patterson)
Rhabdomys pumilio (Sparrman)

Androlaelaps dasymys (Radford)

Androlaelaps glasgowi (Ewing)

Androlaelaps marshalli Berlese

Androlaelaps murinus (Berlese)

Androlaelaps zumpti sp. nov.
Lophuromys flavopunctatus Thomas

Androlaelaps callosus (Berlese)

Androlaelaps murinus (Berlese)
Lophuromys sikapusi (Temminck)

Androlaelaps dasymys (Radford)
Oenomys hypoxanthus (Pucheran)

Androlaelaps murinus (Berlese)
Dasymys incomtus (Sundevall)

Androlaelaps dasymys (Radford)

Androlaelaps murinus (Berlese)
Lemniscomys griselda (Thomas)

Androlaelaps marshalli Berlese

Androlaelaps zulu (Berlese)
Lemniscomys striatus (Linnaeus)

Androlaelaps murinus (Berlese)

Androlaelaps tachyorycles (Radford)
Acomys cahirinus (Desmarest)

Androlaelaps theseus Zumpt
Saccostomus campestris Peters

Androlaelaps marshalli Berlese

Androlaelaps oliffi (Zumpt & Patterson)

Androlaelaps rhodesiensis (Zumpt & Patterson)

W. M. TILL

Androlaelaps theseus Zumpt

Androlaelaps villosissimus (Berlese)
Cricetomys gambianus Waterhouse

Androlaelaps cricetomydis sp. nov.

Androlaelaps galagus (Lavoipierre)
Petromyscus collinus (Thomas & Hinton)

Androlaelaps zuluensis Zumpt

Androlaelaps zumpti sp. nov.
Steatomys pratensis Peters

Androlaelaps marshalli Berlese

Androlaelaps rhodesiensis (Zumpt Patterson)

Androlaelaps theseus Zumpt
Otomys spec.

Androlaelaps murinus (Berlese)
Otomys irroratus (Brants)

Androlaelaps dasymys (Radford)

Androlaelaps glasgowi (Ewing)

Androlaelaps murinus (Berlese)

Androlaelaps later ae (Zumpt & Patterson)

Androlaelaps zulu (Berlese)
Otomys saundersiae Roberts

Androlaelaps dasymys (Radford)
Otomys sloggetti (Thomas)

Androlaelaps dasymys (Radford)
Otomys unisulcatus Cuvier

Androlaelaps zumpti sp. nov.
Parotomys brantsi (Smith)

Androlaelaps dasymys (Radford)
Parotomys littledalei Thomas

Androlaelaps dasymys (Radford)

Androlaelaps zumpti sp. nov.
Mystromys albicaudatus (Smith)

Androlaelaps capensis (Hirst)

Androlaelaps dasymys (Radford)
Desmodillus auricularis (Smith

Androlaelaps marshalli Berlese

Androlaelaps oliffi (Zumpt & Patterson)
Gerbillus paeba Smith

Androlaelaps marshalli Berlese

Androlaelaps oliffi (Zumpt & Patterson)
Gerbillus pyramidum Geoffroy

Androlaelaps centrocarpus (Berlese)
Meriones rex Yerbury & Thomas

Androlaelaps longipes (Bregetova)

ETHIOPIAN MITES OF THE GENUS ANDROLAELAPS BERLESE s. lat. 103

Tatera spec.

Androlaelaps arvicanthis Radford

Androlaelaps marshalli Berlese
Tatera afra (Gray)

Androlaelaps dasymys (Radford)

Androlaelaps marshalli Berlese

Androlaelaps oliffi (Zumpt & Patterson)

Androlaelaps taterae (Zumpt & Patterson)

Androlaelaps theseus Zumpt

Androlaelaps zulu (Berlese)
Tatera nigricauda Peters

Androlaelaps marshalli Berlese

Androlaelaps tateronis (Radford)
Tatera valida (Bocage)

Androlaelaps arvicanthis Radford

Androlaelaps marshalli Berlese

Androlaelaps tateronis (Radford)
Taterillus spec.

Androlaelaps centrocarpus (Berlese)
Taterillus emini (Thomas)

Androlaelaps marshalli Berlese
Unidentified rodents

Androlaelaps arvicanthis Radford

Androlaelaps graingeri Zumpt & Patterson

Androlaelaps hirsti (Keegan)

Androlaelaps marshalli Berlese

Androlaelaps tateronis (Radford)

Androlaelaps villosissimus (Berlese)

REFERENCES

BERLESE, A. 1887. In " Acari, Myriopoda et Scorpiones hucusque in Italia reperta." Fasc. XL,
N. 6.

- 1911. Acarorum species novae quindecim. Redia 7 : 429-435.

- 1916. Centuria seconda di Acari nuovi. Redia 12 : 125-177.

- 1918. Centuria quarta di Acari nuovi. Redia 13 : 115-190.

BREGETOVA, N. G. 1952. New species of mites of the genus Haemolaelaps (Gamasoidea
Laelaptidae) parasitizing rodents. Zoo/. Zh. Moscow 31 : 860-882.

- 1956. Gamasid Mites (Gamasoidea). A cad. Sci. U.S.S.R., Moskva 61 : 246pp.
COOREMAN, J. 1954. Acariens du Congo Beige. Ire serie. Ann. Mus. Congo Beige Tervuren

i : 163-168.
COSTA, M. 1961. Mites associated with rodents in Israel. Bull. Brit. Mus. (nat. Hist.) Zool.

8 : 1-70.
DELFINADO, M. U. 1961. Haemolaelaps travisi, a new species of mite from the Philippines

(Laelaptidae : Acarina). Fieldiana, Zoology 44 : 49-51.
FONSECA, F. DA. 1959. Notes d'acarologie. XLI. Haemolaelaps Berlese versus Atricholae-

laps Ewing et Ischnolaelaps Fonseca; Ornithonyssus Samboa versus Bdellonyssns Fonseca.

Mem. Inst. Butantan S. Paulo (1957/58), 28 : 45-54.

104 W. M. TILL

GROKHOVSKAYA, I. M. and NGUEN-HUAN-HOE. 1961. Gamasid mites of North Viet-Nam.

Part 2. Zoo/. Zh. Moscow 40 : 1633-1646.
HIRST, S. 1916. Notes on parasitic Acari. Description of two new African mites of the

family Gamasidae. /. Zool. Res. i : 76-81.
KEEGAN, H. L. 1956. Ectoparasitic laelaptid and dermanyssid mites of Egypt, Kenya and

the Sudan, primarily based on Namru 3 collections, 1948-1953. /. Egypt, publ. Hlth.

Ass. 31 : 199-272.

- i956a. Original illustrations of Haemolaelaps marsupialis Berlese, 1910 and of five
additional Haemolaelaps species described, but not figured by Berlese. Trans. Amer.
micr. Soc. 75 : 314-319.

LAVOIPIERRE, M. M. J. 1955. A description of a new genus of sarcoptiform mites and of three
new species of Acarina parasitic on primates in the British Cameroons. Ann. trop. Med.
Parasit. 49 : 299-307.

- 1956. A description of a new genus and of three new species of mites (Acarina, Para-
sitiformes) parasitic on West African Mammals. Ann. trop. Med. Parasit. 50 : 291-298.

RADFORD, C. D. 1939. Notes on some new species of parasitic mites. Parasitology 31 : 243-

254-

- 1941. Notes on some new species of parasitic mites. Pt. 4. Parasitology 33 : 306-315.

- 1942. New ectoparasitic mites (Acarina) from Uganda. Parasitology 34 : 185-194.

- 1942^. New parasitic mites (Acarina). Parasitology 34 : 295-307.

1944. New parasitic mites from rodents. Parasitology 35 : 161-166.

STRANDTMANN, R. W. 1949. The blood-sucking mites of the genus Haemolaelaps (Acarina :

Laelaptidae) in the United States. /. Parasit. 35 : 325-352.
TAUFFLIEB, R. and MOUCHET, J. 1959. Notes sur les acariens (Acarina; Laelaptidae et

Spinturnicidae) du Cameroun. Ann. Parasit. hum. comp. 34 : 350-353.
TILL, W. M. 1959. Three new Haemolaelaps species (Acarina : Laelaptidae) from birds in

the Ethiopian region, and a redescription of Haemolaelaps mesopicos Radford. /. ent.

Soc. S. Afr. 22 : 423-435.
VITZTHUM, H. 1943. Acarina. In : Bronn, H. G., Klassen und Ordnungen des Tierreiches,

Bd. 5, Abt. IV, Buch 5, Leipzig (Becker & Erler).
ZUMPT, F. 1950. Notes on parasitic mites. I. Some remarks on the family Laelaptidae

(sensu Vitzthum 1943) with descriptions of three new species from African rodents.

Parasitology 40 : 298-303.

and PATTERSON, P. M. 1950. The Ethiopian species of Hypoaspis subgen. Androlaelaps
Berlese (1903), with description of a new species. S. Afr. J. Med. Sci. 15 : 67-74.

and PATTERSON, P. M. 1951. Further notes on laelaptid mites parasitic on vertebrates.
A preliminary study to the Ethiopian fauna. /. ent. Soc. S. Afr. 14 : 63-93.

- and TILL, W. 1953. The genera Turkiella nov. ( Androlaelaps auct.) and Haemolaelaps
in the Ethiopian region, with keys and descriptions of three new species (Acarina :
Laelaptidae). Ann. Inst. Med. trop. Lisboa 10 : 215-249.

and TILL, W. M. 1956. Notes on Haemolaelaps glasgowi (Ewing) and related forms in
the Ethiopian region, with descriptions of four new species (Acarina : Laelaptidae). Z.
Parasitenk. 17 : 282-291.

- and TILL, W. M. 1958. Notes on the classification and synonymy of gamasid mites
parasitic on vertebrates (Acarina : Mesostigmata) . /. ent. Soc. S. Afr. 21 : 261-273.

and TILL, W. M. 1961. Suborder Mesostigmata. In Zumpt (1961): The arthropod

parasites of vertebrates in Africa south of the Sahara (Ethiopian region). Vol. I
(Chelicerata) . Publ. S.A. Inst. Med. Res. 9 : 17-91.

THE FRANCOLINS,
A STUDY IN SPECIATION

B. P. HALL

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 2

LONDON : 1963

THE FRANCOLINS,
A STUDY IN SPECIATION

B. P. HALL

British Museum (Natural History)

Pp. 105-204 ; 2 Text-figures, n Maps

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 2

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
issued in jive 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.

This paper is Vol. 10, No. 2 of the Zoological series.

Trustees of the British Museum 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued March 1963 Price Thirty-one Shillings

THE FRANCOLINS, A STUDY IN SPECIATION

By B. P. HALL

Associate, Department of Z oology > British Museum (Natural History]

CONTENTS

Page
INTRODUCTION ........... 107

THE GENUS . .......... 108

THE GROUPS . . . . . . . . . . .no

SPECIES AND SUBSPECIES . . . . . . . . .no

METHODS . . . . . . . . . . . .112

CHARACTERS . . . . . . . . . . .115

THE SPOTTED GROUP (francolinus, pictus, pintadeanus) .... 123

THE BARE-THROATED GROUP (afer, swainsonii, rufopictus, leucoscepus) . 125
THE MONTANE GROUP (erckelii, ochropectus, castaneicollis , jacksoni, nobilis,

camerunensis, swierstrai) . . . . . . . .131

THE SCALY GROUP (ahantensis, squamatus, griseostriatus) . . .136

THE VERMICULATED GROUP (bicalcaratus, icterorhynchus , clappertoni,

hildebrandti, natalensis, hartlaubi, harwoodi, adspersus, capensis) . . 140
THE STRIATED GROUP (sephaena, streptophorus) . . . . .148
THE RED-WINGED GROUP (psilolaemus, shelleyi, africanus, levaillantoides,

levaillantii, finschi) . . . . . . . . . .152

THE RED-TAILED GROUP (coqui, albogularis, schlegelii] .... 160

SPECIES NOT INCLUDED IN ANY GROUP (lathami, nahani, pondicerianus ,

gularis) ............ 165

CONCLUSIONS . . . . . . . . . . . 168

SUMMARY . . . . . . . . . . . .172

APPENDIX i . Summary of conclusions reached by R. E. Moreau in his paper

" The Vicissitudes of the African Biota in the Late Pleistocene " . . 173
APPENDIX 2. Taxonomic Notes and Synonymy . . . . .175

MAPS ............ 180

REFERENCES ........... 200

INDEX ............ 201

INTRODUCTION

Thirty-nine species of francolin (Francolinus and Pternistis] were recognised in
Peters' Check List of the Birds of the World 2, 1934, of which thirty-four are African
and five Asiatic: one (F. ochropectus Dorst & Jouanin, 1952) has been described
since. I have rearranged Peters' species slightly, now recognising thirty-six in
Africa and five in Asia. As is inevitable a Check List employing conventional
nomenclature gives no indication of the relationship between the species which must
be the basis for any discussion on speciation. The objectives of this paper are
therefore twofold : firstly to set out in the text and on the maps the present relation-
ship between the species. This is not as formidable a task as might be expected

ZOOL. IO, 2 8

io8 B. P. HALL

from the number of species involved, for thirty-seven of the forty-one fall readily
into eight groups, seven of which are African and one Asiatic. (In this paper the
term " group " is used for either a superspecies, in which all members are largely
allopatric, or for a species group which is a rather looser assemblage of related forms
containing some species which are partly sympatric with other members.) These
groups are discussed and mapped separately with particular reference to their
ecology and the relationship of neighbouring or sympatric forms. The four species
that do not fall readily into any group are discussed separately at the end.

Secondly, from the pattern presented by the birds of each group and by their
distribution I have tried to reconstruct the climatic changes, with the resultant
isolating and rejoining of populations, which could have formed this pattern. In
doing so I am fully conscious that the picture we see today gives only fragmentary
clues to the past and that it is impossible to interpret accurately any but the simplest
situations. Nevertheless, even in the most complex groups, I believe there is
some usefulness in making the attempt, for if a series of studies of this nature suggests
changes that from other branches of science are known to have occurred, it may be
possible to give some tentative datings to various steps in speciation.

Throughout the preparation of this paper I have been particularly fortunate in
being able to discuss the problems with R. E. Moreau, who has been preparing
concurrently a paper on " The Vicissitudes of the African Biota in the Late
Pleistocene ". He has allowed me to quote as Appendix i a summary of his conclu-
sions of the major climatic changes that have taken place in Africa in the past
60,000 years. In acknowledging my gratitude to him for this and for invaluable
advice and criticism I must also absolve him from any responsibility for some of
my interpretations of the evolutionary stages, or the conclusions which I have drawn
from them.

I am also deeply indebted to M. P. Stuart Irwin who prepared distribution maps
for me of all the francolins in the National Museum of Southern Rhodesia: also
to Dr. H. Friedmann, C. W. Benson and I. C. J. Galbraith for criticism of the
manuscript. Others to whom I am grateful for loans of specimens, advice, or other
help, include Dr. Dean Amadon, Professor J. Berlioz, P. Blasdale, Mgr. F. O. Cave,
E. M. Cawkell, Dr. James Chapin, P. A. Clancey, Miss M. Court enay-Latimer,
I. H. Dillingham, H. J. de S. Disney, Sir Hugh Elliot, W. V. Harris, M. J. Hollis,
Cdr. A. M. Hughes (who drew figure i), Gen. Sir Gerald Lathbury, C. W. Mackworth-
Praed, Miles Markus, Col. R. Meinertzhagen, Captain C. R. S. Pitman, O. P. M.
Prozesky, Professor H. Schouteden. Rev. Dr. W. Serle, R. H. N. Smithers, B. W. H.
Stronach, M. A. Traylor, Dr. C. Vaurie, Col. Jack Vincent, Professor V. Van Straelen,
John Williams, Dr. J. M. Winterbottom.

THE GENUS

The francolins are recognised as a branch of the sub-family Phasianinae
(Partridges, Quails, Pheasants) which comprises over fifty genera, most of which
are Palaearctic or Asiatic. My interest for this paper is primarily in the African
members of the genus, but the five Asiatic species are included since they provide

THE FRANCOLINS, A STUDY IN SPECIATION 109

some interesting points of comparison. I am not, however, competent to enter
into a discussion on the limitations of the genus in Asia where there are thirty-
eight other genera in the same sub-family some of which such as Rhizothera, Alectoris,
Perdix, Arboricola seem closely related to Francolinus. They have been dis-
tinguished on various structural characters such as the number of tail-feathers, the
size and shape of the bill, or the length of tail, and I have accepted these distinctions
uncritically. It is sufficient to say that the francolins are characterised by a rather
longer, more hooked bill than members of the other genera (except Rhizothera), a
short tail of fourteen feathers, and an upright stance: in the majority of species the
male, at least, is spurred. The range of colours in the plumage is limited to the
shades of ochre, chestnut, brown, black and white produced by melanic pigments,
but the pattern of individual feathers is complex and varied.

The partridge of Madagascar, Margaroperdix, has also some similarity to the
francolins (particularly to the forest francolin F. lathami) in colour and pattern, but
has various structural differences in bill, tail and legs, which make it doubtful if
the two are very closely related. Apart from this in the Ethiopian region there are
no gamebirds that can be regarded as similar to francolins in either structure or
habits, the only others being the quails (Coturnix and Excalfactoria] , the Stone
Partridge (Ptilopachus) a bird confined to the savanna and steppe belt north of the
Equator the Congo Peacock (Afropavo), confined to the Congo forest, and the
guinea-fowl family (Numididae).

I would suggest therefore that the closer affinities of Francolinus with Palaearctic
and Asiatic genera indicate that it originated in Asia, becoming separated from the
other game-birds as a species particularly adapted to sub-tropical grasslands, and
that later it spread into southern Europe and Africa. Once in Africa lack of
efficient competition allowed it to spread widely and to exploit varied habitats
which, in Asia, are occupied by other related genera.

If this hypothesis of the origin of Francolinus is accepted it is apparent that
there are factors present in Africa that have encouraged speciation in the com-
paratively recent past (since the emergence of the genus). This is readily acceptable
in view of the climatic changes known to have taken place (see Appendix i).
Fluctuations in humidity and temperature would cause redistribution of the
montane and lowland forests, and the woodlands, savannas and steppes, isolating
and rejoining the populations of their indigenous birds.

Among the francolins various generic divisions have been proposed, the most
commonly accepted being Pternistis for the bare-throated francolins (as in Peters',
1934: 84). However, I agree with White (1952) that the bare throat alone is not a
generic character and in other respects these birds are closer to some of the other
large francolins than are other members of the genus. If a generic division was to
be made I would separate the birds of the first five groups (including the Bare-
Throated Group) which have relatively plain or vermiculated backs, from the
birds of the last three groups with striated or quail-type patterning on the backs,
placing F. gularis with the former and F. pondicerianus and F. lathami with the
latter. However, I regard such a division as unnecessary and leading to difficulties
in respect of the rather atypical species F, nahani.

no B. P. HALL

For a study on speciation francolins have the advantage of being remarkably
sedentary birds for their size, not disposed to fly any great distances, and closely
associated with a particular habitat. Morphologically the bold colours and patterns,
particularly on the underparts, have the advantage of indicating where interbreeding
has or has not taken place between neighbouring forms. Finally their " sporting "
and edible qualities have given them an interest to many people other than
ornithologists so that their ranges and habits are probably rather better known than
those of any comparable group: at the same time these qualities have led to a
persecution of the genus over the last fifty years which, combined with the damage
done by intensive agriculture to the natural habitats, has almost exterminated
species from some areas in which they were common and renders it difficult, if not
impossible, to get further information about the habits of species from some areas
which are now " developed ".

THE GROUPS

Grouping the francolins has presented little difficulty, for the close relationship
of most members is evident from their appearance, ecology and distribution and
has been recognised by previous authors. In a few cases where the relationship is
less obvious this is discussed in the text. I have felt that it facilitates discussion
to give descriptive names to the groups rather than numbers, and the names chosen
indicate an important character common to all members though it has not been
possible in all cases to find a character which is exclusive to the group.

The order in which the groups are listed and discussed is not intended to be syste-
matic except that the first five and the last three appear to form related assemblages.

The groups and the species that comprise them are as follows:

1. Spotted Group (francolinus , pictus, pintadeanus) .

2. Bare-throated Group (afer, swainsonii, rufopictus, leucoscepus) .

3. Montane Group (erckelii, ochropectus, castaneicollis , jacksoni, nobilis, camerun-
ensis, swierstrai).

4. Scaly Group (ahantensis, squamatus, griseostriatus) .

5. Vermiculated Group (bicalcaratus, icterorhynchus, clappertoni, hildebrandh
natalensis, hartlaubi, harwoodi, adspersus, capensis).

6. Striated Group (sephaena, streptophorus) .

7. Red-winged Group (psilolaemus , shelleyi, africanus, levaillantoides, levaillan-
tii,finschi).

8. Red-tailed Group (coqui, albogularis, schlegelii).

The four species not assigned to any group are the two African forest francolins
lathami and nahani, and the two Asiatic species pondicerianus and gularis.

SPECIES AND SUBSPECIES

In discussing speciation it must always be borne in mind that this is a continuous
divergent progress starting when two populations become isolated from each other.
In isolation each will develop divergent characters, either morphological or ecological,
or both, but there will be a long period during which the two populations are

THE FRANCOLINS, A STUDY IN SPECIATION in

sufficiently alike for free interbreeding to take place if the isolating barrier is removed
(Phase i). If they remain isolated and divergence continues this stage will
gradually merge into one in which the divergent characters become such that
interbreeding is unlikely (but not impossible) if the two populations rejoin (Phase 2).
Again these characters may be ethological, involving different recognition patterns
or voice ; or ecological, so that the two are unlikely to occur on the same ground ;
or could possibly be a difference in breeding seasons. It is in this stage that the
occasional hybrid may be produced in the wild, and in which hybridisation is
possible and frequent under artificial conditions. Finally the two birds diverge
so completely that they are incapable of interbreeding (Phase 3).

In general most populations in Phase i and in the transition period between
Phase i and 2 would be regarded as subspecies and those in Phase 2 and 3 as species,
but there is no line between them and the decision must rest on the assessment of
the taxonomist. This must be based on the degree of divergence and the degree
to which they are known to interbreed. In the francolins members of different
groups may be assumed to have reached, or almost reached, Phase 3 in their
relationship, but the allopatric members of each group may be only in Phases i and 2.

Each group provides a case in which I have had to make an arbitary decision
whether or not to regard neighbouring forms as conspecific. For example, in the
Spotted Group F. francolinus and F. pictus have diverged appreciably, especially
in males, and hybrids are scarce, so I treat them as species. In the Bare-throated
Group the afer block and the cranchii block have also diverged but interbreed
freely, so are treated as conspecific. In the Montane Group F. ochropectus has
diverged morphologically and ecologically from F. erckelii and is isolated, so is
treated as a species, while atrifrons stands in a similar morphological relationship to
castaneicollis but has the same field habits so is considered conspecific. In the
Scaly Group the Niger provides a barrier between F. ahantensis and F. squamatus
so it is not certain if they are capable of interbreeding but they seem to have diverged
sufficiently to justify specific status. In the Vermiculated Group F. hildebrandti
and F. natalensis have diverged, especially in the females, but interbreed to a
limited extent at one point of contact but not apparently at another so are treated
as species. In the Striated Group " rovuma " type birds (with a streaked abdomen)
apparently do not interbreed with " sephaena " types (unstreaked) in the southern
half of the range but hybridise in the north; the divergence is slight, however, and
they are treated as conspecific. In the Red-winged Group the degree of divergence
and the relationship between shelleyi and its three neighbours, africanus, whytei,
and uluensis varies only slightly, but is just sufficient for me to treat africanus as
a species and the other three as conspecific. In the Red-tailed Group the northern
forms of coqui have diverged slightly, both morphologically and ecologically, from
the southern but seem capable of interbreeding so must be considered conspecific.

The foregoing demonstrates the impossibility of defining a species in such con-
ditions. For this reason I have tried to avoid differentiating strongly between
species and subspecies in the discussion, though greater clarity might have been
achieved by setting out the discussion under specific headings (as has been done in
the Red-winged Group where rearrangement of the species made it necessary).

H2 B. P. HALL

On the other hand I have endeavoured to make a clear distinction between those
subspecies which I term " potential species " namely those which are either
(a) isolated at present, or (6) must be assumed to have been isolated at some period,
since they show a degree of divergence from their neighbours which is too great
to be attributable to local ecological factors and those subspecies whose divergence
can be attributed to ecological factors and which, as a rule, form part of a cline.
The characters and ranges of " potential species " are denned equally with those of
full species, but other subspecies are noted summarily in smaller print in a section
in each group headed " Other variation ", together with any significant local or
individual variation. On the maps as much information as possible has been shown
on the ranges and relationships of species and subspecies, but this has necessarily
been dependent on the complexity of each map. However, I do not consider it
practical or desirable to recognise numerous steps in a cline, or slight local variations,
and have relegated to the synonymy many names given to such intermediates and
micro-populations. There is more justification in naming isolated populations,
however small, such as some of the montane ones, and in these cases I have been
guided by the degree of divergence shown. For instance, I regard names as desirable
for such highly divergent small populations as the montane francolins found on the
Plateau du Day in (French) Somaliland (ochropectus) and at Mega in southern
Abyssinia (atrifrons), but not for others in the same group in other mountains of
southern Abyssinia which differ only slightly from each other. To avoid en-
cumbering the main discussion with irrelevant detail, notes on taxonomy and
synonymy are confined to an appendix.

In the parts of the paper dealing with evolution the ancestral stock of present
species or of more than one species is referred to with the prefix " proto " (e.g.
proto-coqui and pToto-albogularisjschlegelii) : if clarity demands, the name is followed
by " subsp." or " sp.". The name of the author and date of publication are insert-
ed only for those subspecies not listed by Peters (1934).

METHODS

The extensive collection of francolins in the British Museum has formed the basis
for this study, supplemented by loans of critical specimens from other museums.
All the African specimens examined and all the authentic records I have been able
to trace in literature have been plotted over the Vegetation Map of Africa (Oxford
1959), as many species as possible being entered on a single sheet. This has enabled
me to make a direct comparison of the range of any species with the vegetational
belts and with the range of other species, whether in the same group or not. The
Asiatic species have been plotted with accuracy only in the few critical areas, and
for this the distribution maps of Indian birds prepared by Whistler, but never
published, have been of great assistance. The maps of the groups have been compiled
from these key maps. In a distribution map it is always difficult to know to what
extent the range should be shown as continuous between actual collecting points,
and in this respect the maps differ. For the Asiatic species I have shown the

THE FRANCOLINS, A STUDY IN SPECIATION 113

limits of the range only: for the African species I have, as a general rule, shaded
only areas from which specimens have been obtained, though this may give a rather
distorted picture from parts of the country which are least known ornithologically :
in a few cases, most notably for the Bare-throated Francolins (Map 2), I have felt
that greater clarity in presentation could be achieved by shading the whole area
within the limits of the range: this has been done only with species which, from
their known habitat preferences, could be expected to have a reasonably continuous
distribution within a vegetation belt.

The morphological discussion is based on specimens examined personally unless
otherwise stated. For the field notes I have used the standard regional works as a
basis but these have been immeasurably enriched by the numerous conversations
and letters that I have exchanged with the many field workers whose names appear
in the acknowledgments. Where these workers have been responsible for specialised
information I have tried to give the credit in the text, but they are responsible also
for much of the general information which is incorporated without acknowledgment.

I have found it difficult when discussing ecology and habitat to be consistent
in the descriptive terms used, especially in correlating the parts of southern Africa
with which I am familiar with those of the north for which I have to rely on the
descriptions of others. As far as possible I have followed the terminology employed
by the Vegetation Map of Africa (Map n). In this the vegetational types which
mainly concern the francolins are the Woodlands, Savannas and Steppes (Types
16-25). My own experience in southern Africa and the works of such specialists
in ecology as Benson and Stuart Irwin show the importance of the distinction
between the woodlands in which acacia is rare or absent (the " brachystegia wood-
lands " or " myombo " Types 18 and 19) and the acacia country, whether it is
termed acacia " woodland ", " savanna " or " steppe ". This distinction applies
not only to the birds of the actual woodlands but also to those of the grasslands
and savannas (i.e. grasslands with scattered trees) with which the woods are inter-
spersed, many birds being strictly confined to either the brachystegia belt or to
acacia. (In the southern and eastern francolins there are, in fact, several species
indigenous to acacia, such as F. sephaena, F. leucoscepus, F. rufopictus, F. adspersus
and F. levaillantoides , but none which is wholly indigenous to the woodland belt,
though the potential species F. s. shelleyi and F. s. whytei and the Bare-throated
F. afer nearly qualify.) I have therefore made a clear distinction in the discussions
between the brachystegia belt and acacia country and throughout the paper the
term " woodland " is used exclusively for types 17-19 unless " acacia woodland " is
specifically stated.

North of the equatorial forest the woodlands are apparently similar in character
to the brachystegia woodlands though composed largely of different species of trees
and less continuous (Type 17). I am told also that, because of this discontinuity,
the transition from the woodland belt to the acacia is less clearly defined. It is
to be expected that there are therefore less clear-cut ecological distinctions in the
birds, but this is not easy to determine from published works, the term " savanna "
being widely used by different authors without clear definition. I have therefore
been less precise in discussing the habitat of northern francolins. (From the maps it

H 4 B - P- HALL

appears that F. clappertoni and F. coqui spinetorum may be indigenous to acacia, and
F. schlegelii, F. albogularis and -F. icterorhynchus to the woodland belt.)

The distinction between acacia "woodlands", "savannas" and "steppes" is
largely one of aridity and the transition between the types is necessarily gradual.
It will be appreciated from the foregoing that savannas are found in both the
woodland and acacia belts, but they are necessarily different in character. Where
the term is used without qualification it should be plain from the context which belt
is under discussion.

In attempting to interpret in terms of evolution the pattern presented in each
group by divergence, ecology and ranges, I have started in each case with the
distributional maps. From these I have tried to find in each group the factor that
limits the ranges. More often than not the answer lies in the vegetation map, and
the range of a group, species or potential species will be found to coincide with the
limits of a vegetation belt, this coincidence being particularly common where
acacia savanna or steppe changes to woodland. The presence or absence of moun-
tains, hills or rivers is obviously significant in other cases, but here it must be
remembered that, while the reason for the present discontinuity in the range of a
species may be self-evident, the range must have been continuous in the compara-
tively recent past. In one case, the Red-winged Group, the limits of ranges of
species coincide not so much with the vegetational belts as with the isohyets of total
annual rainfall. In another, the Striated Group, the isolated areas from which one
species (F. streptophorus) is found seem to have nothing in common except that they
are the two ends of a ridge of higher country. In several cases where there seems
no ecological reason for the limits of a species' range the bird is replaced by a member
of another group. From facts such as these I have tried to assess in what con-
ditions each group would be most likely to thrive at the expense of other groups,
and, conversely, what conditions would be sufficiently unfavourable to divide it
into isolated populations from which the species developed. In postulating in
each group successive climatic changes to account for the isolating and rejoining
of the various popualtions I have been guided solely by the pattern presented by
the birds themselves, and not by any consideration of what climatic changes are
known from other evidence to have occurred, except that I have not, I hope, postu-
lated any changes which are, on this evidence, outside the bounds of possibility.
However, in the final section of the paper I have tentatively tried to correlate some
of the climatic eras postulated with those known, as listed in Appendix i.

In the discussion on evolution two assumptions are frequently made which are
not necessarily correct but which seem sufficiently probable to justify their use as
a basis for hypothesis. The first is the premise that the degree of divergence shown
by two isolated forms can be correlated with the length of isolation. This, of course,
is not necessarily true, but it can, I think, be accepted that in similar circumstance
there is a likelihood that speciation will proceed at approximately the same pace;
also that a comparatively long interval must elapse for considerable divergence to
take place in isolates. The corollary premise that isolates that have not diverged
extensively have not been long separated is more open to doubt, and examples
(such as the forest owl Phodilus prigoginei] can be quoted of populations that must

THE FRANCOLINS, A STUDY IN SPECIATION 115

have been long isolated but show little divergence. But I regard these as
exceptional.

The second assumption is that the present species and potential species originated
in some part or parts of their present range. This is a premise that is the more
questionable in the case of species which are not members of superspecies. However,
in the francolins these are exceptional, most of the species being part of a super-
species in which all members are allopatric and which has a more or less continuous
distribution over a large part of Africa. In these cases it is highly improbable that
the origin of any member was in the territory of the present neighbouring forms,
although all may have been outside the present range of the group (i.e. when vegeta-
tion belts were pushed north or south).

In such a contingency each proto-species may be presumed to have held approxi-
mately the same position in regard to its neighbours as it does at present.

CHARACTERS

The study of the francolins in species groups has served to demonstrate that many
characters which might be considered as guides to relationship show variable
degrees of uniformity in each group. For example, in the Bare-throated Group
the extent of bare skin on the face and throat is absolutely constant, and can be
regarded as a " group character ", whereas in other groups closely related species
may have a variable amount of bare skin. It seems useful therefore to discuss some
of these characters separately at the start in order to obtain an understanding of
their importance. A summary of some of these characters is set out in Table i.

SIZE

Little emphasis has been put on size or proportions in this paper, for though there
is considerable individual variation there is reasonable uniformity in most related
species. In the Red-winged Group the long-billed and short-billed populations of
F. shelleyi provide the only example of marked proportional variation. On the
whole general variation in size follows Bergmann's Law in relation to altitude and
latitude, equatorial and lowland birds usually being smaller, but there are several
inconsistencies. It is perhaps worth noting especially the incidence (for which I
cannot account) of exceptionally small birds on the coasts of Kenya and southern
Angola, and, to a lesser extent, Natal, since this feature is found in other genera
besides francolins.

PLUMAGE

(a) Feather patterns. The complexity of the feather patterns has already been
referred to. A number of examples are illustrated below which serve to define some
of the descriptive terms I have used, and which have been arranged in sequences to
indicate some of the transitional stages between one pattern and another. This is
a subject which deserves fuller treatment than I can give it here and much of the
discussion by Harrison in his paper " The incidence and origin of spotted patterns
in the Estrildidae " (awaiting publication in the Ibis) is relevant to the francolins.
It is sufficient here to emphasise that comparatively small changes in feather pattern,

B. P. HALL

a s?

O CD
IT) VH

H O

rt ,O

3 o

P X<

.H o

^.22

I S

u
> CD

On H *
fe -

II >

V ~ >

s
S

en g
P .S

en CD

S *

3 CD

4J w ^

o ^

e i^

T3 CD -(->

ft en .0

CD C

> cn -S

<P rt rn

. a3

CD O

3 'S

^ "

3 .3

r^-j CD

en CD XJ

a^
ss s

CD fl T3
O ^j

XH ^ ~

3 CD i
fX, W)

C^ O ft ^ t^.S Xj
^ tuO^ . cd ^^ O

W If t

PH T

I S

w J "

<3 _

Q .* **

J5 co co

< H rf

^88

W 8 S

H^ K
S S

O-^ ~s
o o

X o o

ro O
8 8

o o

o
"3

I
3

to cd "S 5

'o ^)

"{j . ^ c3 O T3 O T3

^ ^

ft'S 5 O "rt S "S S

CD CO

ST ^^ . ^<u^a;

'Sec

"*3

^2 ^ rt >*L c3 5r* C^ ?T*

C 3 S

""ii^ ^^ ni 5i 33

fill -llll

(- IH IH

jsl^l 2 ! 2

! t 2 ^

1 f o

!l!

O _r
Of 3 o

- rt c*

HN Q ^Q

00 CH '
Of Of Qf

H H nT M

2r bo Of "o"
I g - o oi S- 2t

-" T! 3 *" - - H

OH o 5T H M H

o o o

Of Of Of
HN H ^-C.

M HI M

^ 1 "^

*O OT

*O ^O ^O H

S & , *o *o

^ sS- *o *

N (M

III

C o o

*^3

"^ *C

i t t

rX S -S3

CO W3 2n

CD CD

rj CD ~CD~
c3 be to

_O_ d rt

| | | 1

r^> ^

rH CD CD

S bo bo
O o3 o3

^< l-c

T3 O O

C S (3

tn .be

<U T3
^3 li ^ CO

pt
O O O

6 9 H

XJ_ XJ_ Xi

'ts ** "S

ctf cd rt

xi 3 3

XJ T3 (Q rt

2 ^ g

d) u ^^

i 1 3l S

oo - .

O O O T3 T3 H3 i CD

rt rt^o o o bo be
'n'oO* H ^ *~* d

. * 22
d s -

"o " ii tc

J3 T3 'r3 C

3 <i> rt rt

2 3^

2 o . & '

003

."*

"a j^

en c/1 bD
CD CD ^g
>->>-. 05

O O O O
C fl C C

O O O

a a a

v> S

tf)

^ g

J J ^

J J ^ J U 3 3

^ J

to .2

francolinus
pintadecmus
pictus

1 1 : i
s .^ s

1 f 1

I > i |

(/) *-s

IS 1 1 1 -8 1 I

O Q *4 ^* ^i ^

^ "S 2 e S

6 O Q *rv K Q *l

to <o ^

' S '1
111

III

Q (O bo

Q
H

(7)

H
M

THE FRANCOLINS, A^STUDY IN SPECIATION

117

a fr

O c 6C

^ S -S

5 b >>

^
2 5

H 2

s s

totnto to^to^tn

3 S 3 3 .3 3 .2 a

WWW

ptop

3 2

rQ ,0

B. P. HALL

particularly at the tip of the feathers, may greatly alter the appearance of a bird.
(Compare the great differences in the general appearance of the breasts of a female
schlegelii and a male coqui, and in the underparts of a male and female pintadeanus
with the small differences in feathers 2 and i, and 5 and 3 of Figure i.)

(b) Underparts. The greatest variety of colour and pattern throughout the genus
is found on the underparts (and to a lesser degree on the mantle) and considerable
variation is found between closely related birds. The most striking example is
illustrated by Bowen (1930) and shows the difference between the subspecies of

Fig. i. Feather patterns.

i. Barred (c? coqui, breast). 2. Barred with inverted triangle at tip ($ schlegelii,
breast). 3. Transition, barred to spotted ($ pintadeanus, breast). 4. Transition,
barred to spotted (<$ shelleyi, breast). 5. Transition, barred to spotted (<$ pintadeanus,
breast). 6. Spotted (<J francolinus) . 7. Vermiculated (hildebrandti, back). 8. Irreg-
ular double-V-pattern (harwoodi, back). 9. Double-U-pattern (clappertoni, back).
10. U-pattern (clappertoni, back), n. V-pattern (clappertoni, back). 12. Scalloped
(camerunensis , back). 13. Striated (sephaena, back). 14. Quail-patterned (shelleyi,
back). 15. Drop-shaped pattern with windows (bicalcaratus , breast). 16. Streaked
(jacksoni, breast). 17. Drop-shaped pattern (clappertoni, breast). 18. Triangular
pattern at tip (sephaena, breast). 19. Scaly (squamatus, breast).

THE FRANCOLINS, A STUDY IN SPECIATION 119

F. afer with vermiculated underparts which hybridise freely with those with streaked
underparts.

(c) Upperparts. The upperparts are basically either plain, streaked, vermiculated
or " quail-patterned " (Fig. i, 14), sometimes with a more complicated mantle
patterning. While related species may vary between plain, streaked and vermic-
ulated (see Montane Group) " quail-patterning" is more constant throughout the
three groups in which it is found.

(d) Wings and Tail. The wings and tail are, on the whole, not conspicuously
patterned or coloured, exceptions being found in the red wings of the Red-winged
Group and some of the Red-tailed Group, and in the tails of the Red-tailed and
Striated Groups which are reddish and black respectively.

A further exception is the light wing patch under the wing of leucoscepus, and
rather similar light colour in the wings of clappertoni, which are the only members
of their respective groups to have this character. These patches may have functional
significance, for Harrison tells me he has observed the bare-throated leucoscepus
in the London Zoo displaying with its wings carried forward exposing the light patch.
Since other bare-throated francolins have plain wings it would be interesting to
know if they have a different display, and also whether clappertoni and red-winged
birds display in the same way.

(e) Facial pattern. Facial pattern seems to be an important character since it is
the only feature of adult plumage which is recognisable in the chicks (see below), yet
again related species show no constancy. In the Spotted Group two members
have conspicuous facial patterns but pictus is unpatterned (yet hybridises with
francolinus) : in the Red-tailed Group the male coqui is unpatterned but the female
patterned, the two sexes being recognisable in the chicks : in the Red-winged Group
finschi is unpatterned but all other members have a black bridle and necklace,
either well defined or indicated by spotting. Unfortunately chicks of many of the
critical species are not available.

It has proved impossible to include a summary of plumage in Table i, since the
variation in colour and pattern is too complex to put into a few words. In the text
I have tried to give sufficient descriptions to enable the reader to form a picture of
the bird but without going into full detail.

SEXUAL DIMORPHISM

As a rule in francolins the sexes are alike except that the male is more heavily
spurred and the female often slightly more vermiculated, but there is marked
sexual dimorphism in all members of the Red-tailed Group: in two out of nine
members of the Vermiculated Group, one of which hybridises, and in part inter-
grades, with another member in which the sexes are alike: in two out of three
members of the Spotted Group : in two out of seven members of the Montane Group,
and in lathami.

EXTENT OF BARE SKIN

Bare skin on the throat as well as round the eye is found only in members of the
Bare-throated Group (though not in the chicks). Various other species scattered

120 B. P. HALL

through different groups have conspicuous bare patches round the eye, namely
camerunensis and nobilis in the Montane Group, harwoodi in the Vermiculated
Group, and nahani. Others have smaller patches (clappertoni , icterorhynchus, and
adspersus in the Vermiculated Group) or coloured eyelids (the montane jacksoni)
or small coloured patches behind the eye. Unfortunately these smaller patches
have largely been ignored by collectors especially when they are not brightly
coloured and it is difficult to ascertain from skins in which species they are present,
or if colouring is seasonal: to avoid error only conspicuous areas of coloured skin
are therefore mentioned in the table.

COLOUR OF BILL AND LEGS

This is a character that is reasonably constant in half the groups but very variable
in the others, striking variation being found between the three northern species
in the Vermiculated Group which otherwise form a remarkably homogeneous
assemblage.

SPURS

The number of spurs carried by males and females in any species is not constant,
and indeed has been known to differ on the two legs, nevertheless there is sufficient
constancy to justify the term " normal complement ". This varies between species,
from a complete lack of spurs in either sex in pictus of the Spotted Group and nahani
to two fully developed spurs in the males of erckelii in the Montane Group. Blunt,
half-developed spurs are found even in adult birds, often in males as an upper spur,
and occasionally in females of almost all species that are normally without spurs.
In the rare cases where a bird has a normal complement of two fully developed spurs
it seems that the upper is likely to be equal to or longer than the lower, but in the
cases where a bird has one fully developed and one half-developed spur it is the
lower which is fully developed and therefore the longer. Undeveloped spurs are
indicated by swellings and are found normally in fully adult males of the Vermiculated
hartlaubi, and the striated streptophorus. It can be said, therefore, that the number
of spurs has some taxonomic significance, but is not as infallible or as important a
character as some text books lead students to believe.

FIELD HABITS

I have too little personal experience of francolins in the field and too little has
been written, especially of some of the rarer species, for me to make much use of
habits and behaviour as criteria of relationship. There is a certain amount of
information on whether or not the birds are found in coveys, and whether they perch
or roost in trees, but no one group is fully documented. It can be said confidently,
however, that the majority of francolins are found in pairs or small coveys according
to the season, but four species at least are more gregarious: in the Vermiculated
Group bicalcaratus is found in coveys of up to 40 birds (Bannerman, 1930: 321),
and adspersus up to 20 (McLachlan & Liversedge, 1957 : 96) : and in the Red-winged
Group africamis is found in coveys up to 30 birds (Horsburgh, 1912: 47), and

THE FRANCOLINS, A STUDY IN SPECIATION 121

levaillantoides up to 14 (op.cit. : 54). I have found definite records of most members
of the first five groups taking to trees though they do not necessarily use them
regularly for roosting: Hopkinson (1923: 128) found bicalcaratus roosted in trees
only in the rains, and this may well be true of other species. On the other hand I
have no records of any members of the last three groups, the Striated, Red-winged
and Red-tailed, taking to trees even when alarmed and it would seem that these
birds, with their more cryptic (" quail-patterned ") coloration rely for protection
on sitting close.

HABITAT

It will be shown in the discussion on the groups that the majority of the francolins
have very precise ecological requirements, so much so that, though several different
species may be found in the same area, it is rare to find two different species on the
same ground. For instance at Onguati in South West Africa F. adspersus, F.
levaillantoides and F. hartlaubi were found within a few hundred yards of each other,
but adspersus was strictly confined to the riverine vegetation, hartlaubi to the rocky
kopjes and levaillantoides to the grasslands between (Macdonald, 1957: 52). Similar
instances are found throughout southern and eastern Africa where the ranges of
six groups overlap but there are also exceptions. Pitman has found coqui,
levaillantii and afer on the same ground in Uganda: Elliot also has found coqui
with other species in Tanganyika and Blasdale has found it with bicalcaratus in
Nigeria (but not with clappertoni) : Benson and Stuart-Irwin have found natalensis
and sephaena together in riverine thickets in both Northern and Southern Rhodesia,
and Stuart-Irwin tells me that sephaena is also found with adspersus in Ngamiland,
though in Southern Rhodesia the two would never occur together: adspersus also
occurs with swainsonii, sometimes in the same flocks, in Ngamiland. All these
examples are of members of different groups occurring together but hybrid specimens
show that members of the same group may mix where they have a common boundary,
as indeed would be expected.

In any one group the different species may all share the same, or very similar,
ecological preferences, as in the Spotted and Scaly Groups, or different species may
have different preferences, as for example in the Red-tailed and Bare-throated
Groups. There are also individual species which show a wide choice of habitat in
different parts of the range, most notably the red-winged F. shelleyi and the red-
tailed F. coqui, and similar inconsistency is shown in the altitude at which some
species are found, especially among all members of the Red-winged Group (except
the Montane Redwing, F. psilolaemus) and in F. squamatus of the Scaly Group.

VOICE

It would seem that voice should be a guide to relationship in the francolins but
the few direct comparisons that have been made between species indicate that
there are unexpected inconsistencies and affinities. Benson, for instance, records a
remarkable difference in the call of sephaena in Abyssinia and Rhodesia (1958: 191),
yet considerable uniformity in both levaillantii and coqui between the calls of north

ZOOL. 10, 2 9

122 B. P. HALL

and central African birds (1948 : 52). He tells me also that there is a marked family
resemblance in the voice of three members of the Red-winged Group, levaillantii,
shelleyi and the Abyssinian levaillantoides, and in three members of the Bare-throated
Group, afer, swainsonii and leucoscepus. Similarities between members of the
Montane and Scaly Groups respectively have been recorded by Hall (i96oa: 408) and
Collier (1935:666). Furthermore Chapin (1932: 715) found the call of the bare-
throated afer indistinguishable from that of icterorhynchus of the Vermiculated Group.
On the other hand Blasdale tells me that the red-tailed coqui and albogularis in
Nigeria have quite different calls.

CHICKS

I have only been able to examine chicks of less than half the species and it is only
possible therefore to draw a few general conclusions. The most striking variation
is shown in the pattern of the crown and face, but in this respect there is little
difference between the representatives examined of the Bare-throated, Montane,
Scaly or Vermiculated Groups, all of which have a solid brown crown (sometimes
edged black) and a single conspicuous brown or blackish stripe from the bill through
the eye to the neck. In addition the only chick examined of the Montane Group
(erckelii) has a short black stripe below the eye. The chicks of the Red- winged
Group are distinctive in having a narrow and darker crown patch, bordered on each
side by eight alternate white and black stripes over, through and under the eye, the
forerunner of the adult bridle and necklace. Unfortunately no chick has been
examined of F. finschi, the only member of the group without any black-and-white
bridle or necklace, and none of F, psilolaemus in which the bridle is spotted and
indistinct. The female coqui of the Red-tailed Group has a similar pattern to the
red-winged chicks though less clearly denned, but the male chick is like those of the
other groups, though with the brown of the crown extending further forward to
the beak: the chicks therefore bring out the sexual dimorphism of the adults. The
adults of the Spotted Group have, like those of the Red-winged Group, a very
striking facial pattern, and the chicks have a similar narrow crown patch with
alternate stripes on either side, but the pattern is paler than in the Red-wings and
ill-defined. The chick of the ungrouped pondicerianus has an indefinite pattern
intermediate between both main varieties, the dark stripes above the eye being
incomplete and ill-defined. In view of the small variation evident between the
chicks of many species and even groups, it is interesting to note that subspecific
variation is apparent between the paler and darker races of F.francolinus, and
between two " potential species " of the Red-winged Group F. s. shelleyi and
F. s. whytei.

EGGS

The British Museum has a reasonably representative collection of francolin eggs,
including those of at least one member of every group, but as a guide to relationship
they are disappointing. All are rather broad, and are plain or lightly freckled,
ranging in colour from white to light brown. Large series of eggs of both F.

THE FRANCOLINS, A STUDY IN SPECIATION 123

bicalcaratus and F. francolinus show there is considerable variation in colour within a
species, largely correlated with the darker and paler subspecies living in wetter or
drier conditions. There is also some variation in shape. The eggs of the two
African forest francolins lathami and nahani are the most distinctive, those of
lathami (represented by two clutches from the Cameroons) being a plain reddish
brown and more elongated than any others, and those of nahani (represented by
one clutch from Uganda) being the most heavily freckled, brown on a buff ground.
Pitman tells me that the eggs of sephaena are also distinctive in having an excep-
tionally hard shell.

THE SPOTTED GROUP (MAP i)
(F. francolinus, F. pictus, F. pintadeanus)

RANGE AND CHARACTERS

The Spotted Group is the only Asiatic group and is formed by three of the five
Asiatic species. It ranges from south China and the Indo-Chinese countries through
Thailand, Burma, the Indian sub-continent, Ceylon, Persia, Iraq to the eastern
Mediterranean and Cyprus. There is evidence that it has also inhabited many of
the Mediterranean countries and islands in historical times (Lilford, 1862, Lavauden,
1936). Its members are birds of grassland and scrub jungle, and are found also in
cultivations where there is sufficient cover. They occur up to 5,000 ft. but commonly
prefer lower altitudes.

They are mostly medium-sized, with the mantle either spotted or V-patterned
in brown and buff and the back and tail narrowly barred in black and white, or
buff and brown: the wings are barred or spotted. Below they are basically black
or dark brown with white spots, but in the females of two of the species (F. francolinus
and F. pintadeanus) these spots are merged together and elongated to form irregular
bars. The under tailco verts are maroon. The bills are black or dark brown and
the legs vary from yellow to red-brown. Spurs are poorly developed, the normal
complement for the males of F. francolinus and F. pintadeanus being one only, and
that rather short and blunt, while F. pictus is without spurs in either sex.

SPECIES AND POTENTIAL SPECIES

The three recognised species are the only mainland forms with which we need be
concerned for other variation is clinal and ecological and does not suggest that any
smaller populations have been isolated for any length of time from the main body.
The species are quite distinct: the western F '. francolinus has marked sexual
dimorphism, the male having the throat, breast and sides of face largely black,
with sparse white spots on the side of the breast, and white cheeks : it has a complete
broad maroon collar and some black feathers with white spots on the upper mantle ;
the rest of the mantle and wing coverts are V-patterned in buff, brown and
occasionally black.

The female has no black, the throat being whitish, the sides of the face mottled
brown and buff, and the underparts irregularly barred in brown and off-white.

I2 4 B. P. HALL

The upperparts are patterned like those of the male but without black and with the
maroon collar replaced by a small rufous patch on the hind neck.

The range of F.francolinus extends eastwards to Manipur. It is replaced in
southern India by F. pictus, a smaller bird with little sexual dimorphism, but in
females there is a tendency for the patterning on the abdomen to merge into bars
and this seems most pronounced and most constant in the Ceylon population
(watsoni) . The face and throat of F. pictus are unpatterned and rich ochre in colour :
the upperparts are similar to those of the female F.francolinus but with more
spotting in the mantle and no rufous collar patch : the underparts are intermediate
between the male and female oi francolinus being basically black, or very dark brown,
regularly covered all over with large buff spots.

Where the ranges of the F.francolinus and F. pictus meet in Rajasthan (Rajputana) ,
Uttar Pradesh (United Provinces) and Bihar there is no natural barrier between
them, and no ecological distinctions have been noted. It is evident that they do
not interbreed freely but do so occasionally for there are four specimens in the
British Museum which appear to be hybrids: these are from Deesa (the type of
F. intermedius) , " Neemsar " (an unplaced locality probably, from the history of
the collector, near Fateghur) and " Bihar " (no detailed locality).

Ceylon birds are not very different from those of the mainland, but are rather
darker and more closely patterned, with the tendency (already noted) for the
females to be slightly barred on the abdomens.

East of Manipur F. francolinus is replaced by F. pintadeanus which, in both sexes,
differs from the other two species in having the throat and cheeks white, sharply
divided by a black line, and having the mantle largely spotted, not V-patterned,
and with a considerable amount of rufous, especially in the male. Below the male
is patterned all over like F. pictus, but with the spots pure white on a deep black
ground ; the female is barred like the female of F. francolinus but with the barring
narrower and more sharply contrasted.

The higher ranges of Manipur and the Chin Hills seem to form an effective barrier
along most of the boundary between F. pintadeanus and F. francolinus though they
might be expected to meet along the coastal strip. I know of no records of
hybridisation between them.

OTHER VARIATION

Variation is clinal. In F.francolinus western birds are large and dark (francolinus), grading
through arabistanicus of Persia and Iraq to the small, pale, grey birds of Baluchistan and
western India (henrici) . Birds of north-eastern India are darker again (asiae) grading into the
very dark melanotus of Assam.

In F. pictus there is a pale population in the drier north-western areas (pallidus) and other
mainland birds (pictus) are intermediate between pallidus and the darker Ceylon birds (watsoni) .

In F. pintadeanus birds from southern China and northern Indo-China are paler, less heavily
patterned than those of Burma and Thailand (phayrei).

EVOLUTION

The Spotted Group presents a straightforward pattern of three allopatric species
with a continuous distribution across sub-tropical Asia. To interpret this pattern

THE FRANCOLINS, STUDY IN SPECIATION 125

it is only necessary to visualise less favourable conditions in the northern parts of
the present range which would serve to isolate the three species from each other.

I suggest therefore that the three species have arisen from a single species
occupying the present range of the whole group, and possibly much of southern
Europe as well. In an unfavourable era, such as a Glaciation, conditions in the
northern parts of the range became unsuitable and proto-francolinus was isolated
in Palestine and possibly Arabia (if the era was wet as well as cold) : proto-pictus
was isolated in southern India and proto-pintadeanus was isolated in south-eastern
Asia (Map lA). With the return of favourable conditions the three species spread
north again, proto-francolinus invading India from the west and establishing itself
north of proto-pictus. In the period of isolation speciation had developed
sufficiently for the three species not to interbreed freely when they rejoined, but the
presence of hybrids on the boundary between francolinus and pictus indicates that
genetic isolation is not absolute.

A summary of the evolution postulated is, therefore, as follows:
Stage i. The ancestral form of the group widespread, covering present range and
possibly southern Europe as well.

Stage 2. An unfavourable era, probably a Glaciation, pushes the range southwards
isolating the ancestral stock of the three species in south-western Asia, southern
India, and south-eastern Asia (Map IA).

Stage 3. With the return of more favourable conditions the group spreads north-
wards again, but speciation had advanced sufficiently in the three isolated blocks
to inhibit free interbreeding when they rejoin.

THE BARE-THROATED GROUP (MAP 2)
(F. afer, F. swainsonii, F. rufopictus,, F. leucoscepus}

RANGE AND CHARACTERS

The Bare-throated Francolins are widespread in eastern Africa from Eritrea to
Cape Province and westwards, south of the Congo forest, to Gabon, Angola and
northern South West Africa. They are found at comparatively low altitudes,
usually near water-courses or in grasslands with cover of trees and scrub in which
to hide.

They are large francolins distinguished by having a conspicuous patch of bare
skin on the throat and round the eye, varying in colour from scarlet to yellow.
The bill and legs vary from red to black (never yellow): spurs are fairly well
developed, the males of all species having normally a long and strong lower spur;
a short blunt upper spur is found commonly in two of the species (F. leucoscepus
and rufopictus}, less often in F. afer and rarely in F. swainsonii. In the group as a
whole the upper parts are plain brown from the crown to the tail, with some darker
centres to the feathers of the mantle and a few light vermiculations, but in one
species (leucoscepus}, the mantle feathers have white shaft streaks, and in another
(rufopictus} they have broad rufous edges. The underparts are, on the whole,
streaked, but vary widely in colour and in the amount of vermiculation. The
sexes are alike except that the females tend to be slightly more vermiculated.

126 B. P. HALL

SPECIES AND POTENTIAL SPECIES

The Bare-throated Group is a far more complicated assemblage of forms than the
distribution of the four species suggests, for variation in colour and pattern among
the populations of F. afer shows it has had a turbulent history and that at some time
at least six of its subspecies have been isolated for a sufficient period for them to
have diverged considerably.

The species themselves are largely allopatric and ecologically segregated, over-
lapping only in a few marginal areas. The most distinctive member of the group
(F. leucoscepus) is confined to the acacia steppe of the north-east (Map n, Type 25).
In leucoscepus the bare skin is red round the eye, yellow on the throat, the bill
black with a red base and the legs black : the mantle has white shaft-streaks and the
underparts are irregularly streaked all over with plain brown and white, each feather
being brown with narrow white edges and a triangular white patch at the tip,
tapering up the shaft: there is a faint rufous tinge in the brown parts of many
feathers. The primaries have some white on the inner wings which forms a con-
spicuous patch when the wing is displayed: there is no comparable patch in other
members of the group.

F. leucoscepus is replaced by another species, F. rufopictus, in a small area of
Tanganyika stretching from the south-eastern shore of Lake Victoria to the
Wembere. 1 This area is still within the dry belt of acacia steppe but differs from
typical leucoscepus country in the types of Acacia species present (Stronach in litt.},
so it may possibly form an ecological island particularly suited to rufopictus.

In rufopictus the bare skin is coral red, the bill is red and the legs brown: the
feathers of the upperparts are basically grey-brown with dark vermiculations and
dark shaft-streaks, with the feathers of the mantle and wing coverts broadly edged
with chestnut: the breast is mainly grey with black shaft-streaks and the rest of
the upperparts streaked black, white and chestnut; the eye-stripe and sides of the
face are black and white.

In southern Africa there is a third species (F. swainsonii} associated chiefly with
acacia country, ranging in the acacia/mopane savanna belt (Map n, Types 20
and 22) from South West Africa to Mozambique and extending up the Loangwa
valley to the Nyasaland border. In Southern Rhodesia, however, it is found
scattered through the grasslands of the woodland belt, eastwards to the Mashonaland
plateau.

In F. swainsonii the bare skin is red, the bill black above and orange below, the
legs black, with the single spur in the males very long and strong : both above and
below it is a pale brown with some darker shaft-streaks; it has a faint grey wash
over the breast and chestnut streaking on the abdomen. Populations of the
Transvaal and southern Southern Rhodesia (F. s. swainsonii) have a blackish mottled
patch on the abdomen.

The fourth species, F. afer, inhabits rather wetter areas than the other three.
Throughout central Africa its range coincides with the extent of the woodlands,

1 I can find no confirmation, either from museums or from field workers, of the statement by Van
Someren (1925:99) that F. rufopictus occurs in Ruanda, No actual specimens or localities were
quoted,

THE FRANCOLINS, A STUDY IN SPECIATION 127

except in western Southern Rhodesia where these are occupied by F. swainsonii
(see above), and in southern Angola where it has intruded into a small area of the
acacia belt. In South Africa there are two isolated populations in the south-
eastern Transvaal and in the coastal strip of southern Natal and Cape Province.

In F. afer the areas of bare skin, the bill and the legs are red : the back is plain
brown with some darker feather centres. The sides of the face and the underparts
show considerable variation and two diverse blocks of subspecies are easily
recognised, separated by two areas of hybridisation. The first block comprises all
the populations of the southern Congo, northern Angola, Northern Rhodesia east
to the Muchinga Escarpment, extreme western Tanganyika, Uganda and the
shores of Lake Victoria. These birds are heavily vermiculated on the underparts,
with sparse chestnut streaks on the abdomen : the feathers of the face are minutely
patterned in black and grey. I can recognise no significant geographical variation
throughout this area and all populations may be referred to F. a. cranchn, except
for some very dark birds in the Ruzizi valley, north of Lake Tanganyika (harterti)
in which the streaks on the abdomen are maroon rather than chestnut; also the
populations bordering on the hybrid zones in Northern Rhodesia and Tanganyika
have a mixture of white in the abdomen (F. a. intercedens] .

In the second block the birds have no vermiculation and are strongly patterned in
black, white and grey on the face and underparts, and only the subspecies of south-
eastern Cape Province has any chestnut or maroon. The black-and-white sub-
species have a broken distribution, one, F. a. afer, being restricted to the south-
western corner of Angola along the Escarpment and in the Cunene basin. In
F. a. afer the face is white and the underparts broadly streaked in white and black,
the feathers having black centres and white edges.

The rest of the block is found in the eastern half of the country and comprises
an assemblage of subspecies north of the Limpopo and two isolated populations in
South Africa. The northern birds all have the feathers of the breast grey with
black shaft streaks contrasting with the feathers of the abdomen, which are mainly
plain black, and those of the flanks which are streaked black and white. Birds of
the coastal population of Kenya (leucoparaeus} have the eyestripe black and white
and the sides of the face white. They are replaced in northern Tanganyika by birds
with a wholly black face, and similar birds are found south to southern Nyasaland
and extreme south-eastern Northern Rhodesia (melanogaster and loangwae}. There
is no information as to whether or not there is intergradation between leucoparaeus
and melanogaster but birds typical of these forms are found within no miles of each
other at Mombasa and Korogwe respectively.

Birds from the eastern districts of Southern Rhodesia and southern Portuguese
East Africa have the whole face and a necklace above the breast white (swynnertoni) .
In the lower Zambezi valley, the Salisbury area of Southern Rhodesia, and in
southern Nyasaland, between the ranges of the black-faced and white-faced forms,
there is an unstable population with a variable amount of black and white in the
face (humboldtii) ,

Birds of the isolated population in Natal and Cape Province are again differently
patterned, having the face wholly black, and the underparts from the upper breast

128 B. P. HALL

to the abdomen streaked with white, maroon and black, the white lying in parallel
streaks on either side of the shaft (castaneiventer) . These birds intergrade in
the region of Uitenhage with the populations of the southern Cape in which the
maroon is wholly replaced with black (notatus}.

In Natal Vincent tells me that castaneiventer is found only between 2,000 and
4,000 ft., usually on the edges of evergreen forest. It is absent from northern
Natal and the coastal belt (where its place is possibly taken by F. natalensis of the
Vermiculated Group). There is, however, an isolated population in the south-
western Transvaal (lehmanni) which is intermediate in some respects between the
castaneiventer Inotatus forms and swynnertoni, having the face black and the abdomen
with long parallel white streaks as in notatus, but the whole breast grey, with black
shaft streaks, contrasting with the abdomen as in swynnertoni.

Between the block of vermiculated and chestnut forms in the north and the
blocks of black-and-white forms in the east and in southern Angola there are two
hybrid zones, one stretching from Kondoa and Dodoma in central Tanganyika
through central Nyasaland into the Luangwa valley, the other through northern
and central Angola into extreme north-western Rhodesia. The hybrid populations
of Tanganyika and Nyasaland have well-defined streaks on the abdomen varying
in the proportion of chestnut to black-and-white according to the proximity of the
respective parent forms, but most birds from the same area are reasonably constant
and can be distinguished from birds of other local populations. This does not seem
to be true in Angola where micro-populations show greater variability, and the
streaking on the abdomen is less regular and less defined.

In Kenya and Tanganyika F. afer replaces F. rufopictus and F. leucoscepus without
apparently any overlap, but its relations with F. swainsonii in southern Africa vary.
In the Luangwa valley, where both occur, F. afer is found chiefly in the riverine
vegetation and F. swainsonii in the acacia and mopane (Benson and White, 1957 :
23) and the same applies in the Namwala, Mumbwa and Lusaka districts (Benson).
The same is probably true in the Zambezi and Limpopo valleys. However, in
Southern Rhodesia where F. swainsonii has intruded into the woodland belt the two
come together without any ecological segregation, so that both occur on the same
ground. 1 In this area two birds have been collected which show characters of both
species, one from near Salisbury (the type of Pternistis cooperi, Roberts, 1947) and
the second from Hartley's Farm, Rusape, where both species are found (Smithers,
in litt.}. Other specimens of F. afer from the same farm have been noted to have
dark brown instead of the normal red legs. This suggests strongly that some
hybridisation takes place along this border.

There is a record of a third specimen which shows characters of both species
(Markus, 1959), but this was collected at Francistown, three hundred miles from the
nearest race of F. afer, or from any suitable habitat. It was unfortunately not
preserved, but Markus (in litt.} described it as having a red bill and legs, no white on
the face, but both white and chestnut markings on the abdomen, the white markings
being like those of F. a. lehmanni and the chestnut like those of F. swainsonii.

1 Smithers tells me that, with the clearing of land for farming, F. swainsonii has entirely replaced F.
afer in parts of the Salisbury area within the last fifty years.

THE FRANCOLINS, A STUDY IN SPECIATION 129

Since it is unlikely that there is an isolated population of F. afer in country that is
both unsuitable and reasonably well known ornithologically, it must be regarded as
an aberrant specimen of F. swainsonii, emphasising the close relationship of afer
and swainsonii.

The exact ranges of F. afer and F. swainsonii in southern Angola are not known,
nor is there any information on their relationship.

OTHER VARIATION

In F. leucoscepus variation is slight. Birds of the coastal areas of Eritrea and western
Somaliland (F. 1. leucoscepus) have rather more white on the feathers of the underparts than
birds from other parts of the range (F. I. infuscatus), but elsewhere variation in the general
darkness or paleness and degree of patterning of populations is sporadic in local populations.

Apart from the dark birds of the Transvaal and southern Southern Rhodesia (F. s. swainsonii)
which have a mottled black patch on the abdomen, the variation in this species is linked with
ecology. Eastern birds (lundazi White, 1947) are rather browner, less grey than those of the
semi-desert areas of northern Bechuanaland and northern South West Africa (gilli) while
another rather browner population may be recognisable from the Waterburg area (damarensis) .
(Hall, 1956: 100, Benson & White, 1957: 23, 138). Southern birds average larger.

In the stable populations of F. afer there is little variation that has not yet been discussed,
but in the black-bellied, black-faced populations there is some clinal variation, birds of
Tanganyika (melanogaster) being paler and less brown than those of Nyasaland and Northern
Rhodesia (loangwae Grant & Praed, 1934).

EVOLUTION

The pattern presented by the Bare-throated Group is complex, for not only is
there evidence of considerable hybridisation taking place at present but the
morphological characters of the different species, and the degree and type of
variation among the subspecies, suggest incidences of past hybridisation as well.
A simple sequence of evolutionary stages is therefore difficult to envisage and it is
perhaps best to start by discussing the features of the pattern that can be readily
interpreted. It is apparent, for instance, that the present is a favourable era for
F. afer since the hybrid zones indicate that it now occupies areas from which it has
recently been absent. The diverse characters of the afer subspecies point to a
lengthy period in which the two blocks of vermiculated and black-and-white sub-
species were isolated from each other. Since neither apparently developed
ecological preferences in that period they were able to interbreed when the ranges
rejoined. Similarly the variation between many of the black-and-white populations
points to periods in which they have been isolated.

Comparison between Map 2 and the vegetation map shows that the ends of the
hybrid corridor through the Luangwa valley and central Tanganyika coincide with
the intrusive tongues of acacia into the woodland belt. This suggests that the two
afer blocks were divided from each other in a dry era when this corridor was wholly
covered by acacia and occupied by the acacia species. It seems probable that
dry conditions prevailed at the same time in most of Angola eliminating afer from
the present hybrid zone and isolating the small population in the south in the region
of the Angola escarpment where moister conditions might still be expected through
proximity to the sea (see Hall, 1960). Similar dry conditions would serve also to
split the eastern populations.

I 3 o B. P. HALL

It is more difficult to visualise the conditions in which the four species were originally
isolated from each other, and particularly so in the case of afer and swainsonii,
which now occupy neighbouring vegetational zones with a common boundary of
10,000 miles. These two species seem to be the most closely related since they
are alike above and in the colour of the bare skin; furthermore they apparently
hybridise now in one limited area and the dark patch on the abdomen of F. s.
swainsonii may be an indication of more extensive hybridisation in the past. They
may therefore be assumed to have separated most recently. Conversely the
distinctive F. leucoscepus was probably the first to be split from the ancestral stock
of the group. F. rufopictus in some ways combines the characters of the other
three species, having vermiculated, chestnut, white and black patterning in the
feathers and having the bare skin orange, rather than yellow or red. It is possible
that it may represent a relict population approximating more closely to the ancestral
form of the whole group, or else be a product of past hybridisation between species
that has become stabilised and established itself in an ecological island. Both these
alternatives are very speculative and in selecting the latter as slightly more probable
I have been guided by the resemblance of rufopictus on the underparts to some of
the hybrids between the main blocks of afer subspecies. Its present distinctiveness
and ability to live alongside both leucoscepus and afer without interbreeding suggest
an early origin, probably before afer and swainsonii diverged.

I would postulate therefore that the bare-throated francolins were originally
birds of acacia and that they were first split by a spread of woodland or forest into
a north-eastern population (pioto-leucoscepus) and a southern population (proto-
afer / swainsonii). Later the two rejoined briefly and hybridised forming proto-
rufopictus, a relict population which was left in Tanganyika in a subsequent
wetter era when leucoscepus withdrew further north and proto-aferj 'swainsonii
further south.

It is necessary now only to envisage some circumstances in which afer and
swainsonii could develop. I can suggest no likely vegetational changes that would
form a west-east barrier on similar lines to the present boundary, and conclude
that they were split originally into an eastern population (proto-afer) and western
population (proto-swainsonii) divided by a barrier of forest through Natal and the
eastern districts of the Transvaal and Southern Rhodesia. In isolation proto-afer
became adapted to wetter conditions and independent of acacia so that in a sub-
sequent period was able to spread across to Angola through the woodland belt.
It may have hybridised with swainsonii to a limited extent in the east but, for the
most part, was ecologically segregated. It may also have hybridised to a limited
extent with rufopictus, giving northern birds a higher proportion of chestnut and
vermiculations in the feathers than the southern. (This is a possible way to account
for the different directions in which the northern and southern blocks of afer
diverged.) No subsequent changes affected the status of the four species and only
F. afer was subject to further splitting and rejoining, though the ranges of the others
would have spread and contracted in alternate dry and wet spells. It has already
been suggested that a dry spell with a spread of acacia was responsible for isolating
the two main blocks of afer subspecies and also the small population in southern

THE FRANCOLINS, A STUDY IN SPECIATION 131

Angola. This diverged from the northern block along much the same lines as the
southern block in the east, with black-and-white pattern predominating. Sub-
sequently minor fluctuations would serve to isolate the various eastern populations
of afer producing the subspecies leucopareus, melanogaster, swynnertoni, castaneiventer
and notatus. In South Africa the population of the Transvaal (lehmanni), inter-
mediate between swynnertoni and castaneiventer, may have once formed part of
cline between them but has been isolated from both, possibly by the intrusion of
F. natalensis of the Vermiculated Group.

A summary of the evolution postulated is therefore as follows :
Stage i. The group widespread in acacia.

Stage 2. The group split by a spread of forest, proto-leucoscepus isolated in the dry
north-east, proto-afer/swainsonii in the south.

Stage 3. Less humid: the two populations rejoin briefly and hybridise producing
pioto-rufopictus. Proto-afer/swainsonii spreads across the southern part of the
continent.

Stage 4. Proto-swainsonii in the west is separated from proto-afer in the east by a
barrier of forest. In isolation proto-afer becomes adapted to wetter conditions
(Map 2A).

Stage 5. Proto-afer spreads across to Angola through the woodland belt, possibly
hybridising to a limited extent with both swainsonii in the south-east and rufopictus
in the north.

Stage 6. In dry conditions afer is divided into two main blocks, one in the north and
one in the south-east, with a small population isolated in southern Angola. In
isolation the northern block develops distinctive chestnut and vermiculated
patterning below, and the two southern populations a black-and-white patterning
(Map 2B).

Stage 7. Minor fluctuations isolate the various eastern populations of afer.
Stage 8. A favourable era for the group as a whole and particularly for afer, the
vermiculated and the black-and-white blocks of subspecies rejoining and inter-
breeding forming hybrid zones in the east and west. Among the eastern black-and-
white subspecies the white-browed swynnertoni and the black-browed melanogaster
meet and hybridise in the lower Zambezi valley.

THE MONTANE GROUP (MAP 3)

(F. erckelii, F. ochropectus, F. castaneicollis, F. jacksoni, F. nobilis, F. camemnensis,

F. swierstrai]
RANGE AND CHARACTERS

The Montane Group, as its name implies, is formed of scattered communities in
the mountains of north-eastern Africa from Eritrea to Mount Kenya, on the eastern
Congo border, in the highlands of Angola, and on Cameroon Mountain. Its members
are birds which require some trees for cover and roosting and the majority of species
are found in or near montane evergreen forest: only members of the northern
species (F. erckelii} are less associated with forest being found on the grasslands of
the high plateaux of northern Abyssinia where there are neighbouring woods.

In appearance this is the least homogeneous of the groups, so much so that it is

132 B. P. HALL

impossible to designate any " group " character other than that the males have the
crown, lower back, primaries and tail plain brown or red-brown, and the females,
in the species in which they are unlike the males, have the primaries, lower back and
tail vermiculated. However, variation in the other characters seems to follow
geographical trends, the birds of the extreme north-east being the largest and most
heavily spurred with dark bills, yellowish legs, no bare skin round the eye and with
the sexes alike, or only very slightly differentiated by a tendency for the females to
have some vermiculations on the wing and tail. Birds of the two isolated western
populations are the smallest, least heavily spurred with red bills and legs, and with
the sexes quite unlike: in the Cameroon population there is an extensive area of
red bare skin round the eye. The central species have some characters in common
with one or both of their nearest neighbours so that all are linked. The variation
in the colour and pattern of the underparts is considerable, but no more extreme than
that found in one species (F. afer) of the Bare-throated Group. Taking these factors
into account, together with the affinities of other birds and plants of the same
montane communities, I am confident that it is correct to regard the montane
francolins as a superspecies.

SPECIES AND POTENTIAL SPECIES

F. erckelii is the northern representative of the group and has a fairly continuous
distribution in the grasslands of the Abyssinian massif from the Addis Ababa area
northwards to southern Eritrea. An isolated population (pentoni) is found in the
Red Sea Hills at Erkowit, which differs only in being rather greyer. F. erckelii is
rarely found below 6,000 ft. and ranges up to 10,000 ft. It appears to be less
dependent on the presence of evergreen forest than other members of the group,
which would account for its more continuous distribution, and this continuity is
reflected in the birds themselves, which are remarkably uniform in colour and
pattern.

It is the largest of the francolins, with a black forehead and eyestripe, chestnut
crown, the mantle and breast largely grey with maroon streaks and the rest of the
underparts white with maroon streaks : the sexes are alike : the bill is black and the
legs yellowish, the male having two spurs of which the upper is the longer.

On the Plateau du Day in (French) Somaliland there is an isolated population
(F. ochropectus) living in the juniper forest above 3,500 ft. (Dorst & Jouanin, 1950).
It is similar to F. erckelii in many respects of pattern but is dull in colour having
the maroon and chestnut replaced by ochre or brown, and it has some U-patterning
on the lower mantle : the bill is blackish with some yellow on the lower mandible :
the legs are similar in colour and spurs to F. erckelii. It is in fact intermediate in
many respects between erckelii and F. castaneicollis of southern Abyssinia but
possibly sufficiently distinct from both to be granted specific rank.

The nearest population of -F. castaneicollis is found within 150 miles of F.
ochropectus in the mountains of western (British) Somaliland. Other populations
are found in the massif of the Arussi plateau in Abyssinia east of the Rift, and in
the montane " islands " over 6,000 ft. in southern and south-western Abyssinia.

THE FRANCOLINS, A STUDY IN SPECIATION 133

It is apparently more dependent than F. erckelii on cover such as jumper and other
evergreen forest and bamboo and, since forested patches over 6,000 ft. may be
widely separated by unsuitable country, the range is very broken. In conditions
such as these divergence between populations is to be expected and the species is,
in fact, very variable, though individuals from the same populations are reasonably
uniform allowing for the complex patterning of the feathers.

The species as a whole differs from erckelii and ochropectus in having a red bill and
legs, the upper spur equal to or shorter than the lower. It has less black on the face
and forehead, some U-patterning on the feathers of the mantle and breast and the
abdomen largely white. In all the birds of the Arussi plateau south to Alghe
(F. c. castaneicollis) this U-patterning is extensive on the back, wing-coverts and
breast and clearly denned in black and white, mixed with some ochre and chestnut :
Somali birds are similar but rather greyer (ogoensis). Birds from all the isolated
populations west of Lake Zwai vary slightly (but may be grouped for convenience
as F. c. kaffanus) but the U-patterning in all is less well denned and less extensive
and the ochre is largely replaced by rich chestnut. An isolated population at Mega
(atrifrons), less than 50 miles from Alghe where F. c. castaneicollis occurs, is quite
unlike, being without any strong colour or pattern, the U-patterning being only
faintly defined in brown and buff, and the throat and abdomen cream instead of
white. This, like ochropectus, is a population whose taxonomic rank is questionable
but Benson (1945: 393) found it was similar in habits, environment and voice to
other forms of F. castaneicollis and it is therefore perhaps best regarded as a sub-
species in spite of its very distinctive appearance.

There are unaccountably no representatives of the group in the juniper forests
at Yavello (Benson, op. cit.}, or apparently on the mountains of the southern Sudan,
Uganda and northern Kenya (though they may possibly have been overlooked on
the lesser known mountains, and unidentified francolins have been recorded from
Mt. Kulal and Mt. Nyeri at the southern end of Lake Rudolf) but in the Aberdare
Highlands and on Mount Kenya F. jacksoni is common near the edge of the forest
and in the bamboo above 7,500 ft. It has also been found twice west of the Rift
on the Mau Escarpment and in the Cherangani Mountains. It is very similar to
the more chestnut and less heavily patterned populations of F. castaneicollis found
in the west, but has the wings, back and tail a brighter more rufous brown, and the
feathers of the upper mantle, breast and abdomen chestnut edged with white, with
no black and white U-patterning. F. jacksoni has a scarlet eyelid : males usually
have two spurs with the upper poorly developed. Birds of Mount Kenya (F. j.
pollenorum Meinertzhagen, 1937) are rather darker than western birds (F.j.
jacksoni), and in both populations there is individual variation in the extent of white
below.

F. nobilis of Ruwenzori and the mountains from Kivu north to Lake Albert is
rather smaller than the foregoing species and less patterned than any except the
Mega population, jp. c. atrifrons. It has the head, primaries and tail grey-brown
and the mantle and wing-coverts deep maroon, with faint grey scalloping on the
upper mantle : the throat is off-white and the rest of the underparts chestnut with
some narrow grey or whitish edges or scallops replacing the broad white edges of

134 B - p - HALL

jacksoni : there is a conspicuous circle of red skin round the eye : like F. jacksoni and
F. castaneicollis it has a red bill and legs but with the short upper spur of jacksoni.
The width of the greyish edges to the chestnut feathers below varies individually
but, from the limited series available, it seems that in the population of Ruwenzori
(F. n. chapini Grant & Praed, 1934) they are consistently narrow.

The male of the Cameroon Mountain francolin, F. camemnensis, is very like F.
nobilis, except that the maroon of the mantle and wing-coverts and the grey brown
of the back are replaced by a rich dark brown, and the underparts are plain grey
with some darker feather centres: the area of bare skin round the eye is more
extensive. The female is quite different, being mottled and vermiculated above
and below in black, dark brown and buff, with some dark brown or black and off-
white U-patterning, on the mantle, breast and abdomen.

The last species of the group is F. swierstrai which is found associated with patches
of evergreen forest in the Bailundu highlands and along the escarpment in Angola.
It is isolated from all other members of the group by over 1,000 miles and,
morphologically, is the species least readily linked with any other in the group.
Both sexes have conspicuous white eyestripes (not found in other members of the
group) and a dull brown mantle and wing coverts, concolorous with the back, but
which in the female have irregular blotches and transverse markings of darker and
paler brown: the male below is mainly white with the upper breast black con-
trasting with the white throat, and with the lower breast and abdomen streaked
with black on the feather edges : some feathers of the flanks are black with a white
U-pattern. The female below is mainly white with irregular black or brown blotches
or bars, which are most heavily concentrated on the upper breast to form a mottled
band and are sparse in the centre of the abdomen (see Traylor, ig6oa: 142). The
bill and legs are red or orange red as in most other species of the group but the male
has commonly only one spur. It has no red skin round the eye.

EVOLUTION

The distributional pattern presented by these isolated populations associated
with high altitude evergreen forest is only readily interpreted if these populations
are regarded as relicts from a cold, wet era when the montane forest and the group
as a whole had a more or less continuous distribution between the extreme points of
which it is now found (Map 3A) . Any retreat of the forest would cut off the outlying
Cameroon and Angola populations from the eastern birds and they may therefore
be presumed to have been isolated for the longest period, allowing greater opportunity
for divergence.

Similarly the marked, but less extreme, divergence shown by F. nobilis, F. jacksoni
and F. castaneicollis kaffanus is to be expected since the mountains of the Congo
border, Kenya and southern Abyssinia are separated from each other by over
300 miles of comparatively low country. In Abyssinia, on the other hand, the
problem is more complex for the divergence between the northern F. erckelii and
the southern F. castaneicollis is extreme, yet they are not separated from each
other by any greater distance or any greater natural barrier than there is between

THE FRANCOLINS, A STUDY IN SPECIATION 135

the several isolated populations of F. castaneicollis. It seems possible that the
interpretation of this pattern may be found in the ecological differences between the
two species for, from the limited field notes available, it is apparent that F. erckelii
alone of the Montane Group is not dependent on evergreen forest, being found
widely in grassland. (I can find no information on whether it is also found in the
evergreen patches within the range, and it is possible that it exploits both habitats.)
Furthermore, the unusual degree of uniformity of colour and pattern in F. erckelii
argues a continuity of range over a long period, so it can be expected that adaptation
to this type of habitat, which has a less fragmentary distribution than evergreen
forest, took place early in the history of the species soon after it was first isolated by a
retreat of the forest. This adaptation would have raised an ecological barrier
between proto-erckelii and proto-castaneicollis which would inhibit interbreeding
in any subsequent era in which there was a sufficient spread of evergreen forest to
reunite the scattered montane communities in Abyssinia.

It seems unlikely that such a spread has taken place within comparatively recent
times, for there has been time enough for divergence between most of the isolated
populations of F. castaneicollis. This divergence is particularly marked between the
populations west and east of the Rift, F. c. kaffanus and F. c. castaneicollis, and in
the isolated Mega population F. c. atrifrons, and it is likely that breaks were made
earlier between these three groups than between the individual populations of the
south-west.

But while atrifrons can be regarded without much hesitation as a population which
has diverged from F. castaneicollis, any guess at the origin of F. ochropectus in
(French) Somaliland is more doubtful, for it is closest morphologically to F. erckelii
but has some characters of F. castaneicollis and is also a bird of evergreen forest.
It could be postulated that, in isolation, it has descended without much variation
from the common ancestor of erckelii and castaneicollis, or that it is a product of
hybridisation between them at a later reunion. Although the distances that
separate ochropectus from the neighbouring species are not great, the country
between is so low and dry and it would require extreme climatic change to bridge
it with forest or montane grassland, so that the first of these two alternatives seems
the more likely.

There is no such formidable barrier between F. c. erckelii and the greyer F. e.
pentoni, isolated in the Red Sea Hills, and the relatively slight divergence between
them does not suggest lengthy isolation.

A summary of the evolution postulated is as follows :

Stage i. The group widespread in an era in which montane forest covered a large
part of Africa (Map 3A).

Stage 2. With a retreat of montane forest proto-swierstrai and proto-camerunensis
become isolated first and subsequently the ancestral stocks of the other species.
Stage 3. In isolation proto-erckelii becomes adapted to grasslands.
Stage 4. A limited spread of forest allows interbreeding within the ranges of proto-
nobilis, proto-jacksoni and proto-castaneicollis, but though the ranges of proto-
erckelii and proto-castaneicollis may join there is no interbreeding. Proto-
ochropectus remains isolated.

136 B. P. HALL

Stage 5. The forest gradually retreats causing first a split between the populations
of pToto-castaneicollis east and west of the Rift, and isolating the Mega population,
proto-atrifrons , and next the populations of south-west Abyssinia. Subsequently
the forest retreats to its present limits and the country south of the Red Sea Hills
becomes unsuitable for erckelii, isolating the population at Erkowit.

THE SCALY GROUP (MAP 4)
(F. ahantensis, F. squamatus, F. griseostriatus]

RANGE AND CHARACTERS

The range of the Scaly Group lies entirely within 15 of the Equator, and largely
within the limits of the lowland forest in both Upper and Lower Guinea. Inside
the forest these francolins are found in clearings and cultivations. Outside the
perimeter of the forest there are scattered populations north to the Gambia in the
west and to the Jebel Marra in the Sudan, east to Amani in north-eastern Tanganyika,
south to the Vipya plateau, northern Nyasaland, and to the Benguela district in
Angola. These scattered populations may be found in vestigial patches of lowland
forest, riverine forest, montane forest or in cultivations in which there is sufficient
cover in which to hide.

The members of the group are the plainest of the francolins, having little defined
pattern or strong colour. The upper parts are mostly brown sometimes with
underlying vermiculations and with indistinct patterning on the hind neck and
mantle, the feathers having paler edges and darker centres varying from blackish
to orange-brown. There is no defined pattern on the face and the throat is whitish.
The remainder of the underparts are chiefly brown or creamy-buff with the colour
and patterning varying in the different forms but all have very narrow darker edges
to the breast feathers which give a scaly appearance. The bill is orange or red with
some black or brown on the culmen or at the base, and the legs are orange-red, the
males having one main spur and commonly a small upper spur in at least two of
the three species. There is no marked sexual dimorphism but females tend to be
more vermiculated than the males and in some forms tend also to be paler.

SPECIES AND POTENTIAL SPECIES

All forms are allopatric with the three recognised species clearly differentiated,
but otherwise geographical variation in most of the range is not well defined and
tends to be obscured by individual variation. The West African species (F.
ahantensis) is found within the limits of the main block of the Upper Guinea forest
from Sierra Leone to Ghana and in the part of the Lower Guinea forest lying west
of the Niger. Scattered populations are found also in Portuguese Guinea and the
Gambia (and might be expected in former French Guinea).

In ahantensis the upper parts are indistinctly vermiculated with some white
U-patterning on the mantle, the feathers having blackish centres, occasionally with
a reddish-brown shaft streak: the underparts are brown with white and darker
brown U-patterning: the bill is orange with a black base.

The populations of Portuguese Guinea and the Gambia (hopkinsoni Bannerman,

THE FRANCOLINS, A STUDY IN SPECIATION 137

1934) are slightly paler. On the map they appear to be separated from the main
body by about 350 miles, but it is likely that scattered populations exist in suitable
patches of thick bush along the rivers of (French) Guinea, and that variation will
be found to be clinal.

The Niger appears to form a complete barrier between ahantensis on the west
and F. squamatus on the east. The birds found within the boundaries of the
forest from the Niger east to the Ituri (F. s. squamatus) are the least distinctly
patterned of all the group having the upper parts indistinctly vermiculated with
faint greyish U-patterning on the upper mantle, the feathers having blackish centres
tinged with red-brown: the underparts are plain brown with a scaly pattern and
ill-defined darker shaft streaks : the bill is red or orange-red with a darker culmen.

All the scattered populations outside the perimeter of the Guinea forest, namely
those from the Sudan, Abyssinia, eastern Kenya, Uganda, western Tanganyika, the
southern Congo and north-eastern Angola, are sufficiently alike to be united under
the subspecific name schuetti. They differ from F. s. squamatus in being less ver-
miculated, with the pattern on the mantle slightly more clearly defined and with
more red-brown in the centre of the feathers : below they have a less scaly appearance,
the feathers of the abdomen having whitish edges giving an irregular streaky effect.
Within this large range there is much individual and some local variation in the
amount of white below and in the tone of the browns but no significant geographical
variation, although populations are found in varying habitats, in vestigial patches
of lowland and gallery forest and in the lower levels of the montane forests of the
east Congo (up to about 6,500 ft.), and in the mountains of the southern Sudan (up
to 8,500 ft.). In Uganda and Kenya, where much of the forest has been cleared
within the last sixty years, F. squamatus has become largely adapted to cultivations,
but is also found in montane forest up to 9,500 ft. (Mt. Elgon).

West of the Rift in Kenya, southwards to Kilimanjaro, Monduli and Mount Meru
in north-eastern Tanganyika, the birds become increasingly darker and greyer,
with often less white on the abdomen, especially in the males. The darkest, plainest
populations are found on Kilimanjaro (up to 6,500 ft.) (maranensis) and in the
Chyulu Hills.

Although the range of Scaly Francolins in East Africa is now discontinuous and
in these eastern districts the populations are isolated on the mountains (and may be
diverging) the variation exhibited appears fundamentally clinal from a time when
distribution was continuous. However, only 150 miles south-east of Kilimanjaro
birds from the Usambara Mountains (usambarae) show a consistent difference,
having the lores, eye-stripes and cheeks freckled with black and white instead of
being an ill-defined brown. This marking has been described as even more pro-
nounced in a single, rather paler, specimen (the type of uzungwensis) from Kigomo,
Uzungwe Mountains, in south-western Tanganyika, 350 miles to the south. There
are no records of Scaly Francolins from the country between Amani and Kigomo,
much of which is totally unsuitable, but it might be expected in forest patches of
the Nguru and Uluguru Mountains.

Another isolated population is found still further south on the Vipya plateau of
central Nyasaland, in forest between 3,000 and 6,000 ft. (doni Benson, 1939), which

ZOOL. 10, 2 I( >

138 B. P. HALL

is a richer red-brown in the centres of the feathers both above and below, and has
very few white streaks on the abdomen; there is some black in the lores and eye-
stripe. The colour, but not the extent of the red-brown patterning in the mantle,
is matched in a few individuals of schuetti from various parts of its range, but none
approaches this colour below.

The third species of the Scaly Group, F. griseostriatus, is confined to the strip of
rich vegetation with vestigial patches of forest which is associated with the escarp-
ment in western Angola. It is the most distinct member of the group having the
leathers of the mantle and wing-coverts chestnut broadly edged with grey, and the
rest of the upper parts faintly vermiculated, as in F. s. squamatus and F. ahantensis,
but paler: below the feathers of the breast and flanks are chestnut edged greyish
or creamy buff and the abdomen is plain creamy buff. While the coloration is
brighter and more contrasting, the pattern of the feathers of griseostriatus is
fundamentally the same as in the other two species. The bill is mainly blackish
on the upper mandible with a bright red base, and below is orange-red : the legs are
orange-red, and the three males examined have only a single spur.

EVOLUTION

The chief features of the distributional pattern of the Scaly Group are the two
specific divisions which separate ahantensis and griseostriatus from squamatus;
the vast areas in which squamatus exhibits no significant variation : and finally the
indications of speciation in the eastern populations. The fact that the ranges of
F. ahantensis, and F. s. squamatus coincide so closely with the limits of the main
blocks of lowland forest suggest strongly that glades and clearings in this type of
forest must be regarded as the natural habitat for the group, and that the populations
found outside the perimeter of the forest are relicts from a period in which the forest
extended to the extreme limits of the range of the Group. The occurrence of these
populations in montane forest, up to at least 9,500 ft. on some mountains, and the
rapid adaptation of the birds in Kenya to cultivations indicate, however, that their
requirements are not rigid as long as sufficient cover is available. A wide spread of
montane forest might, therefore, not be wholly disastrous for the Scaly Group, though
they could not expect to compete entirely successfully with members of the Montane
Group in these conditions. (It is perhaps significant that the highest altitudes at
which Scaly Francolins have been found are on Mount Elgon, which has apparently
no representative of the Montane Group.)

It seems likely, therefore, that prolonged dry eras have been the chief factors in
speciation. In some such periods (possibly coincidental) early in the history of the
group proto-ahantensis could have been isolated from proto-squamatus by a belt of
savanna stretching to the coast between the Niger and the Volta, and proto-
griseostriastus isolated along the escarpment of Angola, where moister conditions
prevailed, while the interior of the country was dry (Hall, 19606) . With a return
of wetter conditions in West Africa proto-ahantensis spread eastwards. The lack
of divergence between populations on either side of the present gap in the forest east
of the Volta lends support to the conclusions cited by Moreau that the gap between

THE FRANCOLINS, A STUDY IN SPECIATION 139

the two blocks of forest was formerly further east in the neighbourhood of the Niger
delta. The Niger itself at present lies between the two species, and since there can
be little temptation for sedentary birds such as francolins to fly across a broad
river it can be regarded as a barrier sufficiently formidable to inhibit interbreeding.
It is possible that the Lower Congo could have acted as a similar barrier between
proto-squamatus and pitoto-griseostriatus if in a wet era the escarpment forests of
Angola were reunited with the forest of Gabon and Cabinda.

In F. squamatus the remarkable lack of variation in the whole species argues that
the last spread of forest which linked the present outlying populations was of
comparatively recent date. If degree of divergence can be accepted as a guide to
time then it can be postulated that the earliest populations to be isolated by the
retreat of the lowland forest would be those of Nyasaland and Tanganyika: but in
this instance especially this precept must be posed very tentatively for the isolated
populations concerned have also become adapted to rather different conditions at
higher altitudes which might accelerate divergence. Since the most geographically
remote of the remaining populations, those of the Jebel Marra (known from one
specimen only) and southern Abyssinia, do not differ significantly from typical
schuetti in north-eastern Angola and the southern Congo, the forest may have
persisted longest in north-eastern Africa.

Finally the distinction and the fairly sharp transition between squamatus and
schuetti on the perimeter of the Congo forest merits attention. It suggests that at
some time western and eastern birds were divided (presumably in a dry era) and on
reuniting the eastern birds tended to favour rather more open habitat on the
forest fringes, and the western birds clearings in the deep forest.

A summary of the evolution postulated is as follows :

Stage i. The group developing in lowland forest, and probably extending to East
Africa.

Stage 2. In a dry era proto-ahantensis in the Upper Guinea forest becomes cut off
from proto-squamatus in the Lower Guinea forest somewhere in the region of the
Niger. P-roto-griseostriatus in the escarpment forests of Angola also becomes cut
off from proto-squamatus.

Stage 3. In a wetter, more favourable, era, the group spreads again but no inter-
breeding takes place between the incipient species, the barriers formed by the
Niger and Lower Congo possibly reinforcing any tendency to genetic isolation that
has developed while they have been separated.

Stage 4. A dry era divides the Lower Guinea forest into a western block, in which
proto-squamatus subsp. develops and an eastern block in which the birds become
adapted to slightly less humid conditions. These eastern birds, proto-schuetti,
keeping more to the forest fringes when the two are reunited.
Stage 5. A wide spread of lowland forest extends the range of proto-ahantensis to
the Gambia, and proto-squamatus sp. to embrace all the outlying localities at which
it is found at present (Map 4A).

Stage 6. With the retreat of the forest, populations become isolated first in Nyasa-
land, proto-doni, and Tanganyika, proto-uzungwensis and proto-usambarae, then on
the mountains of the Kenya/Tanganyika border, and finally in the Sudan, Uganda

I 4

B. P. HALL

and Abyssinia. These isolated populations survive in isolated patches of lowland
and riverine forest, and also become adapted to montane forest up to high altitudes,
especially where there is no competition from members of the Montane Group, and,
in some areas, to cultivations.

In West Africa populations of ahantensis become isolated in the Gambia and
Portuguese Guinea and the main block of the species is divided by a break in the
forest in the region of the Volta.

THE VERMICULATED GROUP (MAP 5)

(F. Ucalcaratus, F. icterorhynchus, F. clappertoni, F. hildebrandti, F. natalensis, F.
hartlaubi, F. harwoodi, F. adspersus, F. capensis)

RANGE AND CHARACTERS

The Vermiculated Group is the most widespread of all the African groups. I
regard it as consisting of a superspecies (containing six allopatric species) and
three related species. The superspecies has a more or less continuous distribution
from Senegal to Eritrea and southwards to Natal with outlying populations in
Morocco and in the mountains of South West Africa and southern Angola. Of
the three extraneous species one (F. harwoodi} is confined to a small area of Abyssinia
almost encircled by the superspecies: the second (F. adspersus) replaces the super-
species in northern Bechuanaland but is partly sympatric with the outlying
population in South West Africa: the third (F. capensis] is completely isolated in
the extreme south.

The ecology of the group is variable and will be discussed in detail under the
individual species but it can be said in general that from West Africa to Abyssinia
and Uganda the members of the superspecies (F. bicalcaratus, icterorhynchus and
clappertoni) are birds of grasslands and cultivations in the woodland, acacia savanna
and steppe belts. However, in east and southern Africa this habitat is largely
occupied by members of other groups, in particular F. afer of the Bare-throated
Group (Chapin 1932: 714), and it appears that members of the Vermiculated Group
have been driven into more rocky country on hillsides and into thickets in river
valleys and the distribution is consequently more broken (compare Maps 2 and 5).

In choosing the term " vermiculated " as the descriptive name for this group I
have given emphasis to the most striking character which the majority of the
component forms have in common although in some subspecies of F. clappertoni
the vermiculations are largely or wholly replaced by extensive U-patterning. All
members have brown or grey-brown heads, backs, wings and tail, with lighter
vermiculations and/or V- and U-patterning, and sometimes some rufous mottling:
the lores are black or blackish and most forms have a white eye-stripe : the under-
parts are patterned with dark brown (and sometimes chestnut or maroon) on a
whitish or cream background, except in the females of the two sexually dimorphic
forms which are plain orange-ochre below: the colours of the bill and legs, and the
colour and extent of bare facial skin, is variable and so is the number of spurs (see
Table i).

THE FRANCOLINS, A STUDY IN SPECIATION 141

SPECIES AND POTENTIAL SPECIES

The superspecies consists of six clearly differentiated forms which are usually
regarded as species, though it can be argued that two of these, F. hildebrandti and
F. natalensis, could be treated as conspecific. The three northern members,
F. bicalcaratus, F. icterorhynchus and F. clappertoni, form a homogeneous assemblage
occupying grasslands and savannas of various types. F. bicalcaratus is the repre-
sentative in West Africa, being common in the acacia and woodland belts from
Senegal to the Cameroons, extending also into cleared parts of the Upper Guinea
forest. There is also an isolated population in Morocco. Above, it is faintly
vermiculated with extensive V- or U-patterning on the mantle and wing-coverts:
below, each feather is cream with some chestnut on the edges and with a black drop-
shaped centre that is broken by small cream " windows " (Fig. i, 15), giving a
partly streaked and partly spotted effect: the bill and legs are greenish, and two
fairly well-developed spurs are common in the males.

The Morocco population (ayesha) is not conspicuously different, though isolated
by 1,500 miles, but may be slightly more rufous and with rather smaller " windows "
in the patterning below. Other variation in the species is clinal and will be
discussed later.

In the Cameroons, at about longitude I5E., F. bicalcaratus is replaced abruptly
(but not apparently along any ecological boundary) by a plainer bird, F.
icterorhynchus, which is more vermiculated and less V-patterned above, and which is
irregularly V-patterned, blotched or spotted with dark brown below: the bill and
legs differ from bicalcaratus in being mainly orange, and there is a small patch of
yellow skin round the eye. Its range extends eastwards through the woodland belt
(Map n, types 16 and 17) to Uganda, where a few specimens are found which
have some chestnut streaks on the flanks (as in the type of "F. ugandensis) ." In
the acacia belt to the north (Map n, type 20) F. clappertoni replaces F. bicalcaratus
between northern Nigeria and Lake Chad, and replaces F. icterorhynchus in the Sudan.
It is common also in the Nile and Blue Nile valleys and on grassy hillsides of western
Abyssinia and Eritrea below 7,000 ft. (above which it is replaced by the montane
francolin, F. erckelii). Western birds (F. c. clappertoni) are very similar to F.
icterorhynchus above but a more orange brown, while below they resemble F.
bicalcaratus except that there are no " windows " in the dark feather centres and the
feathers are edged with maroon rather than chestnut. In the Nile valley and southern
Sudan birds become darker and greyer, less vermiculated, but heavily U-patterned
above and more heavily patterned below on a whiter ground. In the Blue Nile
valley and eastwards birds lack vermiculations entirely but are clearly and regularly
V-patterned above and on the breast, giving a lanceolate appearance, and they
lack the moustachial streak which is present in the western birds.

A single specimen collected at " Ngeem ", Lake Chad (possibly Nguigmi on the
north-west edge of Lake Chad), the type of "F. tschadensis ", is very like some of the
specimens of F. clappertoni from the Nile valley but the underparts are more buffy
and lack any maroon markings, and it is rather less heavily patterned on the back.
Neumann suggested (c/Bannerman, i : 327) that it is a hybrid between F. clappertoni
and F. icterorhynchus. Since F. icterorhynchus does not now apparently occur as

i 4 2 B. P. HALT.

far north as this, and since this specimen is rather more heavily patterned than the
northern populations of icterorhynchus and clappertoni it might perhaps be better
regarded as an aberrant specimen that indicates the close relationship of the two
species.

East and south of Lake Victoria F. icterorhynchus is replaced by F. hildebrandti,
which is a bird largely associated with rocky ground and thickets in river valleys
and on hillsides, from sea level to about 8,000 ft., and which is found equally in the
acacia savanna and the brachystegia woodland belts. It is common from central
Kenya south to the Tabora and Dodoma districts of Tanganyika but in the southern
part of Tanganyika and northern Mozambique its distribution is sporadic : however,
it is again common in Nyasaland at varying altitudes, and in the upper Luangwa
valley. The males are very like F. icterorhynchus above, but have more solid,
blotchy, brown markings below: the bill is reddish with a brown culmen and
yellowish base, and the legs reddish. The female is quite different having the
underparts plain orange ochre. The sexual differences are most complete in the
most southerly birds from southern Nyasaland, Mozambique and south-west
Tanganyika (F. h. johnstoni] in which the female has no patterned plumage on either
the upper mantle or the breast, whereas those from most of Kenya (F. h. altumi)
have some patterned " male " plumage on both. Females from the centre of the
range (F. h. hildebrandti) are intermediate, having the patterned feathers mostly
confined to the upper mantle.

F. hildebrandti reaches the southern limits of its range in the Luangwa and Shire
valleys, and is replaced further south by F. natalensis, a bird with apparently the
same ecological preferences. Birds from the Transvaal and Natal and the southern
parts of Southern Rhodesia (F. n. natalensis} differ from hildebrandti in having
the sexes alike, the backs darker brown and more heavily patterned and vermiculated
and having regular double U-patterning below, unlike the brown blotches of the
male hildebrandti. The bill and legs of natalensis are similar to those of hildebrandti
except that there is no black on the culmen and the male has commonly one spur and
the female none, whereas in hildebrandti the male has commonly two (once three)
spurs and the female one or two vestigial spurs. However, the intensity and
regularity of the patterning, particularly in the females, decreases northwards,
showing some approach to hildebrandti. The populations of the northern districts
of Southern Rhodesia, the Zambezi valley and the Luangwa valley (neavei) are
plainer and more rufous above and the females tend to have the underparts washed
with buff, the abdomen plainer and the U-patterning less defined giving a more
spotted or streaked effect, but there is much individual variation.

As far as is known the ranges of F. hildebrandti and F. natalensis only approach
each other in two areas, in the Luangwa valley between latitudes 12 and I3S and
in the neighbourhood of the Shire/Zambezi confluence. In the Luangwa valley
there is some indication that the two may sometimes interbreed, or have interbred
in the recent past, for some of the specimens of natalensis are more blotched, less
U-patterned below than those of F. n. neavei, and some of the females are washed
with buff below and have plainer abdomens (see Benson and White, 1957 ' 138) :
one male from Jumbe (13 i6'S., 32 07'E.) lent by the National Museum of

THE FRANCOLINS, STUDY IN SPECIATION 143

Southern Rhodesia is close to F. n. neavei below but has, like hildebrandti, a dusky
ridge to the culmen and two spurs. Nevertheless, there are in this area specimens
typical of the two forms taken from within 50 miles of each other so it is probable
that there is only sporadic interbreeding.

In the Zambezi/Shire area the situation seems rather different. F, natalensis
has been collected from points along the Zambezi as far down as Tambara (60 miles
below Tete) but not as low as the junction with the Shire, and not from the left
bank below the junction of the Mushonganende River, 280 miles above Tete.
Within 40-60 miles of Tambara F. hildebrandti has been collected in southern Nyasa-
land and neighbouring Mozambique at Zobue, " on rocky hillsides " (Vincent 1934:
336), on Mlunganyama hill " at 500 ft. but not at 200 ft." (Benson, C. W. & F. M.,
1948: 3), on Mwananbidzi Hill at 2,500 ft., and near Port Herald in the Shire valley
(Long, 1960: 100), but not as far down the Shire as the confluence. There is no
evidence in this area of interbreeding between F. natalensis neavei and F. hildebrandti
and it has been suggested that here the Zambezi itself, which is as much as two miles
wide when in flood, may be a barrier (Benson, et alia: in press).

The last member of the superspecies, F. hartlaubi, is confined to the mountains of
northern South West Africa and southern Angola and is strictly a bird of the rocks.
It is the smallest member of the group and has distinctive rufous mottling above.
The females are rufous below like those of F. hildebrandti but the male is finely streaked
from chin to abdomen in brown and white. The bill is brownish above and yellow
below and at the base: the legs are yellow with spurs poorly developed but often
with two present vestigially in both sexes.

The first of the three members of the group which are not included in the super-
species is a francolin about which little is known, F. harwoodi. Only males have
been collected and from only three localities, all in the gorges of the Upper Blue
Nile or its tributaries. Of members of the superspecies it resembles F. natalensis
most closely, being similar above but with more defined U-patterning on the nape,
and below having similar double U-patterning, but rather darker and more clearly
defined and extending to the chin, and being absent from the centre of the abdomen,
which is cream. It is distinctive in lacking the white eye-stripe but having a circle
of bare scarlet skin round the eye. The bill is recorded as red but from the appear-
ance of skins probably has a black culmen; the legs are red with two fairly well-
developed spurs. F. harwoodi appears to be cut off from other members of the
group by the presence of the montane francolin F. erckelii on the plateau above the
gorges, in the type of open country which, at lower altitudes, is occupied by F.
clappertoni both to the north and the south.

In southern Africa along the borders of Rhodesia and Bechuanaland the super-
species (represented by F. natalensis) is replaced in similar habitat by F. adspersus,
which ranges westwards through both acacia savanna and steppe (Map u, types 20
and 25) to South West Africa where it overlaps the range of the isolated member
of the superspecies, F. hartlaubi. Here the two forms are found in the same localities
but differ in their ecology and habits, adspersus living in coveys along watercourses
and hartlaubi usually in pairs on the rocks. F. adspersus is a rather different-looking
bird from those members of the group that have been discussed: its appearance is

144 B. P. HALL

grey, the pattern being composed of minute vermiculations on the upper parts
and with broader black-and-white vermiculations over all the underparts. It has
no eyestripe but an area of bare yellow skin round the eye: the bill and legs are
orange-red, the male having usually a single long spur.

The last member of the Vermiculated Group, F. capensis, is isolated in the extreme
south-west. Like many other members of the group it is found in thick cover in
rocky river valleys chiefly in the coastal districts of the Cape. It is not known
from Little Namaqualand (possibly because of the lack of suitable rivers) but has
been recorded from the gorge of the lower Orange River at Assenkjer. It is the
largest member of the group and most like F. natalensis and F. harwoodi in the
predominance of brown and white double U- or double V-patterning. However,
in capensis the patterning is distinctive in being composed of very fine irregular
white lines on most of the back (the rest being vermiculated), and similar on the
underparts but with distinct white shaft streaks, giving a streaky effect: it has no
eye-stripe and no bare skin round the eye : the bill, as in hildebrandti, is red with a
dark culmen, and the legs red, the male having one or two spurs and the female
one.

OTHER VARIATION

In F. bicalcaratus variation is largely clinal, the palest birds being found in the drier savanna
areas and the darkest and most heavily patterned in the cleared parts of the forest areas in
southern Cameroons. Among the dry country birds those from Senegal, Gambia and the
hinterland of Ghana (bicalcaratus) are paler with more rufous heads than those of northern
Nigeria and northern Cameroons (adamauae) . The birds from within the borders of the Upper
Guinea forest from Sierra Leone to Ashanti (thornei) are slightly whiter, less creamy below than
those within the borders of the Lower Guinea forest from south-eastern Nigeria and southern
Cameroons (ogilviegranti) which become increasingly dark eastwards. Birds from Accra, between
the two forest blocks, are closer to bicalcaratus than to thornei: those of most of southern
Nigeria are intermediate between adamauae and ogilviegranti. The isolated ayesha in Morocco
(already discussed) is closest to the paler birds above but has heavier patterning below.

In F. icterorhynchus variation is clinal and ecological, pale, lightly patterned birds being
found in the southern Sudan and in the region of the Albert Nile in northern Uganda (icterorhyn-
chus) : dark, heavily patterned birds are found in the more humid areas from the Oubangi and
Uelle to southern Uganda (dybowskii).

In F. clappertoni the clinal variation is greater than in the two previous species, and features
some changes in pattern that warranted discussion in the main section. West of the Nile
and in the Nile Valley the paler, less patterned clappertoni merges through the intermediate
heuglini to the dark U-patterned gedgii of the southern Sudan. Among the more V-patterned,
lanceolated populations of the east, those of Eritrea, and northern Abyssinia (sharpii) are pale
above (though not as sandy as clappertoni), those of south-western Abyssinia (nigrosquamatus)
are darker and more patterned, and those of the Blue Nile valley (konigseggi) are somewhat
intermediate, closest to sharpii but with a greater suffusion of black markings below.

In the southern populations of F. natalensis there is some variation according to the habitat,
darker, more patterned birds being found in the richer vegetation of Natal, Zululand, and the
Zoutspansberg, and paler birds in the drier savanna country of the Transvaal (see Clancey,
J 953 : 59)- (Since the range of the darker birds is discontinuous it is impracticable to distinguish
the two variations by name.)

The available series of F. hartlaubi is inadequate for detailed study, but it would seem that
the populations of southern Angola (hartlaubi) are rather smaller than those of South West

THE FRANCOLINS, A STUDY IN SPECIATION 145

Africa, and that the western birds from the Kaokoveld and Erongo (crypticus Stresemann,
1939) are rather paler than those from the Waterburg and Otavi further east (bradfieldi) .

EVOLUTION

It has already been suggested that the rather different ecology of the northern
members of the Vermiculated Group compared with those from Kenya southwards
may be due to competition with the Bare-throated Group, and this possibility has
considerable bearing on the interpretation of the pattern presented by the ranges
of the various species. I believe, for example, that while the present can be regarded
as generally favourable for the Vermiculated Group as a whole, it is even more
favourable for the Bare-throated Group and that the distribution of the southern
Vermiculated species is heavily restricted because of this, especially in the areas
where alternative habitats to the savanna grasslands are limited.

This would account for the absence of any member of the group from the grass-
lands south of the Congo forest, also for the sporadic distribution in eastern Africa,
and possibly also for the limited ranges of F. natalensis and F. capensis in the
coastal districts of the south-east, where the gap between them is occupied by the
bare-throated F. afer.

In the north-east it is apparent from the isolated population in Morocco that the
group is not at present as widespread as it has been at some time in the not distant
past (to judge by the lack of divergence in the Morocco population). Elsewhere
in the north the range appears to be expanding, for there is good reason to suppose
that F. clappertoni has only recently extended its range in the acacia savanna south
of the Sahara, where it has supplanted F. coqui (see discussion under Red-tailed
Group and Maps 5 & 8), and also has recently extended its range eastwards towards
the Red Sea coast, cutting off the Erkowit population of the montane francolin
(F. e.pentoni] from the main body of the species (compare maps 3 and 5).

From the foregoing it can be assumed that grasslands and scrub in savanna and
woodland are the natural habitat for the group, but it is adaptable : the presence of
several members in areas where the rainfall in under 10 inches a year shows especially
that it can be tolerant of dry conditions if there is adequate cover in watercourses.
It is likely, therefore, that humid eras in which there has been a wide spread of
forest have been less favourable to the group than the dry eras, and that it is chiefly
in these wet eras that the group has been split and speciation has developed in the
isolated populations. The morphological differences between F. adspersus and
other members of the group, combined with the fact that it has diverged sufficiently
to be able to live alongside another member, F. hartlaubi, in complete ecological
segregation, suggest that probably the group developed in the east and south-east
and that pToto-adspersus was isolated earliest from the original stock. This could
conceivably have taken place at the time when a spread of montane forest linked
the Natal forests with those of Southern Rhodesia and Nyasaland, cutting off
proto-adspersus in the Kalahari/South West African region. Pioto-capensis and
proto-harwoodi can be presumed also to have been isolated at an earlier period than
any of the six members of the superspecies, but several factors could have caused
their segregation. Proto-capensis, as suggested, could have been cut off by the

] 4 6 B. P. HALL

intrusion of the bare-throated F. afer. F. harwoodi on the other hand appears to
be a " relict " population that has survived in the Blue Nile Gorges at a period when
the group was overrun on the plateau above. The fact that the high parts of the
plateau immediately surrounding the range of harwoodi are at present occupied by
F. erckelii of the montane group suggests that it was first split from other members
of the group in an era when montane conditions prevailed all over the plateau, as
they do now.

Proto-hartlaubi in South West Africa may also have been cut off at an early date
from the remainder of the superspecies, in which case it must be regarded as a
coincidence that it has diverged along the same very distinctive lines (with marked
sexual dimorphism) as the Tanganyika population, hildebrandti. It seems more
likely that, following the periods in which adspersus, capensis and harwoodi were
isolated, there was a period which favoured the superspecies, during which it became
widespread through the grasslands of the tropics and south to Natal and the
Transvaal. Meanwhile, proto-adspersus had become adapted to rather drier con-
ditions in acacia steppe to the south-west. A subsequent drier era in southern
Africa which spread acacia steppe through the Limpopo and Zambezi valleys could
isolate proto-natalensis from the rest of the superspecies at a time when proto-
hartlaubi and proto-hildebrandti were still linked. A further spread of steppe
northwards bringing competition with proto-adspersus or a spread of the Bare-
throated Group might then eliminate the superspecies from the areas between the
Rhodesias and South West Africa, where there is little alternative habitat. The
mountains of the west, however, would provide sufficient ecological variety for
proto-hartlaubi to find a different habitat and co-exist in South West Africa with
proto-adspersus, and in southern Angola with the bare-throated afer.

The tendency to approach hildebrandti shown by the Zambezi population of
natalensis (neavei) suggests that these birds are either closest to the ancestral stock
of natalensis I hildebrandti or else are the result of a period in which the two had
become reunited after their distinctive characteristics had developed, but before
there was any bar to free interbreeding. However, the lack at present of an inter-
mediate population showing all stages of intergradation between hildebrandti and
natalensis suggests that the ranges of the two were again separated in the not too dis-
tant past : the presence in the Luangwa valley of a few apparent hybrids among other-
wise typical birds indicates that there has been only sporadic interbreeding now that
the ranges again join. Minor fluctuations in climate would serve to bring about this
splitting and reuniting.

The differences between the various populations of hildebrandti are not great
enough to suggest any lengthy periods of isolation, but, combined with the present
discontinuous distribution, indicate a limited amount of splitting and reuniting
since the species developed. The greater amount of "male" plumage exhibited
by the females of hildebrandti in western Kenya (altumi) indicates some approach to
F. icterorhynchus of Uganda, suggesting that the break between them was at a
comparatively recent date.

The three northern members of the superspecies, icterorhynchus, clappertoni and
bicalcaratus form a homogeneous and completely allopatric assemblage from which

THE FRANCOLINS, A STUDY IN SPECIATION 147

it can be postulated that the present extensive range has not been occupied for a
long period, since otherwise climatic fluctuations would have caused complex
splitting and rejoining of populations over such a wide area. To interpret the main
lines of the pattern it is only necessary to visualise one period in which the ancestral
stock of the three members spread westwards from Abyssinia to Senegal, followed
by one unfavourable period in which it was divided into a western (proto-bicalcara-
tus), central (pToto-icterorhynchus) , and eastern (pioto-clappertoni] population.
This period seems likely to have been a humid one in which Lake Chad was extensive
and forest extended from its southern tip southwards, forming a north-south barrier
between the western and central parts of the savanna belt. At the same time it
can be postulated that there would have been sufficient riverine forest along the
Nile to form a second barrier further east, and that forest in Kenya and Uganda cut
off proto-icterorhynchus from proto-hildebrandti in the south. Speciation developed
in the three pockets and was almost complete when dry conditions returned. Proto-
clappertoni became adapted to rather drier conditions than the other two species
and spread eastwards over the Nile north of proto-icterorhynchus. The differences
in pattern between the eastern and western populations of clappertoni suggest
that at some subsequent period the Nile and its tributaries again formed a barrier
between east and west, allowing birds with the distinctive V-patterning of F. c,
sharpii to develop in the north-east. At present it appears that clappertoni is
extending its range westwards, north of the ranges of icterorhynchus and Ucalcaratus,
at the expense of the red-tailed coqui.

In the west it is apparent that there must have been a period when sufficiently
humid conditions prevailed along the coast to allow the francolins to spread to
Morocco. The lack of marked divergence in the Morocco population (ayesha)
suggests that this period was of a more recent date than the humid period postulated
for the isolation of the ancestral stocks of the three northern species.

A summary of the evolution postulated is as follows:

Stage i. Ancestral stock of the group develops in the east and south-east.

Stage 2. Proto-adspersus isolated in the south, possibly by a spread of montane
forest.

Stage 3. The rest of the group extends northwards to Abyssinia, and widely through
grasslands south of the Congo. Pioto-capensis isolated, perhaps by competition
with the Bare-throated Group.

Stage 4. The group spreads westwards to Senegal but withdraws from the high
plateau of Abyssinia, which is occupied by the Montane Group, leaving a relict
population in the Blue Nile gorges, proto-harwoodi.

Stage 5. (South). Pioto-natalensis isolated, possibly by extension of acacia steppe
in Limpopo or Zambezi valleys bringing competition with proto-adspersus.

Stage 5. (North). Proto-bicalcaratus in the west, proto-icterorhynchus in the centre
and pioto-clappertoni in the east, isolated from each other, probably by a belt of
forest extending northwards to Lake Chad (at a time when its area was more exten-
sive), and by riverine forest in the Nile valley. Pioto-clappertoni becoming adapted

148 B. P. HALL

to drier conditions than the other two species. Forest in Uganda and Kenya
separates proto-hildebrandti from proto-icterorhynchus.

Stage 6. (South). Proto-hartlaubi isolated in the mountains of South West Africa
and Angola, possibly by an invasion of the central areas by proto-adspersus or by
the Bare-throated Group.

Stage 6. (North). Minor fluctuations, allowing pToto-clappertoni first to spread
westwards across the Nile and then splitting into a western and eastern population,
possibly during a return of more humid conditions when the Nile valley again
became a barrier.

Stage 7. (South). Minor fluctuations causing rejoining and re-splitting of proto-
natalensis and proto-hildebrandti, and isolating populations of hildebrandti for short
periods.

Stage 7. (North). A humid period in which proto-bicalcaratus spreads to Morocco,
followed by a drier period in which the Morocco population is isolated.

Stage 8. (South). The ranges of natalensis and hildebrandti rejoin, at least in the
Luangwa valley, and there is sporadic interbreeding.

Stage 8. (North). In the Sudan clappertoni spreads westwards through the drier
acacia belt.

THE STRIATED GROUP (MAP 6)
(F. sephaena, F. streptophoms)

RANGE AND CHARACTERS

The Striated Group is unlike all others in consisting of only two species which are
partly sympatric. The range of the group as a whole is mainly in eastern Africa,
from Somalia to Natal, but extending across to South West Africa and southern
Angola, and with an isolated population in the Cameroons.

Both species are mainly brown above with some chestnut, black and white pattern
in the collar, and white shaft streaks on the mantle: the eye-stripes are white,
edged with black and there is some chestnut on the sides of the face: the throat is
white bordered with chestnut: the underparts are creamy buff with contrasting
maroon, chestnut, or black-and-white patterning on the breast (different in the two
species) and lighter patterning on the flanks and abdomen.

The habitat of the two species is different and will be discussed later.

SPECIES AND POTENTIAL SPECIES

The two species are quite distinct both ecologically and morphologically. F.
streptophoms has a puzzling distribution, being found commonly in the grasslands
and " orchard bush " of northern Uganda and sporadically in north-western Kenya
usually on stony scrub-covered hillsides (Jackson, 1938: 241). In north-western
Tanganyika Dillingham has found it also in clearings of brachystegia. These are
the only areas from which it is known in the east but 1,500 miles to the west a single
specimen was obtained by Bates among grass and rocks between 3,500 and 4,000 ft.,

THE FRANCOLINS, A STUDY IN SPECIATION 149

in the Cameroon highlands 40 miles west of Fumban (on the borders of former
British and French Cameroons).

It is a darker bird than F. sephaena the back being grey-brown with dark brown
patches and with fewer and narrower white shaft-streaks. It has a solid chestnut
collar on the hind neck above a broken black-and-white collar of barred feathers :
there is a considerable amount of chestnut in the facial pattern but only a narrow
band bordering the base of the throat: the feathers of the breast are barred black
and white occasionally with a chestnut tip: the remainder of the underparts is
pale cream marked with some dusky vermiculations and broad blackish streaks on
the flanks : the bill is black with yellow at the base and the legs yellowish with only
a vestigial spur in the males. The sexes are alike. The Cameroon bird is like those
of East Africa but smaller (wing i<$ 145 against 6^ 151-160).

In northern Uganda, where both species are found, Pitman tells me that sephaena
is strictly confined to the driest areas of scrubby thorn country while streptophorus
is in savanna grasslands. Both Jackson (1938: 241) and Dillingham (in litt.} stress
that the call of streptophorus is quite unlike that of any other francolin known to
them, and is a melodious trill or whistle. Dillingham on the other hand found
sephaena' s call closer to other francolins though Vincent, on a label of a bird from
Swaziland, describes it as "a loud piping trill " and Benson tells me he has
found it very variable, Rhodesian birds being hardly recognisable as the same
species as those in Abyssinia. Comparison between the eggs would also be useful
for Pitman confirms Jackson's statement that the eggs of sephaena are exceptionally
hard, and it would be interesting to know if those of streptophorus, at present un-
known, have the same character.

F. sephaena is widespread in the acacia savanna and steppe from eastern Abyssinia
and (British) Somaliland westwards to the Albert Nile, southwards through Kenya
and along the Tankanyika coast to Mozambique and Natal, westwards to South
West Africa and southern Angola. It is never far from water and in the driest
parts of the range is confined to watercourses.

Above it is a rich red-brown with broad white shaft-streaks : the chestnut collar
is broken with white and the patterned feathers of the upper mantle are rich dark
brown or blackish with elliptical white centres : the females are slightly vermiculated.
There is no black on the underparts, the feathers of the breast being cream with
triangular chestnut or maroon markings on the tips : the bill is wholly black and the
legs carmine, the males having a single long sharp spur.

The birds here designated as F. sephaena have sometimes been considered to
belong to two species, for two distinct types are found, one, mainly coastal, which
has fine, drop-shaped maroon streaks on the abdomen, and the other in which the
abdomen may have some dusky patterning but which has no maroon streaks. In
the Shire valley of southern Nyasaland birds of the two types appear to behave as
species being found close to each other without an ecological division, but without
any intergrading : specimens from the southern tip of Lake Nyasa and eastwards
are streaked, and so are those from the lower Shire valley at Chiromo and Tangadzi,
but a specimen from Lengwe, 35 miles upstream, is unstreaked, as are those from
all localities westwards in the Zambezi valley (Benson, C. W. and F. M., 1948: 3).

1.50 B. P. HALL

In northern Tanganyika a similar clear division is found between the two types,
birds from inland, at altitudes above 1,500 ft., being unstreaked, those from the
coastal plain being streaked. However, from the Kenya boundary northwards
there is no clear division either geographically or morphologically: in western Kenya,
southern Abyssinia (Benson, 1945: 391) and throughout Somalia both types are
found, as well as some intermediates with sparse, fine streaks. In (British) Somali-
land streaked birds predominate at lower altitudes near the coast but both types
are found in the highlands and streaked specimens have been found as far inland
as Harar in north-eastern Abyssinia. There is no evidence that the two types
behave here as species and it is more probable that they represent a hybrid population.
(For this population it is convenient to use the name spilogaster, which was given
to a streaked bird from Harar see Rand, 1950 : 384 while rovuma should be confined
to the unmixed streaked populations from Tanganyika southwards.)

It will be seen from the map that inland birds of the unstreaked type are found
in two blocks from northern Tanganyika northwards and from southern Nyasaland
southwards, the species being absent from inland areas through most of Tanganyika,
Nyasaland and Northern Rhodesia. Birds of the northern block (grantii) have the
patterning on the collar and on the breast more restricted (especially in the females)
than the southern birds. In birds of the southern block the patterning on the
collar is brighter and more extensive and the patterning below extends to the lower
breast: birds from the Transvaal, Natal and extreme southern Mozambique
(Coguno) have the rest of the underparts fairly heavily vermiculated (sephaend),
but northwards through Southern Rhodesia the vermiculations become lighter and
birds from South West Africa and southern Angola across the continent to southern
Nyasaland are appreciably less patterned (zambesiae) .

OTHER VARIATION

Throughout the range of F. sephaena there is considerable local variation in size with
equatorial birds on the whole being smaller. Exceptionally small birds are found on the Kenya
coast and exceptionally large ones in the Upper Zambezi and Chobe valleys. Among birds
without the chestnut streaks on the abdomen there is considerable variation in the amount
of vestigial patterning on the abdomen, some having ill-defined whitish streaks and irregular
dusky U-patterning. On the whole eastern birds in, and bordering, the hybrid zone tend to
be more patterned than western.

In South West Africa some rather greyer birds are found, probably associated with the
limestone pans, but the variation is not constant.

EVOLUTION

The peculiar distribution of F. streptophorus is the major problem set by the
Striated Group. The fact that the eastern and western populations of streptophorus
are not strongly differentiated and yet the country in between appears totally
unsuitable for the species indicates that these are relicts from an era when the group
had a wide distribution north of the Congo forest. Being relicts they may not now
be in their natural habitat and it can, therefore, be reasonably postulated from the
distribution of sephaena, that the natural habitat of the Group is in acacia steppe,
and that proto-streptophorus once occupied this belt from the Sudan westwards,

THE FRANCOLINS, A STUDY IN SPECIATION 151

north of the forest and woodland belts. One thinks usually of these belts as moving
north and south with successive climatic fluctuations, but it is possible to visualise
that the ridge of higher ground that stretches at about latitude 6N from the
Cameroons eastwards to Uganda may sometimes dislocate regular north-south
movement. It seems possible that, following a dry era in which the belt of acacia
steppe lay as far south as this ridge, the return of more humid conditions and richer
vegetation might come first to the lower country of the Shari basin and Bahr-el-
Ghazal to the north of the ridge, leaving temporarily a pocket of dry country on
the ridge while the main belts moved north. If proto-streptophorus became isolated
in this pocket it could be expected that, as conditions became more humid and less
and less suitable, it would be driven to the only parts where a tolerably dry habitat
could still be found, and one which would be unattractive to the indigenous francolins
of the encroaching woodlands. Such refuges were provided in the higher ground at
the western and eastern extremes of the previous range.

It is possible that streptophoms may yet be found in the country between the
Cameroons and Uganda, but if so could be expected only in some patches of sparse
vegetation. Since so little divergence is shown by the Cameroon bird it can be
postulated that it has been separated from the Uganda population at a comparatively
recent date.

The pattern presented by sephaena points to at least one unfavourable era in
which the species was divided for long enough to allow the streaked and unstreaked
types to develop in isolation, almost to the specific level. The present distribution
shows how small an extension of woodland towards the coast in Tanganyika would
divide the species. In a considerably wetter era with a wide spread of woodland
or forest it is likely that the northern populations would be confined to the coastal
plain of Somalia while the southern populations might develop on the fringes of the
Kalahari. If this was followed by a dry era with a spread of acacia the unstreaked
southern populations might spread northwards up the Luangwa valley, through
Tanganyika and Kenya along the valleys of the rivers which run westwards to the
lakes, and then up the Abyssinian Rift to the Somali highlands.

Meanwhile the streaked northern populations spread southwards, chiefly in the
valleys of the rivers flowing to the sea. It is difficult to account for the fact that the
two types seem to behave as species in the south but mix and interbreed to a limited
extent in the north. It may be that the two came together at a later date in the
south when speciation was further advanced, and it is also possible that in northern
Tanganyika the sharper change of altitude on the edge of the coastal plain may
provide here a barrier that is lacking in Kenya and eastern Somalia (though not in
the mountains of (British) Somaliland).

The present gap in distribution between the unstreaked birds of Kenya and
northern Tanganyika (grantii) and those of the Rhodesias and southern Nyasaland
(zambesiae) is sufficient to account for the divergence between them.

The scattered distribution of sephaena along watercourses in the drier areas would
account for much of the local variation in both size and patterning that is exhibited,
for such populations would be isolated and might be expected to diverge in periods
when there is little suitable vegetation between one river valley and another.

152 B. P. HALL

A summary of the evolution postulated is as follows:

Stage i. Ancestral stock occupying the acacia belt north, east and south of the
central forest and woodland block.

Stage 2. The stock split into two, one, proto-streptophorus , in the acacia belt north
of the central forest, the other, proto-sephaena, in acacia in the eastern districts from
Somalia to southern Africa. This split probably taking place in a humid era with
a spread of woodland or forest between Uganda and Abyssinia.

Stage 3. F. sephaena split into a northern group, probably confined to lowland in
Somalia, and a southern group, probably on the borders of the Kalahari. The
northern group being the ancestral stock of the birds with streaked abdomens, and
the southern group of those with unstreaked abdomens. This split taking place
in an even wetter era than Stage 2 with a wide spread of forest and woodland over
most of central Africa. Proto-streptophorus still isolated south of the Sahara.

Stage 4. (East). A return of arid conditions with increased acacia allows the
northern streaked birds to spread down the coast, and the southern, unstreaked
birds to spread north up the Luangwa valley, through western Tanganyika and
Kenya and the Abyssinian Rift. The two types meet first in the north, and in
Somalia they mix and interbreed to a limited extent forming a hybrid zone but
with streaked birds always predominant near the coast. Subsequently they meet
in northern Tanganyika and later still in southern Nyasaland, but do not interbreed.

Stage 4. (North). In a dry era the acacia belt and proto-streptophorus move south
to the ridge of high ground from the Cameroons to Uganda and north-western
Kenya (Map 6A).

Stage 5. (East). An increase in humidity with a spread of woodland in central
Africa cuts off the northern unstreaked birds, proto-grantn subsp., from the
southern unstreaked birds, proto-zambesiaej sephaena subsp.

Stage 5. (North). An increase in humidity, coming first to the low ground of the
Shari basin and Bahr-el-Ghazal, north of the ridge, isolates proto-streptophorus
along the ridge. Increased humidity and competition with indigenous woodland
species drive it into refuges at the extremes of its range where more arid conditions
persist.

THE RED-WINGED GROUP (MAP 7)
(F. psilolaemus, F. shelleyi, F. africanus, 1 F. levaillantoides , F. levaillantii, F.finschi]

RANGE AND CHARACTERS

This is a very homogeneous group which ranges over most of eastern and southern
Africa from Eritrea to the Cape and westwards to Angola. The various species
have a complex distribution but are largely allopatric though the ranges of three
overlap in the Transvaal. Members of the group are found in a wide variety of
habitat and at any altitude, and even individual species are found in varying habitats
in different parts of their range.

1 F. afer of Praed & Grant see appendix.

THE FRANCOLINS, A STUDY IN SPECIATION 153

All birds of the group have " quail-type " plumage on the backs, with a lattice-
work pattern of irregular white or buff bars and shaft streaks on a basic colour of
mixed black, brown, grey and chestnut: all have the under wing-coverts and part
of the wing chestnut (including the South African form F. africanus, commonly
known as the Greywing, in which the chestnut is reduced) : all have a marked facial
pattern but this, and the pattern on the underparts, varies in the different forms:
all have black bills, usually with a yellowish base, and yellowish legs, the males
usually with a single spur. The sexes are alike.

SPECIES AND POTENTIAL SPECIES

The distribution and relationships of the various forms of this group would be
easier to discuss if no previous attempt had been made to designate species or
subspecies in the East African forms. I cannot agree with any of the classifications
made for this group which seem to place too great a reliance on minor morphological
characters, particularly the size of the bill, without regard to the ecology of the
forms of their likely evolution. The rearrangement I propose I believe to be more
realistic in these respects, though I appreciate that conventional nomenclature fails
to give more than an inadequate picture, and it is necessary as well to discuss the
relationships in some detail. To facilitate the recognition of the species as re-
arranged they will be discussed under separate headings and I have proposed new
common names for each.

F. psilolaemus The Montane Redwing.

F. psilolaemus comprises the montane populations of Kenya, Uganda and
Abyssinia known by Praed & Grant as F. shelleyi theresae, F. s. elgonensis, F. afer
psilolaemus and F. a. ellenbecki.

All these birds are found in heath and grasslands above 8,000 ft. All have a high
proportion of rich chestnut in the wings and differ from other red-winged francolins
in having some barring on the tips of the primaries: the underparts are rich buff
mottled with chestnut and with some dark brown or black markings: the facial
pattern is broken and ill denned. Birds from the Shoa mountains round Addis
Ababa (psilolaemus) are small with small bills (wing ^164-174, bill ^30-34), and
have blackish spots on the throat, indistinct black spotting on the upper breast
and rather sparse chestnut and blackish patterning on the abdomen. Birds from
the Arussi plateau south of the Abyssinian Rift (ellenbecki) are slightly larger than
Psilolaemus, darker on the head and back, the lighter markings being greyer: below
the spots on the throat and breast are darker and the patterning is stronger and
richer in colour. Birds from Mount Kenya and the Aberdares (theresae) are large
(wing (174-191, bill ^35-38) : they have only occasional speckles on the throat
but are heavily spotted on the upper breast : they are darker above than Abyssinian
birds and the chestnut markings on the abdomen are richer. Birds from Mount
Elgon (elgonensis) are similar to theresae but darker and richer.

F. shelleyi Shelley's Redwing.

F. shelleyi comprises the forms known by previous authors as F. s. shelleyi,

ZOOL. 10, 2 II

154 B - P. HALL

F. s. whytei, F. afer uluensis, and F. a. macarthuri Van Someren, 1938, but excludes
theresae and elgonensis which were previously ascribed to shelleyi (see above).

All the birds included in this assemblage differ from F. psilolaemus in having the
breast maroon, chestnut and grey, unmarked with black below the necklace and
gorget : the remainder of the underparts are patterned all over with black and white,
the pattern basically of white spots on a black ground, but the spots frequently
merging to form irregular bars. In Kenya and northern Tanganyika birds of this
type with rather small bills, white throats, well-defined facial pattern and necklace
and rather small patterning on the abdomen are found on hillsides, grassy plateaux,
and clearings in montane forest from Mount Kenya southwards to the Crater
Highlands and Mount Meru. On Mount Kenya and the Aberdares they are found
at lower levels than F. psilolaemus keeping below the montane forest. Since they
are confined to uplands above 3,000 ft., their distribution is necessarily discontinuous
and some divergence is shown by isolated populations, birds of the Loita Plains in
south-western Kenya being rather greyer and those of the Chyulu Hills (macarthuri)
rather darker than birds from central Kenya (uluensis).

The most southern representative recorded of this small-billed form is from
Mbulu at 4,900 feet in the highlands south of Lake Manyara in northern Tanganyika.
Fifty miles south-east at Salanga on the Bereku Ridge at about 5,500 feet a large-
billed bird has been collected which has larger, more open, patterning on the
abdomen: between the two localities there is lower, rather open country probably
unsuitable to these francolins. Birds of the large-billed type are found on the
Uganda/Tanganyika borders and at scattered localities through western and
southern Tanganyika, Mozambique, southern Nyasaland, southern Northern
Rhodesia, Southern Rhodesia the Transvaal and Natal. The small-billed, small-
patterned form of Kenya (uluensis) and the large-billed, large patterned form of
Tanganyika (shelleyi) have been regarded as belonging to different species and the
proximity of the two in the Mbulu area lends weight to this view. Nevertheless,
they seem to have similar ecological requirements and field habits; and, while in
the Mbulu area the topography of the country may inhibit interbreeding, there is a
bird from Amani, 200 miles to the west, which is intermediate between the two
forms in the patterning on the abdomen, though closer to shelleyi in bill size, which
suggests there may be no intrinsic barrier.

The long-billed shelleyi is found in varying habitats throughout its range. In
southern Nyasaland, as in northern Tanganyika, it may be found on montane grass-
land up to 7,000 feet, but it is also found in Nyasaland and Northern Rhodesia in
grass under brachystegia or even mopane at low altitudes, though it is nowhere
common. In Southern Rhodesia it is common in Mashonaland and the Midlands,
in open grass of the plateau, but scarcer westwards in Matabeleland (Smithers
et al, 1957: 46). It occurs at least as far south as Pretoria and Legogot, in the
Barberton district of the Transvaal, in presumably the same type of country as in
Southern Rhodesia, but in Natal is found in acacia country below 2,000 ft. along
the foothills of the Drakensburg.

In northern Nyasaland, the south-eastern Congo and northern Northern Rhodesia
the species (as proposed) is represented by another long-billed form (like shelleyi), but

THE FRANCOLINS, A STUDY IN SPECIATION 155

with a small patterning (like uluensis) and which differs from both in having a buff
throat with the facial pattern and necklace freckled and indistinct. This form
(whytei], like shelleyi, is found at varying altitudes and in both brachystegia and
montane grassland and, though there is no ecological barrier between them and
they have been collected from within 40 miles of each other at Lundazi and Mzimba,
no intermediates are known (Benson, 1951: 79). Benson tells me that he believes
this is more likely to be due to lack of collecting and to the scarcity of both forms,
rather than to the existence of any barrier to interbreeding. The situation may be
found to be comparable to that between uluensis and shelleyi in northern Tanganyika
but further research is required.

F. africanus (F. afer) the Greywing.

I regard the South African Greywing as a monotypic species, and do not associate
any of the northern short-billed forms with it. It differs from F. shelleyi in having
a white throat flecked with black, the facial pattern obscured by black-and-white
freckling, the patterning on the underparts very small, and the rufous in the wing
very much reduced. The bill is small as in uluensis. It is found in the grasslands
of the mountains of the southern Transvaal, Orange Free State and Cape Province,
reaching Natal on the high spurs of the Drakensburg. McLachland & Liversidge
(1957 : 94) show an overlap in the ranges of F. shelleyi and F. africanus in the
southern Transvaal but I cannot substantiate this, the most northerly specimens o'f
africanus recorded being from Potchefstrom and Wakkerstrom, 100 miles south of
shelleyi at Pretoria, and in rather different country.

In Natal Vincent tells me there is a clear altitudinal distinction between them,
africanus being found above 6,000 ft. and shelleyi below 2,000 ft., while the inter-
mediate levels are occupied by a third member of the group, F. levaillantii. Further-
more F. shelleyi has an earlier breeding season than the other two species, which
would contribute to segregation.

It will be appreciated therefore that africanus has, in many respects, the same
relationship to shelleyi in the Transvaal as shelleyi has to whytei in Northern Rhodesia
and to uluensis in Tanganyika, in that it appears to be completely allopatric and to
exhibit distinct morphological differences. But, whereas there appear to be no
ecological barriers between shelleyi /uluensis and shelleyi /whytei and the morpho-
logical differences are slight, there is a distinct ecological barrier between shelleyi
and africanus and morphological differences are greater, representing a greater
degree of divergence. This to me is just sufficient to warrant giving africanus
specific rank.

F. levaillantoidss (olim F. gariepensis) Acacia Redwing.

F. levaillantoides comprises all the subspecies of southern Africa usually ascribed
to this species and in addition the northern populations known by Praed and Grant
as F. afer stantoni, F. a. friedmanni, F. a. archeri, F. a. lorti and F. a. gutturalis.

It will be seen from the foregoing that I regard the red-winged francolins of the
northern and southern acacia steppe as conspecific. They differ from F. psilolaemus
and from F. shelleyi in being paler and less patterned below, lacking on the abdomen

156 B. P. HALL

the rich chestnut markings of psilolaemus or the black-and-white patterning of
F. shelleyi, having instead sparse chestnut and blackish flecks or streaks, more
common in the northern birds.

In southern Africa they are found in the arid zones from coastal Angola through
South West Africa, Bechuanaland, Orange Free State and the Transvaal to the
east coast in extreme southern Mozambique, associated, at varying altitudes, with
open grassland, hillsides and kopjes, and edges to limestone pans. There are
sight records (not included on Map 7) from the Lower Orange River (Winterbottom
& Courtenay-Latimer, 1961: 11-12 and in litt.}. The range of levaillantoides
overlaps that of africanus in the Orange Free State and that of shelleyi in the
Transvaal. There is little information on their relationship here (perhaps because
their numbers have been heavily reduced since the country was settled) but it can
be expected that levaillantoides would be found at lower altitudes than africanus,
and in drier areas than shelleyi.

In the northern acacia belt the red-winged francolins are found in much the
same habitat as southern birds, in sparse grass on rocky hillsides and mountains
between about 2,000 and 7,000 feet. In central Abyssinia, where the Acacia
Redwing is found in the same areas as the Montane Redwing, it occupies the lower
slopes and F. psilolaemus the higher (Friedmann, 1930: 112-116).

In northern Abyssinia and Eritrea there is an isolated population (gutturalis)
found in scrub above 6,000 feet (Smith, 1957: 20). These birds appear to be a
link between the acacia species and the Montane Redwing F. psilolaemus, for, like
the montane bird, they have the facial pattern freckled and ill denned. They also
differ from other Acacia Redwings in having the abdomen regularly marked with
heavy blackish streaks on the feather centres, though the actual shafts are white,
while the flanks are broadly streaked with chestnut. However, in colour and in
the distribution of patterning below they seem closer to the Acacia Redwings and
are tentatively regarded as conspecific with them.

Otherwise variation in pattern among the Acacia Redwings is slight. All have
the facial pattern and necklace well defined, but in birds from southern Angola
(jugularis) south at least to the Cunene, the black-and-white necklace is enlarged
on the breast to a broad gorget. The birds from the Benguela area are exceptionally
small, and also rather grey, while Cunene birds are larger and rufous. Throughout
both the northern block and other members of the southern blocks there is similar,
very marked variation in colour between grey and rufous populations, largely
associated with ecology, but there is remarkably little consistent difference between
northern and southern birds in spite of the great distance separating them; those
from the north have commonly, however, fine dark markings on the abdomen that
are usually lacking in those from the south and the abdomens are whiter, less buff.
The greyest populations are found in the mountains of eastern (British) Somaliland
(lorti), in the Kalahari (kalaharica see appendix), and in the Benguela district
(as already noted) . The most rufous birds are found in the Transvaal and Orange
Free State (levaillantoides}, while in both South West Africa (pallidior} and southern
Abyssinia and southern Sudan (archeri) both rufous and greyish micro-populations
are found haphazard (see appendix).

THE FRANCOLINS, A STUDY IN SPECIATION 157

It will be seen that, as far as the northern populations of the red-winged francolins
are concerned, the subspecies of the acacia zone could be considered conspecific
with F. shelleyi, since they are aUopatric. However, since in southern Africa
F. levaillantoides and F. shelleyi both occur in the Transvaal, two species must be
recognised, and as the northern acacia populations are closer both morphologically
and ecologically to the southern acacia birds than to F. shelleyi, it seems logical to
associate them in the same species.

F. levaillantii Levaillant's Redwing.

I have made no rearrangement in this species, which is universally recognised
as comprising scattered populations from north-western Kenya to the Cape and west-
ern Angola and which differs from all other red- winged francolins in having an ochre
collar and the sides of face and edges of the throat ochre inside the black-and-white
facial pattern. Such populations are found in north-western Kenya, where west
of the Rift F. levaillantii replaces F. shelleyi in grasslands and valleys of the plateaux
between 6,000 and 8,000 feet. In Uganda and Ruanda it is found at varying
altitudes being common in the lower grasslands west of Lake Victoria (south of
Bukobo) and on the plateaux westwards to Lakes Edward and Kivu.

It is absent from most of Tanganyika but reappears on the Njombe and Iringa
mountains in the south-west, and on the Nyika Plateau of northern Nyasaland,
where it is common above 7,500 feet alongside the rarer F. shelleyi whytei (Benson,
in litt.}. Another isolated population is found on the Kibera Plateau in Upemba,
southern Congo, at about 6,000 feet (Verheyen, 1953 : 256), and scattered populations
are found in the grasslands of the Balovale and Mongu districts of north-western
Northern Rhodesia westwards through Angola to the Mombolo plateau and the
Bailundu highlands. In southern Africa it is found only from the Zoutspansberg
southwards through the eastern Transvaal, Natal (between 2,000 and 6,000 feet),
Basutoland and the coastal districts of eastern Cape Province, keeping to the
moister grasslands and even reed-beds in river valleys. It seems, therefore, in the
Transvaal, Natal and Orange Free State, where the four species of Redwing occur,
that F. levaillantii is associated with the wettest habitat.

The distinctive ochre patterning of the face and throat of F. levaillantii has
already been noted : it is otherwise most like F. levaillantoides but is long-billed
and darker and richer in colour, with the black-and-white patterning of the necklace
and gorget extending in a complete collar round the hind neck below the ochre
collar, and the black-and-white stripes from above the eye, which in levaillantoides
and shelleyi run down the side of the face, in levaillantii run behind the head to join
at the back. In spite of the wide range and broken distribution there is little
variation between the populations but in the populations of southern Africa and of
the Nyika plateau these stripes continue down the back of the neck to join the
black-and-white collar, while in the western and northern populations the ochre
collar is unbroken. The Nyika birds (crawshayi) are richer in colour than those of
southern Africa (levaillantii), with more rufous in the wing, and with more black
markings on the abdomen, while the northern birds without the stripe on the hind
neck (kikuyuensis) are otherwise similar to crawshayi.

I 5 8 B. P. HALL

F. finschi Finsch's Redwing.

The last of the red-winged species, F. finschi, is monotypic. It has a very limited
distribution being recorded only from about six localities in western Angola, in a
strip about 200 miles long from near Nova Lisboa in the south to Vila Salazar in
the north, and again from Brazzaville, in the Congo, 350 miles to the north. On
Mount Moco in the Bailundu highlands it was found in the brachystegia woodlands
and on the bare upper slopes between 6,500 and 7,000 feet (Hall, i96oa: 408) but
the specimen from Vila SaJazar (American Museum of Natural History) must have
been obtained in very different habitat at not more than 2,000 feet, and at Brazza-
ville Malbrandt and Maclatchy (1949: 142) found it in grassland with neighbouring
woods in which the birds took refuge.

It is possible that the different habitats of F. finschi may be partly due to com-
petition with F. levaillantii, which occupies grasslands in the southern part of the
range, but is not found north of the Mombolo plateau.

F. finschi differs from other red-winged f rancolins in lacking any black-and-white
patterning on the face and neck and having the breast grey: it is closest to F.
levaillantii in having a long bill and the sides of the face and border of the throat
ochre, but in F. finschi the ochre is not so extensive on the hind neck : the abdomen
is a mixture of buff and pale chestnut.

OTHER VARIATION

Considerable individual variation is to be expected in any birds as intricately patterned as
the red-winged francolins and is found in all species. There is some local variation, usually
in an increase or reduction of rufous and grey in the plumage, associated with ecology. This,
as noted, is particularly marked in F. levaillantoides, but since the problems it raises are largely
nomenclatorial it is discussed fully in Appendix 2.

EVOLUTION

The complex pattern displayed by the seven species of the Red-winged Group is
not easy to interpret but I believe may be partly explained by comparison with a
rainfall map (Map 7A) . From this it will be seen that the range of F. levaillantoides
falls largely in the parts where the rainfall is below 20 inches in the year, the range
of F. shelleyi falls between the 20-40 inch isohyets, the range of F. levaillantii falls
between the 40-60 inch isohyets, while the range of F. finschi coincides remarkably
with the only two patches south of the Congo rainbelt where there is over 60 inches
in the year. Thus a slight increase or decrease of rainfall may be expected to favour
the spread of one species at the expense of others. The discontinuous distribution
of all species, and the scarcity of red-winged francolins in so many parts of their
range, suggests that the present cannot be regarded as a favourable era for any
members. This might be due to competition with other groups, but if so it could be
expected that F. levaillantoides would be more common in the Kalahari, where
there is no competition. From the association of so many members of the group
with mountains I believe that a colder era would be more favourable to the group
as a whole, and it was in some such era or eras, with fluctuations in rainfall, that

THE FRANCOLINS, A STUDY IN SPECIATION 159

continuity in the ranges of the various species was achieved. Conversely warmer
eras could be regarded as least favourable and ones most likely to produce isolated
populations in which speciation developed.

Even with this hypothesis as a basis a good many interpretations of the pattern
present themselves which are largely dependent on which species one believes are
most closely related. The following sequences are put forward tentatively on the
premise that plumage pattern is the best guide to relationship. I can offer no
explanation of the variation in bill size in terms of evolution and relationship but it
is worth noting that in this group it is the species and, in shelleyi, the subspecies,
associated with the higher rainfall areas that have the long bills. The length may
perhaps be correlated with the hardness of the ground in which the birds search for
food.

The distinctive patterning on the abdomens of F. africanus and F. shelleyi suggests
that their common ancestor split at an early date from the rest of the group. The
present distribution suggests they developed in the south ; later spreading north to
Kenya in a favourable era; later splitting into two blocks, proto-africanus and
proto-shelleyi ; later proto-shelleyi splitting into three from which the subspecies
shelleyi, whytei and uluensis developed. A favourable era must have followed in
which their distribution was continuous over the present range.

The species with relatively unpatterned abdomens may have originated in the
north. A warm era could be expected to drive members of this block to the
extreme north of the range, and also to high altitudes: it can be postulated that
proto-psilolaemus developed in such an era to specific status and became adapted to
montane conditions : later in a favourable era other members of the northern block
spread south and west as far as Angola and the lower Congo. A subsequent un-
favourable era isolated first proto-finschi in the Angola mountains, then proto-
levaillantii in the mountains of central Africa, and proto-levaillantoides in Abyssinia,
Somaliland and Eritrea: in isolation each became adapted to different degrees of
humidity: subsequent fluctuations in rainfall allow first proto-levaillantoides to
spread widely through the continent and then proto-levaillantii. The small degree
of divergence shown between the isolated populations of both species (with the
exceptions of F. levaillantoides gutturalis in Eritrea which perhaps was isolated
earliest) suggests that these dry and wet cold spells were comparatively recent.

A summary of the evolution postulated is as follows:

Stage i. The group widespread in eastern Africa in a cold era.
Stage 2. The group split into a northern and southern block.

Stage 3. The southern block spreads north, at least to Tanganyika, and the northern
block south to Kenya (Map 76).

Stage 4. Warm and unfavourable, both blocks split into two : in the south proto-
africanus isolated in the mountains of South Africa, proto-shelleyi possibly in the
mountains of Nyasaland: in the north proto-psilolaemus isolated at high altitude,
possibly on Mount Kenya or Elgon, and the ancestral stock of the other species
isolated in the extreme north.
Stage 5. Favourable. Proto-shelleyi spreads, possibly to the limits of its present

160 B. P. HALL

range. Members of the northern blocks (with unpatterned abdomens) spread south
and west to Angola.

Staged. Unfavourable. Proto-finschi isolated first in Angola: proto-levaillantii
isolated next in some mountains of central Africa not occupied by proto-shelleyi,
while proto-levaillantoides is cut off in the north. Each becomes adapted to different
requirements of humidity.

Stage 7. In a long cold era favourable to the group as a whole, fluctuations in rainfall
and temperature favour each species in turn. In some unusually cold spell proto-
psilolaemus spreads to cover a large part of Abyssinia and Kenya but is later split
into isolated populations in the highest mountains. Proto-shelleyi following a
favourable era is split into three, from which the subspecies shelleyi, whyteiand
uluensis develop almost to specific status. An unfavourable spell in the north
isolates proto-gutturalis in Eritrea from the main body of proto-levaillantoides: in
a subsequent dry spell (during which the acacia spreads to link the northern and
southern belts) proto-levaillantoides spreads south to Bechuanaland. In a wet
spell pToto-levaillantii spreads at the expense of shelleyi and levaillantoides, and
proto-finschi spreads: where the two come together finschi is pushed into the
mountains.

Stage 8. The present. Too warm to be favourable to any members of the group,
the ranges of all species being discontinuous and the birds uncommon in most
areas.

THE RED-TAILED GROUP (MAP 8)
(F. coqui, F. albogularis, F. schlegelii}

RANGE AND CHARACTERS

This is a group of small francolins (wing under 150 mm.) which ranges from
Senegal to the Sudan, and from central Kenya west to the central Congo and
Angola, and south to the Transvaal and Natal: there is an isolated population in
the Rift of Abyssinia. Members of the group are found in woodland, savanna and
steppe and, in parts of the range, often on the same ground as other francolins, which
suggests that their smaller size may make them non-competitive. The group is
formed of three species which are allopatric except in the southern Congo and
eastern Angola where isolated populations of albogularis are found within the range
of coqui.

All members of the group have a considerable amount of ochre on the sides of the
face, spreading as an ill-defined collar on the hind neck : the crown is plain brown or
chestnut, contrasting with the back which has a quail-type pattern (except in the
females of schlegelii in which the striations and cross bars are reduced to a minimum),
varying in colour from grey and black to vinous rufous: in the redder birds the
rufous in the tail is conspicuous but in some of the greyer forms is no more than a
pinkish wash on the outer feathers: the underparts are variable but all birds have
a white or buff throat contrasting with the breast: all have black bills with some
yellow at the base and yellow legs : the males have a single spur.

THE FRANCOLINS, STUDY IN SPECIATION r6r

SPECIES AND POTENTIAL SPECIES

F. coqui is the most widespread of the three species which comprise the group.
It is found in short grass in brachystegia woodland from southern Uganda and
northern Tanganyika, southwards to the Rhodesias and westwards to Brazzaville
and the edge of the Angola escarpment, but is absent from mountains and from the
drier acacia steppe of Tanganyika. South of the brachystegia belt it is found in
the acacia savanna skirting the Kalahari, from northern South West Africa to the
Transvaal, and in the coastal districts of Natal. Scattered populations in Kenya,
southern Abyssinia and the Harar area are also found in acacia (Benson, 1945 : 390)
finding it common around Yavello and Mega in " open short-grassed country with
scattered low thorn bushes at 4,000-4,500 feet ".

In West Africa there is a population in the acacia belt of Senegal, the Upper
Volta and northern Nigeria that is isolated by 1,000 miles (and the Lower Guinea
forest) from the Brazzaville birds, and by nearly 2,000 miles from the Abyssinian
birds. In this area coqui is found commonly in cultivations, often alongside F.
bicalcaratus (Blasdale: unpublished). However, it seems that at Potiskum, 150 miles
east of Kano in northern Nigeria, coqui is replaced abruptly by F. clappertoni which
occupies the acacia belt eastwards to Abyssinia (Hollis: unpublished). In
Abyssinia clappertoni occupies the northern and western areas to the centre of the
Rift but is replaced by coqui in the Harar area in the east and in the Mega/Yavello
area of the south.

In coqui the quail-type patterning on the back is very consistent and well defined,
though the basic colour, and the colour of the head, vary from grey to brown or
rufous in different populations. The sexes differ, the males having the sides of the
face and throat ochre or light buff without a black eye-stripe or necklace, the
females having the sides of the face similar in colour to the males but the throat
whiter and having a black necklace and a black eye-stripe that continues as a black
line down the sides of the face. The males also have the breast barred while in
the females the barring is replaced or overlaid by a pink or grey wash.

There is little significant geographical variation in birds of the southern part of
the range, from Angola, the Congo, central Tanganyika and coastal Kenya south to
South Africa (coqui}. All have grey wings and barred abdomens (though the degree
of barring is variable) and females have a pink wash on the breast. In the
Abyssinian population (maharao) the wings are rufous and both sexes have light
narrow barring from breast to abdomen, the females having only a faint wash of
pinkish grey over the barring on the breast. The isolated West African population
(spinetorum) also has rufous wings, but is pale above, washed with pink, and there
is no barring on the abdomen in either sex: the breast of male is broadly barred
and that of the female plain and pinkish-grey. In Uganda, Kenya and Tanganyika
the pattern of variation is confused. East of the Rift in Kenya, south to the Arusha
area of Tanganyika, there is a population (thikae, Grant & Praed, 1934) known only
from a few specimens, which shows some approach to the Abyssinian birds in having
rufous wings, light barring, and vestigial barring on the breasts of the females, but
with no barring on the abdomens in either sex. From the Rift westwards to the
eastern shores of Lake Victoria the birds (hubbardi) are similar to both the West

162 B. P. HALL

African spinetorum and the eastern thikae in having plain abdomens but the wings
are grey like the southern coqui, the breasts of the females are plain and grey, and the
whole upperparts are rather dark and grey. Birds from west of Lake Victoria
(ruahdae) are close to the southern coqui, but have the black barring broader and more
widely spaced. These birds intergrade with hubbardi in the Tabora area.

The second species in the group, F. albogularis, has a scattered distribution.
There is a population in the Gambia, found mostly on the edges of cultivation, but
it is rare (Hopkinson, 1923: 129). Other populations are found on the coastal
plain near Accra and in the hinterland of Ghana, eastwards through Nigeria to
Garoua in the (French) Cameroons. Field notes on these birds are scarce and
somewhat contradictory, Serle (1957:389) finding them common in "the open
savannas and rolling hill country " of Eastern Nigeria, whereas Blasdale tells me
that further north he has found them only in thick woodland where there is
abundant Isoberlinia doka. Most localities from which they have been collected
lie within the woodland belt (Map n, Type 17) but Garoua lies within the acacia
savanna belt (Map n, 20) and the specimen was obtained in " open, sandy, sparsely-
treed country " (Good, 1952: 67).

F. albogularis is absent from the woodland and savanna belt north of the Congo
forest (where F. schlegelii is found) but reappears again 1,500 miles away in the
south-eastern Congo on the plateaux of Marungu and Kibara, in the Upemba National
Park, on the plains of the Balovale district of Northern Rhodesia, and on the treeless
dambos of the Zambezi/Congo watershed in eastern Angola.

Above F. albogularis is very like F. coqui but the quail-type patterning is less
well defined in the females, the shaft streaks and barring being narrower: the wings
are more rufous (similar in this respect to the east Kenya population, thikae). As
in coqui the females have a black facial pattern and necklace, but it is ill defined,
and the males have none, though they differ from coqui in having a white throat
contrasting with the ochre cheeks, as in the females. In the rest of the underparts
the males of albogularis are quite distinct from any form of coqui being chestnut on
the breast with ochre shaft streaks and rich ochre on the abdomen, lacking any dark
barring. The females on the other hand are barred, in greater or lesser degree
according to the population, with a faint wash of pink or rufous on the breast, and
resemble closely the females of the Abyssinian and east Kenya populations of
coqui. The population of albogularis in which the females are least patterned is
in the Gambia (nominate albogularis) the barring being light and restricted to the
breast and flanks (though varying individually). In birds from Ghana to the
Cameroons (buckleyi) the barring is generally heavier, usually with only the centre
of the abdomen unpatterned. Birds of both sexes from the populations south of
the Congo forest are slightly larger and more richly coloured than those from West
Africa and in the females of the south-east Congo (dewittei) the barring is heavier
but no more extensive than in buckleyi: the darkest birds with the most heavily
barred females are found in eastern Angola and possibly the Balovale district
(meinertzhageni White, 1944 see appendix).

The third species of the group, F. schlegelii, has been considered by Chapin and
Peters to be a subspecies of F. coqui. I believe it is distinct enough to warrant

THE FRANCOLINS, A STUDY IN SPECIATION 163

specific status and furthermore is closer in several respects to F. albogularis.

It is a rare bird, found scattered through the woodland and savanna belt from the
eastern Cameroons to the western Bahr-el-Ghazal. From what Cave tells me it
seems to be more consistently a woodland bird than albogularis and he has the
impression that it is found in rather denser woodland than the sympatric F.
icterorhynchus of the Vermiculated Group, especially where there is Isoberlinia doka.

Above schlegelii is closer to albogularis than coqui, though the quail type patterning
is much reduced and the sexual dimorphism is more pronounced, the males having
broad white shaft streaks but few transverse bars while the females are almost
unpatterned: both sexes have a soft vinous wash. Below the male of schlegelii is
like that of southern races of 'coqui, with ochre sides of the face and throat and
narrow black and white barring over the breast and abdomen : the female resembles
the male on the head and throat but has the breast grey, patterned with cream
triangular marks at the base of the feathers, the flanks sparsely barred, and the
centre of the abdomen plain cream.

OTHER VARIATION

In the southern populations of F. coqui there is local and ecological variation in size and
colour; exceptionally small birds are found in coastal Kenya and Natal; grey, pale birds in
Bechuanaland (vernayi) and pale, rather pinkish birds in South West Africa (hoeschianus
Stresemann, 1937) : other exceptionally grey birds occur in southern Mozambique (Coguno) and
may be expected locally in dry conditions, while pockets of more richly coloured birds are
associated with richer vegetation. There is otherwise little variation between birds from Natal
and the Transvaal north to coastal Kenya and west to Northern Rhodesia, except in the
extreme west: here and in Angola the birds are consistently less rufous and the bars on the
underparts are narrower and rather closer together (angolensis) . Birds from Kasai and
western Katanga (kasaicus White, 1945) are similar to coqui above and angolensis below: they
intergrade with coqui in the eastern Katanga.

In F. albogularis of West Africa individual variation in patterning has already been noted.
Serle (1957: 388) also found the general colour individually variable between rufous and grey,
though there seems a tendency for the coastal population of Accra to be greyer than inland
birds.

Individual variation is also found in F. schlegelii especially in the females (see Tray lor, ig6ob:
86).

EVOLUTION

The most striking features of the pattern presented by the Red-tailed Group
are the enormous distances separating the north-western and south-eastern sub-
species of both F. albogularis and F. coqui: also, in F. coqui, the complexity of
variation in East Africa in contrast to the uniformity of southern birds. I feel
confident that the discontinuous range of F. albogularis is largely attributable to
the presence of F. schlegelii in the intervening segment of the same vegetational belt,
and suggest that the ancestral stock of the two species once had a continuous
distribution in the savanna and woodland encircling the Congo forest. To account
for the split into two species I suggest this era was followed by a very dry period
in which the forest was divided into a western and eastern block each encircled by
woodland and separated by a corridor of acacia linking the northern and southern

164 B. P. HALL

acacia belts. It can reasonably be postulated that proto-albogularis developed in
the belt round the western forest and proto-schlegelii in the belt round the eastern
forest (Map SB). When the two forest blocks rejoined with a return of humid
conditions and the surrounding belts became one again, the central segment of the
belt was occupied by proto-schlegelii and the western and southern segments by
proto-albogularis.

It is more difficult to interpret the pattern of F. coqui but a logical sequence can
be found which will fit into that postulated for the other species. Proto-coqui
may have been separated from the ancestral stock of the other two species at a
time of spread of montane forest and developed in the east while proto-albogularis /
schlegelii developed in the north. With the retreat of the forest proto-coqui spread
south and west chiefly in acacia savanna. In an unfavourable spell, either too
wet or too dry, proto-maharao in Abyssinia was isolated. In the very dry era
postulated above, in which there was a corridor of acacia in the central Congo,
proto-coqui could have spread into the northern acacia savanna belt. Subsequently
with a return of humid conditions the northern population would be cut off from
the southern by woodland and forest. The southern populations nearest the Equator
prove adaptable so that, with the return of woodland and savanna and the in-
digenous albogularis to the southern Congo and Angola, coqui competes successfully,
occupying the woodlands and albogularis the more open grasslands and plateaux.

Meanwhile the isolated northern population diverges from southern birds, develop-
ing the distinctive unpatterned abdomen in both sexes, and spreads through the
northern acacia belt west to Senegal and east through the Sudan to Kenya and
Uganda. Here it encounters both the Abyssinian maharao and the southern coqui.
The characters of the present populations of Kenya (hubbardi and thikae) and of
southern Uganda (ruahdae) suggest that some interbreeding took place at some
period.

The present appears to be favourable to the Vermiculated Group in the north,
at the expense of the Red-tailed Group. In the immediate past it seems that
F. clappertoni has spread west through the acacia belt in the Sudan, eliminating
coqui between the Nile Valley and northern Nigeria. In the woodland and savanna
belt F. bicalcaratus flourishes in the west and F. icterorhynchus in the east, and
competition with these two may account for the scarcity and discontinuous dis-
tribution of albogularis and schlegelii. However, in West Africa coqui, though
apparently unable to compete with clappertoni, is yet able to live on the same
ground as bicalcaratus on the southern fringe of the acacia belt. The populations of
coqui in eastern and southern Africa also seem able to co-exist on the same ground
as other francolins.

A summary of the evolution postulated is as follows :

Stage i. The group develops in savanna, perhaps in north-eastern Africa.

Stage 2. An unfavourable era, possibly a spread of montane forest. The group
splits into two with proto-coqui isolated in the east and proto-albogularis I schlegelii
to the north of the forest.

Stage 3. More favourable. Proto-coqui spreads south and west in rather dry

THE FRANCOLINS, A STUDY IN SPECIATION 165

savanna, mostly acacia. PToto-albogularis/schlegelii spreads in the wetter savanna
and woodland encircling the lowland forest of the Congo Basin (Map 8 A).

Stage 4. A very dry era. The Congo forest is divided into a western and eastern
block, each surrounded by its own woodland and savanna belt, with a corridor of
acacia between. Proto-albogularis develops in the belt surrounding the western
block of forest, proto-schlegelii in the belt surrounding the eastern block. Proto-
coqui spreads through the acacia corridor into the northern acacia belt. In an
unfavourable period in the north-east proto-maharao becomes isolated (Map SB).
Stage 5. More humid. The two blocks of forest rejoin and the two savanna belts
become one again, the eastern segment being occupied by proto-schlegelii and the
western and southern segments by proto-albogularis. In the southern segment, as
woodland and savanna return to districts previously acacia, proto-coqui is not
pushed out but becomes adapted to woodland and competes successfully with
proto-albogularis which takes to the more open grasslands. In the north proto-
spinetorum subsp. spreads west to Senegal and east to Abyssinia and Kenya and
diverges from the southern birds. There is limited interbreeding where the northern
birds re-encounter other members of the species in the east.

Stage 6. Favourable for the Vermiculated Group, especially in the north. F.
clappertoni spreads, eliminating F. coqui from the acacia belt in the Nile Valley and
Sudan westwards to northern Nigeria. In the woodland and savanna belt F.
bicalcaratus and F. icterorhynchus flourish at the expense of F. albogularis and
F. schlegelii. In the south F. coqui co-exists more successfully with francolins of
other groups.

SPECIES NOT INCLUDED IN ANY GROUP (MAPS 9 & 10)

Francolinus lathami

F. lathami is a forest francolin ranging from Sierra Leone to western Uganda and
the southern Sudan.

It appears at first sight quite unlike any other African francolin (closer, except in
size, to F.francolinus of Asia), being, in the male, largely black below with regular
white spots, in the female brown with irregular white spots, the throat black and
face patterned, the upper parts mottled rufous and brown with some white streaks
on the mantle, the bill black, finer than in other francolins. However, comparison
of a pair of lathami with a pair of schlegelii suggests that lathami may be more
closely related to the Red-tailed Group than to any other francolin. It will be
seen that the pattern and the colouring on the upper parts are basically the same,
though lathami, as would be expected in a forest bird, is darker and more mottled.
Below the patterning of the females has some similarity in the irregular, rather
triangular shaped spots: while the spotted male of lathami looks very different
from the black-and-white barred male of schlegelii, this may not be a very significant
difference for in odd abdomen feathers of several lathami the spots are less regular
and merge as vestigial bars. The chief differences in fact lie in the shape of the
bill and the pattern of the face and throat. Furthermore the first plumage of
lathami chicks is very similar to the plain vinous mantle of the female schlegelii.

166 B. P. HALL

While being confident that there are affinities between lathami and schlegelii I
am not prepared to guess at what period in history proto-lathami became isolated
and adapted to forest. The likeness to one particular species in the Red-tailed
Group suggests that its isolation could have been subsequent to the emergence of
species within the group, the adaptation to such different habitat accelerating diver-
gence. On the other hand F. schlegelii is the species of the Red-tailed Group that
lives in the thickest woodlands and, if proto-lathami had broken from the ancestral
stock of the Red-tailed Group before the three species had diverged, the similarities
between lathami and schlegelii might be due to convergence. For this reason, and
with regard to both its morphological and ecological distinctions, I feel it best to
treat lathami independently.

It will be seen that I am not prepared to regard the black-and-white spotting
in the males of lathami as indicative of relationship with F.francolinus. It has
been amply demonstrated in the discussion on groups that the pattern of the
underparts is of less significance than size and the pattern above, and while it is
interesting that such a striking pattern is repeated in the genus in two such different
birds, it cannot be relevant to the grouping of species.

From the distribution plotted on Map 9 it would appear that there is a wide gap
between the populations of the eastern and western Congo, but this is more likely
to be due to lack of collecting in the central area than to a genuine break. There
is little geographical variation, in the males of eastern birds (schubotzi) the black-
and-white pattern extends further down the abdomen, and the females have the
cheeks more rufous than grey, but there is considerable individual variation.

Francolinus nahani

This little francolin is confined to the heaviest forests of the north-eastern Congo
and Uganda. It is a distinctive bird, predominantly dark brown or blackish, with
white streaks on the underparts and some white spots on the wing-coverts and
flanks. The bill is black at the tip and crimson at the base; there is a patch of
crimson bare skin round the eye; the legs are crimson, and without spurs in either
sex.

It is difficult to assess its affinities with the other francolins. Superficially it
appears closest to F. lathami in its small size, general dark coloration and suggestion
of spotting, and they occupy the same type of habitat. On the other hand the
similarities are in those characters that are to be expected in forest birds, and could
well be due to convergence, and the fact that F. nahani in the whole of its range is
sympatric with F. lathami, without apparently being ecologically segregated, does
not suggest close relationship. I would suggest very tentatively that F. nahani
may have broken off at a remote period from the ancestral stock of the Scaly Group
(in the same way as I have suggested that lathami may have broken off from the
Red-tailed Group), and become adapted to heavy forest instead of the forest clearings.
I suggest this because nahani, although much smaller, seems to have some characters
in common with the Scaly Group as a whole and with each of the three individual
species. In both nahani and all species of the group the sexes are alike, and the
legs are red: nahani is like F. squamatus above, having an unpatterned head and

THE FRANCOLINS, A STUDY IN SPECIATION 167

vermiculated back; it is like F. ahantensis in the white streaking below; and like
F. griseostriatus in the crimson base to the bill. Furthermore in all species of the
Scaly Group there is a small area of bare skin below and behind the eye, though this
is not brightly coloured and usually overlooked.

Chapin considered the absence of spurs to be a generic character and proposed
the monotypic genus Acentrortyx for F. nahani, but the fact that F. pictus in India
is also without spurs, but yet is so closely related to F. francolinus as to hybridise,
demonstrates that this character is not necessarily significant. I see no other reason
for regarding F. nahani as generically distinct.

F. pondicerianus

F. pondicerianus is the smallest of the Asiatic Francolins and is widespread from
the shores of the Gulf of Oman eastwards through the plains of India and the dry
north-western districts of Ceylon, living in grassland or scrub. It appears to have
some affinities with both the Red-tailed and the Striated Groups of Africa and in
colour and pattern is closest to F. coqui, in particular to the geographically nearest
form F. c. maharao of Abyssinia. It has " quail-type " plumage on the back,
but is duller than coqui, with broader transverse white bars and narrower shaft
streaks, and with chestnut markings on a grey or huffish ground: the underparts
are whitish, irregularly barred with very fine blackish bars: the throat is buff
surrounded by a blackish necklace, which is smaller than that found in the females
of coqui and starts from the chin and not the sides of the bill. In other respects
it is closer to the Striated Group (F. sephaena and F. streptophorus) than to the
Red-tailed Group, having a longer and more graduated tail (the tail more than half
the length of the wing), and stronger (though not necessarily longer) legs and larger
feet, the legs being reddish as in F. sephaena. The bill is distinct from either group
being brownish in colour and heavier in front of the nostrils. The tail in colour
combines characters of both groups being reddish at the base but otherwise largely
blackish.

The affinities with two of the African Groups suggest that the ancestral stock
of F. pondicerianus was isolated from the African stock before the Red-tailed and
Striated Groups had diverged and for this reason I have discussed it independently.
It seems that, while the African stock underwent many vicissitudes of climate with
consequent isolating and regrouping of populations resulting in speciation, the
Indian stock had a comparatively tranquil history for in pondicerianus there is
little variation, and that mostly clinal and ecological: birds from the dry districts
of Persia and Baluchistan, east to the western Punjab (mecranensis) , are the greyest
and palest, and those from southern India (pondicerianus) the darkest and least
grey with the greatest amount of chestnut markings. Those from northern and
central India.(interpositus) are intermediate. Ceylon birds (ceylonensis Whistler, 1941)
are greyer than either pondicerianus or interpositus but darker than mecranensis.

F. gularis

F. gularis is the swamp francolin of India, being confined to reeds and marshes
in the plains of the Ganges, Brahmaputra and their tributaries from central Uttar

168 B. P. HALL

Pradesh eastwards into Assam. It is a large francolin, comparable in size and
general proportions to the largest member of F. francolinus, but with rather a shorter
bill and larger feet (as would be expected in a marsh bird). It is, however, quite
different in colour and pattern being predominantly a brown and white bird, with a
barred back, streaked underparts and a rufous throat, wings and tail. The sexes
are alike.

Proto-gularis seems likely to have become isolated from other francolins in a dry
period and taken refuge in river valleys, becoming adapted to these conditions.
The dry era postulated may even have occurred before the link between the Asian
and African francolins was broken since the differences from other francolins shown
by F. gularis, both in ecology and morphology, suggests divergence over a long
period.

CONCLUSIONS
THE GROUPS

The chief lesson to be learnt from comparison of the relationships between the
members of each group re-emphasises the difficulty of assessing from morphological
characters the level of speciation which any form has attained. This is brought
home most forcibly by finding that such dissimilar birds as cranchii and afer in the
Bare-throated Group will interbreed freely, while the three Redwings in the Trans-
vaal shelleyi, levaillantoides, and levaillantii, which look more alike, are able, on
account of ecological divergence, to live in a small area without interbreeding.

It seems that a relatively slight ecological divergence among the francolins may
be effective against interbreeding because their sedentary habits and specialised
choice of habitat make them reluctant to move out of their selected type of vegeta-
tion. These same habits allow also for comparatively small barriers to > act as
isolating factors between two populations that are not necessarily ecologically
divergent. Two instances have been given where rivers appear to form such a
barrier, the Niger between ahantensis and squamatus of the Scaly Group and the
lower Zambezi between natalensis and hildebrandti of the Vermiculated Group.
In the Red-winged Group fifty miles of unsuitable country in Tanganyika between
F. s. shelleyi at Salanga and F. s. uluensis at Mbula serves as barrier to interbreeding
between them.

The difficulty of designating in these circumstances which members of groups
should be called species and which subspecies was discussed at the start of the
paper, for the status of most lies somewhere between the two. This should be
borne in mind in the subsequent discussion on evolution, for these " species ", which
are allopatric members of superspecies, for the most part appear to have attained
a level of speciation far below that of those species which are able to live with each
other in full sympatry.

EVOLUTION

The stages of evolution postulated for each group have been based on what seem
to me to be the simplest and most likely interpretations of the patterns shown, and
they represent the minimum number of stages which can be envisaged to produce

THE FRANCOLINS, A STUDY IN SPECIATION

169

such a pattern. They represent a sequence rather than a time-table of change and
as such could, in theory, be spread over any number of years. However, there is
evidence from the distribution of many montane plants and trees, as well as birds,
which shows that there have been periods in which montane forest has spread
over vast areas of Africa, linking up the present isolated patches in the Cameroons,
Angola, East Africa, Abyssinia, Nyasaland and South Africa, descending as low as
2,000 ft. (see Map 3 A). At the same time (by inference a wet era) there would be
extensive lowland forest below that level. This would have the effect of reducing
the refuges for grassland and woodland species to comparatively small areas isolated
from each other by forests. Such conditions would not only be catastrophic for
many species but would obliterate any traces of the pattern of distribution and
evolution prior to that period. Thus, although such a spread of forest may have
occurred during any extensive cold wet era, it follows that, if the stages of evolution
I have suggested are accepted, the conclusion cannot be evaded that they have
been accomplished since the last spread. This conclusion carries with it implications
on the rate of speciation which need to be examined closely.

In the course of these stages I have suggested that twelve species of francolins
(the ancestral stock of the eight groups and the four independent species) have
increased to forty-one, all the new species being members of superspecies. The
accompanying dendrogram (Fig. 2) summarises the relationship postulated of the

squamatus
ahantensis
griseostriatus

adspersus

capensis

natalensis
hartlaubi

hildebrandti

icterorhynchus
clappertoni
bicalcaratus
bar wood i

swainsonii
rufopictus

leucoscepus

francolinus

pictus
pintadeanus

gularis-

africanus
shelley i
.psilolaemus
levaillantoides
levaillanti
finschi

streptophorus
sephaena

.pondicerianus

Francolinus

FIG. 2. Dendrogram showing the relationship postulated of the various groups and species,

and their date of origin in respect to the last major spread of montane forest.
ZOOL. 10, 2

iyo B. P. HALL

various groups and species to each other and their date of origin in respect of this
last spread. Unfortunately no date can be given yet to this last spread of forest,
and there is indeed still uncertainty on the date of onset of the Last Glaciation,
opinions varying between 100,000 and 50,000 years ago. Authorities, however,
agree that it ended about 18,000 years ago and that there have been at least four
glaciations in the course of the Pleistocene. I am suggesting therefore that these
changes have taken place in the comparatively short period of the Late Pleistocene
(i.e. about the last 100,000 years).

If we accept that Francolinus may (among other modern genera of birds) have
originated in the Oligocene (approximately forty to thirty million years ago) (Van
Tyne & Berger 1959 : 6-17) and has taken all the time prior to the last glaciation
to divide into twelve species, it seems that I am postulating an incredible acceleration
in the rate of increase of species. But this is not so, for, as stated, the concept of
periodic spreads of forest over vast areas of Africa is inseparable from the acceptance
of a high casualty rate among the non-forest species at that period. If, for instance,
the present was followed by a cold, wet era that brought with it a wide spread of
forest, it is highly unlikely that any of the less successful non-forest species of
tropical Africa would survive (e.g. rufopictus, streptophorus , finschii, albogularis,
schlegelii) .

Competition between members of different groups in the limited refuges left
would eliminate several of the apparently successful species as well (just as F.
dappertoni appears to have eliminated F. coqui from the Sudan) , while members of
the same group, forced into unprecedented proximity, would probably interbreed
freely. Thus the total number of African species could again be reduced to few
(if any) more than the nine which survived the previous forest spread. These would
be the ancestors of future groups.

Thus while I believe the number of members in any group may fluctuate widely
with successive glacial and inter-glacial eras it seems that any increase in the number
of groups and independent species would be very slow and not incompatible with
the rate of evolution since the inception of the genus.

It could also be suggested that any further increase is unlikely in this figure,
which represents the number of birds which have reached a sufficiently high level of
speciation to be able to live sympatrically without interbreeding, for it is evident
that only a limited number of such birds can be supported in any given area. In
birds like the francolins, in which two species are rarely found actually on the
same ground, this limit is governed largely by the number of alternative types of
habitat, and the unequal distribution of the groups in tropical Africa gives some
reason to suppose that this limit may have been reached.

Disregarding the Montane Group, with its extremely specialised ecology, it will
be seen that, apart from a relict population of the Striated Group, only three groups,
the Scaly, Vermiculated and Red-tailed, are fully represented in the north-western
section (from West Africa to the Sudan and Uganda) . There is little overlap between
the ranges of these groups and where two do occur together one is thriving at the
expense of the other. Furthermore over a vast area it appears that members of the
most successful group, the Vermiculated, have totally driven out representatives of

THE FRANCOLINS, A STUDY IN SPECIATION 171

other groups, namely the red-tailed F. coqui from Lake Chad to the Nile, and the
striated F. streptophorus between the Cameroons and Uganda. This north-western
section is relatively flat with monotonous vegetation, and no very clear-cut divisions
between the vegetation belts : the east and south on the other hand, from Abyssinia
to South Africa and Angola, is more mountainous in many parts with more varied
vegetation, and here no less than six groups are sympatric or partly sympatric.
In this section of Africa I have suggested certain instances in which the range of one
group has been restricted by competition with another, but it is also evident that
several special may thrive in a small area.

I am not competent to discuss the situation in Asia where the development and
status of the francolins must be dependent on those of the numerous other genera
of gamebirds which I have not studied, but it seems unlikely that there is much
room for an increased number of full species.

In arguing that the members of groups may have developed to the level of members
of superspecies within the Late Pleistocene I am conscious that this is postulating
a more rapid rate of speciation than is usually accepted, though Rand (1951: 567)
has suggested that some of the forest subspecies of West Africa may have evolved
in the last 12,000 years, while Moreau (1930) gives examples of Egyptian subspecies
which have probably evolved in 5,000 years. However, some weight is added to
my argument by the evidence collected by Moreau on the climatic fluctuations
that have taken place in this period (Appendix i), which shows that there has been
sufficient variation in both temperature and humidity to bring about the wide
redistributions of the montane and lowland forest, woodland, savanna and steppe
that would be necessary to the different stages proposed. Furthermore, it shows
that all the vegetational changes postulated in the evolutionary sequences in the
groups (which were founded entirely on the evidence of the birds alone) did, or could,
take place, though it is not always easy to satisfy the sequence.

To support some of the major changes which I have envisaged, he has shown, for
instance, that there have been very dry periods prior to 10,000 years ago in Angola
and Katanga (at least) which, if they prevailed elsewhere, could have been sufficient
to connect the northern and southern acacia belts in the way postulated in the
evolutionary sequences of the Bare-throated, Striated and Red-winged Groups. He
gives corroboratory evidence for the existence of such a dry corridor at some period
in Tanganyika, and Benson and White (in press) suggest this corridor may have been
through the Luangwa and Rukwa valleys, and they list many species of birds and
mammals whose present distribution has led them to this conclusion.

Moreau shows also that there was a period (undated) at which Lake Chad was
600 miles long, probably with forest at the southern end joining the main Congo
forest, so forming a barrier between eastern and western woodland or grassland
species, such as was postulated in the Vermiculated Group to split proto-bicalcaratus
from proto-icterorhynchus. He quotes too a record that suggests there was probably
lowland forest at Oldowai in northern Tanganyika in the early Pleistocene where there
is now acacia steppe. Though this dating can hardly be correlated with the spread of
lowland forest envisaged in the later stages of the evolutionary sequence of the
Scaly Group, it suggests that such a spread has taken place once and could

172 B. P. HALL

therefore be repeated in similar climatic conditions: perhaps achieved in even one
of the recent minor pluvials in East Africa (less than 35,000 years ago).

Similarly he says that late in the Middle Pleistocene there is believed to have been
a dry period when the Congo forest was reduced to strips in the east and west (and
possibly also on the northern rim) and that there was acacia in the country between.
This makes the sequence of evolution postulated for the Red-tailed Group less
improbable than it seemed when I first proposed it (before I had seen Moreau's
paper), though this sequence remains the most difficult to correlate with known
climatic sequences.

Finally, the undated dry period in the Late Pleistocene when the Sahara advanced
300 miles southward may perhaps be correlated with the stage postulated in the
Striated Group when dry conditions prevailed on the ridge of higher country between
the Cameroons and Uganda on which proto-streptophorus may eventually have
been isolated.

All the other climatic changes I have postulated in Africa are relatively small
and the occurrence of similar changes is unquestionable, though not necessarily in
correct sequence. However, some of the dating for these in West Africa is interesting
for the most recent humid period in the north-west (5,000-2,000 years ago) may be
the time at which F. Ucalcaratus spread into Morocco up the north-west coast, and
if so the Morocco population, which shows only slight divergence, may have been
isolated for only 2,000 years. The lack of divergence between the populations of
F. ahantensis on either side of the present gap in the Upper Guinea forest also gives
support to Moreau's evidence that the gap was once further east " probably as
late as the Late Pleistocene ".

In the Spotted Group of Asia, as in the African groups, it was postulated that the
members developed to specific status after the last glaciation. While it is easier
here to argue that the same sequence of isolating and rejoining of species was likely
to take place in any glaciation it seems unlikely that the simple pattern which
remains now could be the product of more than one such sequence.

Finally, lest any of the foregoing should sound dogmatic, it is necessary to end on
the same note of caution that was sounded in the introduction. I am fully aware
that all the conclusions are based on circumstantial evidence drawn from my
interpretations of the picture as we see it today, which are necessarily fallible since
so much of the picture is missing. That it has proved possible to construe the known
facts into a logical and coherent story gives some verisimilitude to it and so does
the evidence that many of the postulated vegetational changes have actually taken
place. However, the story remains hypothetical, but further studies of this nature
and researches in other branches of science will corroborate or disprove others of
the possibilities suggested so that eventually a picture may emerge that approximates
to reality.

SUMMARY

i. It is found that the genus Francolinus (including Pternistis) is comprised of
eight groups, the members of which are largely allopatric, and four independent
species.

THE FRANCOLINS, A STUDY IN SPECIATION 173

2. Some of the difficulties are discussed of denning the relationship between the
members of the groups.

3. The value of different characters as a guide to relationship is illustrated by
comparison between the groups.

4. Each group is studied in turn and the characters and ranges of its component
members defined. In doing this it was found necessary to review the systematics
of the Redwing and Grey wing Francolins of the species africanus (olim. afer),
levaillantoides (olim. gariepensis] and shelleyi.

5. An attempt is made to postulate the sequence of evolutionary stages through
which each group has developed, and the climatic and vegetational changes which
have accompanied each stage.

6. The four independent species are discussed.

7. Comparison between groups suggests that slight ecological divergence may be
an especially effective barrier to interbreeding.

8. It is concluded that most members of groups reflect in their characters,
relationship, and present ranges, the climatic changes that took place in the Late
Pleistocene, subsequent to the last glaciation, or, in Africa, the last extensive spread
of montane forest. Consequently it is suggested that the number of species of this
status is liable to fluctuate considerably in glacial and inter-glacial periods, though
the number of groups and independent species may remain more or less constant.

9. These conclusions are supported to some extent by evidence provided by
R. E. Moreau of climatic changes that have taken place in the Late Pleistocene
(Appendix i), but more evidence is needed from all branches of science to evaluate
them fairly.

APPENDIX i

Summary of some conclusions reached by R. E. Moreau on the climatic fluctuations
in Africa as set out in his paper " Vicissitudes of the African Biota in the Late
Pleistocene " (in press : to be published in Proceedings Zoological Society, London).

Africa generally (in conformity with world temperatures)

Colder than the present (with associated reduced humidity) in varying degrees for
at least 50,000 years prior to 18,000 years ago. In the last extreme phase, extending
from about 24,000-18,000 years ago the cooling amounted to about 5^- Around
7,000-5,000 years ago probably 2C. warmer, and for some centuries prior to 150
years ago 1-2 C. cooler, than the present.

Mediterranean Africa and Northern Sahara

A major pluvial period affected at least the western half of the area in the late
Middle Pleistocene, prior to perhaps 100,000 years ago. Palaearctic plants and land-

Footnote : Mr. Moreau has very kindly allowed me to include this summary to illustrate my paper
but it must be appreciated that most of the conclusions quoted have been qualified in the full text, where
reference is given for the authority on which they are based. This summary must not therefore be used
as a base for further premises without consulting the full paper. It must be emphasized that the time-
correlations of the events referred to in different parts of Africa remain to be worked out.

1 74 B. P. HALL

molluscs advanced some 400 miles southward and persisted in Ahaggar until the
Neolithic, some 6,000 years ago.

Between 5,000 and 2,000 years ago it was more humid (at least in the north-west)
than at present.

Southern edge of Sahara and sub-Saharan West Africa

The southern edge of the desert has fluctuated in the course of the Late Pleistocene
(i.e. about the last 100,000 years) between about 300 miles south and 300 miles north
of the present position and these changes probably extended across to the White
Nile. Since the Neolithic, i.e. in about the last 5,000 years, a belt of country some
300 miles in width has reverted to desert.

These fluctuations affected the vegetation of coastal West Africa, especially the
Upper Guinea forest, and when the desert came south of its present location the area
of forest was greatly reduced and further dissected.

At some stage in the Late Pleistocene, not yet dated, Lake Chad increased to an
area 600 miles long, extending nearly to Tibesti. This implies a contemporary north-
ward extension (or movement) of the equatorial rain-belt by some 300 miles.

Prior to its capture by the Lower Niger, some 10,000 years ago, the Upper Niger
expended itself in large lakes northwest of Timbuktu, in what is now extreme desert.

Congo basin and Angola

Late in mid-Pleistocene very dry, with aeolian re-distribution of Kalahari sand.
Congo forest reduced to a strip towards the coast, another along the eastern rim of
the basin and perhaps a third along the northern edge. Throughout Late Pleistocene,
humid except in Katanga and Eastern Angola. Here arid in the latter part of the
Late Pleistocene down to about 10,000 years ago.

East Africa

The " Gamblian " pluvial, with fluctuations in intensity was contemporary with
the Last Glaciation and extended down to about 18,000 years ago, and there have
been subsequently at least two short periods more humid than the present. Owing
to the broken topography the vegetational effects would have been, in the main,
local advances and retractions of vegetation types. In Somaliland the fluctuations
were minor and the country always to some degree dry. (Note : There is evidence
that at some period in the Early Pleistocene lowland forest extended at least to
Oldowai in northern Tanganyika. Also at some period in the Late Pleistocene a
corridor of acacia probably extended through Tanganyika to Northern Rhodesia).

Rhodesia and southern Africa

During the Late Pleistocene the rainfall fluctuated between about half and one and
a half times the present in the Rhodesias and the Transvaal. Kalahari sand was
blowing east as far as the Victoria Falls some 12,000-9,000 years ago. The Orange
Free State was drier than the present some 40,000 years ago. South West Africa was
never better than semi-arid,

THE FRANCOLINS, A STUDY IN SPECIATION 175

APPENDIX 2

Taxonomic notes and synonymy

SPOTTED GROUP

F. francolinus

Five clinal races have been recognised and further sub-division does not seem useful. There-
fore caucasicus and sarudyni are placed in the synonymy of francolimts ; festinus Koelz, 1955
and bogdanovi in the synonymy of henrici ; parkerae Van Tyne & Koelz, 1955 in the synonymy
of asiae. Also billypaynei Meinertzhagen, 1933, which is a name given to an exceptionally dark
micro-population from the Lake of Antioch, is included in francolinus.

F. pintadeanus

The characters on which wellsi was based do not seem constant and it is accordingly regarded
as a synonym of pintadeanus.

I accept Delacour's opinion (1951 : 10) that -F. boineti Bourret, 1944 is an aberrant specimen of
F. pintadeanus.

BARE-THROATED GROUP

F. leucoscepus

The amount of variation exhibited by the whole species is so slight that numerous sub-
divisions are not practical. Accordingly holtemulleri, muhamedbenabdullah, keniensis,* kili-
mensis, tokora, oldowai Van Someren, 1939, are placed in the synonymy of infuscatus.

F. swainsonii

Since Pternistis has been submerged in Francolinus, P. s. chobiensis Roberts, 1932, is pre-
occupied and F. s. lundazi used instead (see Benson & White, 1957: 138).

A record of the occurrence of F. swainsonii at Karonga, at the north end of Lake Nyasa
(Long, 1961 : 25) was based on a misidentification, the specimens being F. afer (Benson,
in lift.).

F. afer

In the block of vermiculated subspecies there is much individual but little geographical
variation. Accordingly punctulata and nyanzae are placed in the synonymy of cranchii.

In the black-and-white birds of Angola and South West Africa variation between the northern
and southern birds is too slight to warrant recognition, so palliditectus White, 1958 (nom. nov.
for cunenensis Roberts) is placed in the synonymy of afer. In South Africa an intermediate
population between castaneiv enter and notatus was named krebsi, which I place in the synonymy
of castaneiv enter. In eastern Africa the name humboldtii was given to a bird from Tete on the
Zambezi, a locality lying in the centre of the zone of hybrids between the black-faced melano-
gaster and the white-faced swynnertoni : it is here restricted to the unstable population.

If names are required for the hybrids between the vermiculated and black-and-white sub-
species it is convenient in Angola to refer those populations which are closest to cranchii to
manueli White, 1945 (synonyms camabatelae Meise, 1958, and mackenziei White, 1945). an ^
those which are closest to afer to benguellensis (synonyms angolensis Bocage, angolensis Roberts,
and chio Meise, 1958 see Traylor ig6oa: 143-146). Similarly in Tanganyika, Nyasaland and
the Luangwa valley those populations closest to intercedens may be known as bohmi (synonym
tornowi Meise, 1933), and those closest to melanogaster may be known as itigi (synonym
aylwinae White, 1947).

With the merging of Pternistis in Francolinus F. coqui angolensis Rothschild, 1902, is ante-
dated by Pternistes afer var. angolensis Bocage, 1893, and by the laws of homonymy should be
renamed. However Bocage 's name angolensis was given in error to the same bird which he had
named benguellensis at the head of his article, and the name was forgotten until listed in the
synonymy of Pternistis afer benguellensis by Bowen (1930: 15?) Jt thus has remained unused
as a senior synonym in the primary zoological literature for more than fifty years and must
be considered a forgotten name (nomen oblitum). (Int. Code of Zool. Nomen. 1961. Article

* Also pre-occupied by F. squnmatus keniensis Mearns, 1910.

176 B. P. HALL

23b.) Application is being made to the International Commission for it to be placed on the
Official Index of Rejected Names.

MONTANE GROUP

F. castaneicollis

It is probable that most of the micro-populations in isolated patches of evergreen forest are
slightly distinct, but it is not practical to recognise by name all those that show minor
differences: accordingly vottegi and gofanus are placed in the synonymy of castaneicollis, and
patrizii Toschi, 1958, in the synonymy of kaffanus.

F. jacksoni

It has not been established that variation in the amount of white in the underparts is geographi-
cal or associated with altitude. Therefore gurae is placed in the synonymy of jacksoni.

F. nobilis

F. camerunensis ruandae Van Saceghem, 1942, is placed in the synonymy of F. n. nobihs.
It is also pre-occupied by -F. coqui ruandae Jackson, 1938.
F. swierstrai

F. cvuzi Themido, 1935, is placed in the synonymy of F. swierstrai.

SCALY GROUP

F. squamatus

Although there is considerable individual variation no significant geographical variation is
recognised from the greater part of the range. Accordingly tetraoninus, zappeyi and dowashanus
are placed in the synonymy of schuetti and whitei Schouteden 1954 (nom. nov. for confusa
Schouteden 1954) * n tne synomy of squamatus.

It is not thought practical to recognise by name any micro-populations isolated on the
mountains of the Kenya/Tanganyika border, since the main trend of variation is clinal.
Therefore, kapitensis, keniensis and chyuluensis Van Someren, 1939 are placed in the synonymy
of maranensis.

VERMICULATED GROUP

F. bicalcaratus

It has been shown that birds become darker and more heavily patterned towards the south-
east of the range, but the differences exhibited in this cline do not seem great enough to warrant
separating the birds of the extreme south-east from those of the southern (British) Cameroons.
Accordingly molunduensis Grote, 1949, is synonymised with ogilviegranti.

F. icterorhynchus

The clinal variation does not warrant more than a division into two subspecies. Accordingly
grisescens is placed in the synonymy of icterorhynchus and emini and ugandensis in the synonymy
of dybowskii. Birds showing a few chestnut feathers on the flanks, the character on which
ugandensis was founded, occur sporadically in south-eastern Uganda and must be considered
aberrant specimens perhaps indicating past hybridisation with clappertoni.

F. clappertoni

F. tschadensis is placed in the synonymy of F. c. clappertoni (see text).

The clinal variation shown in southern populations does not warrant extensive splitting;
accordingly cavei Macdonald, 1940, and testis are placed respectively in the synonymy of gedgii
and sharpii.

F. hildebrandti

Numerous subdivisions are not practical; accordingly helleri is placed in the synonymy of
hildebrandti, and grotei and lindi in that of johnstoni.

F. natalensis

The name thamnobium Clancey, 1953, was proposed to distinguish paler birds inhabiting drier
areas from those of wetter habitats in Natal and the Zoutspansberg (natalensis), but since this

THE FRANCOLINS, A STUDY IN SPECIATION 177

gives natalensis a discontinuous range it is not practical to recognise the differences by name.
F. hartlaubi

ovambensis is considered a synonym of bradfieldi.
F. adspersus
kalahari is considered a synonym of adspersus.

STRIATED GROUP

F. sephaena

The name somaliensis Grant & Praed, 1934, was given to streaked birds from Somaliland
and was based on the misapprehension that the type of spilogaster was an unstreaked bird
(Rand, 1950: 384). As noted in the text I believe spilogaster must be used for the hybrid
populations from Somaliland to southern Kenya. However, if it is established that there still
exists an unmixed population of streaked birds on the coastal plains which is distinct from the
southern rovuma then the name somaliensis is available. Such a population can be assumed
to have existed once since it must be regarded as one of the parents of the hybrid spilogaster
but it may have been integrated into the hybrid population.

Local variations in size, colour and pattern throughout the species do not seem great enough
to distinguish by name. Accordingly zuluensis is placed in the synonymy of sephaena :
thompsoni, chobiensis, mababiensis in the synonymy of zambesiae : schoanus, ochrogaster Hartlaub,
1882, delutescens, and jubaeensis in the synonymy of grantii: Peters (1934: 73) shows that
rovuma antedates kirkii.

RED-WINGED GROUP

F. psilolaemus

On the author's own showing fricki is regarded as a synonym of F. p. ellenbecki (Friedmann,
1930: 113).

F. shelleyi

There is not sufficient variation between the populations of South Africa and Tanganyika
to justify the recognition of trothae which is regarded as a synonym of shelleyi, as is sequestris
Clancey, 1960.

F. africanus

Praed & Grant have been among the few modern authors who have correctly regarded
Perdix afra Latham, 1790, as the senior name of the South African Grey wing, known by others
as F. africanus Stephens, 1890. However, since the submersion of Pternistis in Francolinus,
Perdix afra Latham is pre-occupied by Tetrao afer Muller 1776, the senior name for the bare-
throated francolin formerly known as Pternistis afer. The Greywing must therefore now be
known as F. africanus.

Within the range of the species there is slight local variation but it is not great enough for
me to recognise any subspecies: proximus Clancey, 1957, i- s therefore placed in the synonymy.

F. levaillantoides

Roberts (1936: 321) shows that F. levaillantoides is the original name for this species and
gariepensis is a synonym.

In this species the haphazard distribution of rufous and grey micro-populations in the
southern districts of the Sudan and Abyssinia and in Angola and South West Africa makes it
difficult to know how many subspecies can usefully be recognised. I believe it is impractical
to attempt to differentiate between these variations if no other differences are shown, and,
furthermore, names should not be used for intermediate populations. Accordingly I recognise
in the south

(a) F. I. levaillantoides. All districts east and south of the Kalahari, intergrading with
kalaharica on the fringes of the Kalahari. Richly coloured (even in the greyest specimens)
and heavily patterned. Synonyms, ludwigi and gariepensis.

(b) F. I, kalaharica (Roberts) 1932. The Kalahari, Pale and grey, birds from central

178 B. P. HALL

districts almost unmarked on the abdomens; birds from the fringes rather darker with
more markings on the abdomen. Synonym, langi Roberts, 1932. (Although langi has
line priority over kalaharica I use kalaharica, as I am entitled to do as first revisor, since
the name was given to a very pale, grey bird from the Damara Pan near the centre of the
Kalahari, and thus represents the extreme of variation, while langi was given to a rather
darker bird from Nkate near the eastern border.)

(c) F. I. pallidior. South West Africa. Either rufous or grey but the rufous birds paler in
tone than levaillantoides, and the grey birds darker and more heavily marked than
kalaharica. As known at present the population of the Windhuk area is more rufous,
and that of the Tsumeb area greyer (approaching kalaharica). Synonym wattii
Macdonald, 1953. (See Macdonald, 1953, for detailed discussion.)

(d) F. I. jugularis. Southern Angola to the Cunene. Differs from all other subspecies in
having a broad black-and-white gorget: typical birds from the Benguela area are small
and grey but those from the Cunene are larger and more rufous showing an approach to
some of the populations of pallidior. Synonyms, cunenensis Roberts, 1932, stresemanni
Hoesch & Niethammer, 1940.

In the north

(e) F. 1. archeri. Extreme southern Sudan to the Golis range of western (British) Somaliland.
Similar to F. I. pallidior in having both rufous and grey populations but in both the
abdomens are paler and commonly marked with fine lines. Synonyms friedmanni
Grant & Praed, 1934, stantoni Cave, 1940.

(/) F. I. lorti. Mountains of eastern (British) Somaliland. Very grey, with a very pale
abdomen fairly heavily marked with fine lines.

(g) F. I. gutturalis. Northern Abyssinia and Eritrea. Quite distinct (see text). Synonym
eritrea Zedlitz.

F. levaillantii

A number of names have been given to the birds of the northern and western populations
which have no black-and-white stripe down the ochre collar on the hind neck. However,
while the few specimens available indicate that there is a tendency for Angola birds to be
paler, the difference is trivial and the series otherwise is remarkably uniform. Accordingly
mulemae, adolffriederici, benguellensis Neumann, 1908, clayi White, 1944, and momboloensis
White, 1952 (nom. nov. for benguellensis) are placed in the synonymy of kikuyuensis .

THE RED-TAILED GROUP

F. coqui

In the populations of central and south-eastern African there is considerable individual and
local variation in size and colour so that consistent geographical variation is hard to determine.
I do not believe it is practical to distinguish from nominate coqui the small coastal birds of
Kenya since equally small birds are found in Natal, or to distinguish the populations of Natal,
Zululand and southern Mozambique, which on the whole are less rufous, for individuals can be
matched with others elsewhere. The names stuhlmanni and campbelli are therefore placed in
the synonymy of coqui. I regard lynesi Sclater, 1932, also as a synonym of coqui since the type
is intermediate between coqui and kasaicus but comes from an area in which the majority of
specimens are closest to coqui.

For the Uganda race the original spelling " ruahdae " has been used since, while the author
may have intended the name to be " ruandae ", there is no clear evidence of this intention in
the original publication. (Int. Code of Zool. Nomen. 1961, Article 32a (ii) .) F. coqui ruandae
Jackson, 1938, is accordingly a junior objective synonym. (Article 33a (ii) .)

F. albogularis

Serle (1957: 388) shows that gambagae must be considered a synonym of buckleyi.

The isolated populations of Marungu and Upemba in the south-eastern Congo, the Balovale
district of Northern Rhodesia, and of eastern Angola seem to show an increase from east to
west in the amount of patterning in the females, and the Angola birds seem also to be darker,

THE FRANCOLINS, STUDY IN SPECIATION 179

less rufous, than others (Hall, ig6oa: 409). Too few specimens are available at present on
which to judge the extent of variation, and until there are more, especially of topotypical
meinertzhageni from Rhodesia, it seems best to use meinertzhageni for Rhodesian and Angola
birds and dewittei for the Congo birds, though I believe that the Rhodesian birds may prove to
be closer to those from the Congo than to those from Angola.

F. schlegelii

Traylor (1960: 86) has shown that confusus Neumann, 1933, should be considered a synonym
of schlegelii.

SPECIES NOT INCLUDED IN ANY GROUP

F. pondicerianus

It has been shown that variation in this species is slight and largely clinal and ecological.
It is not practical to recognise more than one intermediate subspecies. Accordingly paganus
Koelz, 1954, ti* ar Koelz, 1954, prepositus Koelz, 1954, are au placed in the synonymy of
interpositus.

SPOTTED GROUP MAP 1

The shaded areas represent the overall ranges of the species. The thick red line indicates
divisions between species. The thin red line indicates divisions between subspecies, the
zig-zag line denoting intergradation. The red crosses indicate localities from which
hybrid specimens have been obtained.

F. francolinus

F. pictus

1. francolinus

2. arabistanicus

3. henrici

4. asiae

5. melanotus

6. pallidus

7. pictus

8. watsoni

F. pintadeanus 9. phayrei

10. pintadeanus

MAP. IA. Hypothetical distribution of the ancestral stock of the three species during a
glaciation (Stage 2).

BARE- THROATED GROUP MAP 2

Overall ranges are shown rather than precise collecting localities. Thick red lines
denote specific boundaries. Thin red lines denote subspecific boundaries, plain lines
being boundaries between non-intergrading subspecies, zig-zag lines boundaries between
intergrading subspecies, toothed lines boundaries of hybrid zones. The red " X " marks
the area of hybrids swainsonii x afer. Numbers have been given only to species and
subspecies which are important to the discussion on speciation, but the ranges of other
subspecies are indicated.

i. F. leucoscepus (leucoscepus, infuscatus)

iiillii-iiililiiliil 2 - F. rufopictus

V.V.'.'.'.'.J 3. F. swainsonii (swainsonii, lundazi, damarensis, gilli)

F. afer. Black-and-white, " afer-type " subspecies

4. castaneiv enter & notatus

5. lehmanni

6. swynnertoni

6x7. hybrids (humboldtii)

7. melanogaster & loangwae

8. leucoparaeus

9. afer

Vermiculated, " cranchii-type " subspecies

10. cranchii & intercedens

11. harterti

Hybrids between " afer-type " and " cranchii-type " subspecies.

MAP 2A. Hypothetical distribution of the ancestral stock of the four species during a
spread of forest (Stage 4).

MAP 2B. Hypothetical distribution of the ancestral stock of the four species and two main
blocks of afer subspecies during a later dry period (Stage 6) .

MONTANE GROUP MAP 3

The letters indicate precise localities from which specimens have been collected.
Plain red lines denote divisions between isolates which are regarded as species, and
broken lines between those regarded as subspecies. The thickness of the lines is relative
to the degree of divergence between neighbouring forms.

e F. erckelii i. pentoni

2. erckelii

o F. ochropectus 3. ochropectus

c F. castaneicollis 4. ogoensis

5 . castaneicollis

6. kaffanus

7. atrifrons

j F. jacksoni 8. jacksoni

g. pollenorum

n F. nobilis 10. nobilis

1 1 . chapini

x F. camerunensis 12. camerunensis

s F. swierstrai 13. swierstrai

MAP 3A. Hypothetical distribution of montane forest (shaded black) in a cold, wet era
(Stage i).

ZOOL. 10, 2.

SCALY GROUP MAP 4

A continuous distribution has been shown for members of the Scaly Group through the
Upper and Lower Guinea forests, along the escarpment of Angola, and in the wetter areas
of Uganda and Kenya, for within these areas the birds may be expected in suitable
clearings or cultivations. Outside this range only general localities from which specimens
have been collected have been shaded. Broad red lines indicate divisions between species :
thinner red lines denote divisions between subspecies, those that intergrade being indicated
by a zig-zag line. Thicknesses of the subspecific lines are relative to the divergence shown
by neighbouring forms.

F. ahantensis

1. hopkinsoni

2. ahantensis

F. squamatus

- griseostriatus

3. squamatus

4. schuetti

5. maranensis

6. usambarae

7. uzungwensis

8. doni

MAP 4A. Hypothetical map showing extent of lowland forest in a warm, wet era (Stage 5).
Limits of forest shown by broken line.

ZOOL. IO, 2

VERMICULATED GROUP MAP 5

The distribution of many members of this group is known to be discontinuous.
Accordingly only general areas from which specimens have been collected are shaded,
but the birds may be expected to occur in some of the intervening areas in which little
collecting has been done. Thick red lines denote divisions between species: where there
is evidence of hybridisation these lines are toothed. Thin red lines denote divisions
between subspecies, those that intergrade being shown with a zig-zag line. The red
" X " denotes a hybrid between hildebrandti and natalensis. The red " T " denotes the
aberrant or hybrid specimen F. tschadensis. The red " U " denotes the area in which
the aberrant " ugandensis "-type birds are found.

F. bicalcaratus i. ayesha

2. bicalcaratus

3. adamauae

4. ogilviegranti

5. thornei

F. clappertoni 6. clappertoni

8. heuglini

9. sharpii

10. konigseggi

11. nigrosquamatus

F. icterorhynchus 12. dybowskii

13. icterorhynchus

F. hildebrandti
F. natalensis

F. hartlaubi
F. capensis
F. adspersus
F. harwoodi

14. altumi

15. hildebrandti

1 6. johnstoni

17. neavei

1 8. natalensis

(hartlaubi, bradfieldi, crypticus}

tfll

STRIATED GROUP MAP 6

Only general areas from which birds have been collected are shaded. The broken thick
red line indicates the outer limits of the range of F. streptophorus , and the unbroken thick
red line the outer limits of the range of F. sephaena. Thin red lines indicate subspecific
divisions, plain lines denoting divisions without intergrading, zig-zag lines denoting
intergrading, and toothed lines denoting the boundary of a hybrid area.

F. streptophorus

F. sephaena i. sephaena

2. zambesiae

3. grantii
3x4. spilogaster

4. rovuma

MAP 6A. Hypothetical distribution of the ancestral stock of the two species at the end
of a dry era (Stage 4). Arrows indicate the line of advance of encroaching woodlands.

,vif

RED -WINGED GROUP MAP 7

The distribution of all members of this group is discontinuous, accordingly only general
areas from which birds have been collected are shaded. Thick red lines denote the
extreme limits of the range of each species : these lines are broken where they link remotely
isolated populations, except in F. levaillantoides where arrows indicate where the lines
linking the northern and southern populations might fall. Lineal divisions between
subspecies are not shown but different shading has been used for those subspecies which
are strongly differentiated: others are numbered. The red cross indicates the Amani
specimen which is intermediate between F. s. shelleyi and F, s. uluensis.

F. psilolaemus (psilolaemus, ellenbecki, elgonensis, theresae]

F. shelleyi uluensis (including macarthuri)
F. shelleyi shelleyi
F. shelleyi whytei

F. africanus

F. levaillantoides i. levaillantoides

2. kalaharica

3. pallidior

4. jugularis

5. archeri

6. lorti
F. levaillantoides gutturalis

F. levaillantii

F. finschi

7. levaillantii

8. crawshayi
g. kikuyuensis

MAP yA. Mean Annual Rainfall (taken from Fitzgerald 1950: 35)
under 4 inches

MAP ye. Diagrammatic map showing hypothetical colonisation routes and centres of
speciation (Stages 3-6).

RED-TAILED GROUP MAP 8

The distribution of all members of this group is discontinuous and only general areas
from which specimens have been collected are shaded. Thick red lines denote the extreme
limits of the ranges of species, those linking remotely isolated populations being broken,
except between the north-western and southern populations of F. albogularis where the
link is indicated by arrows. Subspecific divisions are shown lineally only in the critical
area of Kenya and Tanganyika. Here a toothed line surrounds a population which
appears to be a result of hybridisation but which does not now intergrade with neigh-
bouring subspecies: zig-zag lines indicate intergrading. Outside this critical area all
subspecies, if not isolates, intergrade.

F. coqui

W///A

F. schlegelii

F. albogularis

1. coqui

2. vernayi

3. hoeschianus

4. angolensis

5. kasaicus

6. ruahdae

7. hubbardi

8. thikae

9. maharao
10. spinetorum

11. albogularis

12. buckleyi

13. dewittei

14. meinertzhageni

MAP SA. Hypothetical distribution of the ancestral stock of F. coqui and proto-albogularis/
schlegelii (shown by mixed diagonal and horizontal shading) before the two latter species
diverged (Stage 3). Distribution of lowland forest shown in black.

MAP SB. Hypothetical distribution of the ancestral stock of the three species in a sub-
sequent dry era (Stage 4).

SPECIES NOT INCLUDED IN ANY GROUP
MAP 9

F. nahani
F. lathami

1. lathami

2. schubotzi

F. pondicerianus

F. gularis

MAP 10

1. mecranensis

2. interpositus

3. pondicerianus

4. ceylonensis

VEGETATION MAP OF AFRICA MAP 11

Adapted from " Vegetation Map of Africa south of the Tropic of Cancer", Oxford
University Press, 1959.

The numbering and terminology of the original map have been used as far as possible,
excluding Madagascar. Explanatory notes indicating geographical areas are given in
the key in cases where it is not possible to differentiate on the map between vegetation
types.

KEY

1-4. EVERGREEN FORESTS & MONTANE COMMUNITIES.

1. Montane Evergreen Forest.

2. Temperate and Subtropical Evergreen Forest. Knysna and other
districts of the extreme south only,

3. Montane Communities undifferentiated. These include evergreen
forest (type i), grassland (type 5) and woodland.

4. Montane Communities with alfro-alpine communities. High moun-
tains of central and east Africa and Ethiopia only.

(It has not been possible on this scale to distinguish between types i ,
3, and 4.)

5-6. MONTANE & HIGH ALTITUDE GRASSLAND.

5. Montane Grassland above 6,000 ft. North-eastern and central Africa.

6. Temperate and Subtropical Grassland. Pure grassland above 3,5oo//.
in South Africa.

MOIST FOREST AT LOW AND MEDIUM ALTITUDES.

FOREST-SAVANNA MOSAIC.

8. Moist forest surrounded by savanna of tall grasses. Found in belts

surrounding type 7.
9. Coastal Forest-Savanna Mosaic. East coast only.

10. DRY & DECIDUOUS FOREST with abundant Baikiaea plurijuga. Found

only in parts of N. & S. Rhodesia, Bechuanaland and the Caprivi Strip,
and sometimes regarded as woodland rather than forest.

12 & 14. THICKETS.

12. Itigi type. Central Tanganyika only.

14. Ethopian evergreen type. Ethiopia only.

15. CAPE MACCHIA.

16. RELATIVELY MOIST WOODLANDS. Types in which the genera

Isoberlinia, Brachystegia, and Julbernadia are absent or rare.

17-19. WOODLANDS (interspersed with Savanna).

17. With abundant Isoberlinia doka and /. dalzielii. Northern areas.

18. With abundant Brachystegia and Julbernadia " myombo ". Tan-
ganyika to central Northern Rhodesia, Southern Rhodesia and Mozambique.

19. Similar to type 18 but with a number of distinctive species. Prin-
cipally on Kalahari sand. Angola to western N. Rhodesia.

20-22. RELATIVELY DRY WOODLANDS & SAVANNAS.

20. With frequent Acacia but other species as well, and savannas of tall
grass with certain species of Acacia.

21. Ethiopian types. A heterogeneous assemblage of low deciduous
shrubs on stony ground. Ethiopia only.

22. With abundant Colospermum mopane. Southern Africa only.

24. GRASS STEPPE ON KALAHARI SAND.

25. STEPPES.

25. Wooded steppe with abundant Acacia and Commiphora.

26. Grass steppe with thicket clumps. Western Uganda only.

27. Grass steppe Luanda type. Coastal Angola only.
28-31. KAROO & SUBDESERT STEPPE.

28. Karoo Succulent Steppe. Western Cape Province only.

29. Karoo shrub and grass. Cape Province and southern South West Africa.

30. Transitional and mixed Karoo. East-central Cape Province.

31. Northern areas and northern South West Africa and southern Angola.
32. DESERT.

33- SWAMPS.

200 B. P. HALL

REFERENCES

BANNERMAN, D. A. 1930. The birds of Tropical West Africa, i. London.
BENSON, C. W. 1945. Notes on the birds of southern Abyssinia (part). Ibis 87: 366-400.

- 1948. Geographical voice-variation in African birds. Ibis 90: 48-71.

1951. Breeding and other notes from Nyasaland and the Lundazi district of Northern

Rhodesia. Bull. Mus. Comp. Zool. Harv. 106: 69114.

1958. Some additions and corrections to a " Check List of the Birds of Northern

Rhodesia ". Oc. Pap. Nat. mus. S. Rhod. 22b: 190-197.

BENSON, C. W. & F. M. 1948. Notes from southern Nyasaland mainly from the Lower

Shire Valley at 200 ft. altitude. Ostrich 29: 1-16.
BENSON, C. W., IRWIN, M. P. S., WHITE, C. M. N. in press. The significance of valleys as avian

zoogeographical barriers. Symposium on causes and problems of animal distribution with

special reference to southern Africa. July, 1961. Zool. Soc. of southern Africa.
BENSON, C. W. & WHITE, C. M. N. 1957. Check List of the Birds of Northern Rhodesia.

Lusaka.

BENSON, C. W. & WHITE, C. M. N. in press. Discontinuous Distributions (Aves).
BOWEN, W. W. 1930. The relationships and distribution of the bare-throated francolins

(Pternistis). Proc. Acad. Nat. Sci. Phil. 82: 149-164.
CHAPIN, J. P. 1932. The birds of the Belgian Congo. I. New York.
CLANCEY, P. A. 1953. Miscellaneous taxonomic notes on African birds. Durban Mus.

Nov. 4(4): 57-64.
COLLIER, F. S. 1935. The Ahanta Francolin in Ibadan Province, Nigeria. Ibis (13) 5:

665-666.
DELACOUR, J. 1951. Commentaires, modifications et additions a la liste des oiseaux de

1'Indochine Francaise (2). L'Oiseau (5) 21: 1-32.
DORST, J. & JOUANIN, C. 1954. Precisions sur la position systematique et 1'habitat de

Francolinus ochropectus. L'Oiseau (5) 24: 161-170.
FITZGERALD, W. 1950. Africa (seventh edition). London.
FRIEDMANN, H. 1930. Birds collected by the Childs Frick expedition to Ethiopia and

Kenya Colony. Pt. i. Non-passeres. U.S. Nat. Mus. Bull. 153.
GOOD, A. I. 1952. The birds of French Cameroon (Part i). Mem. Inst. Fran. Afr. Noire

Nat. 2: 1-203.
GRANT, C. H. B. & PRAED, C. W. M. 1946. Notes on Eastern African Birds. Bull. Brit.

Orn. Cl. 66: 74-81.
HALL, B. P. 1956. Notes on a small collection of birds from Panda Matenga, N.E. Bechuana-

land. Ostrich 27: 96-109.
HALL, B. P., ig6oa. The ecology and taxonomy of some Angola birds. Bull. Brit. Mus. (Nat.

Hist.) Zool. 6 (7): 370-453.

I96ob. The faunistic importance of the scarp of Angola. Ibis. 102: 420-442.

HARRISON, C. J. O. In press. The incidence and origin of spotted patterns in the Estrildidae.

Ibis. 105
HOPKINSON, E. 1923. The game-birds and pigeons of the Gambia. Avic. Mag. (i) 4: 125-31,

166-72, 183-7, 207-15, 23 1 -?-

HORSBURGH, B. 1912. The game-birds and waterfowl of South Africa. London.
JACKSON, F. 1938. The birds of Kenya Colony and the Uganda Protectorate. Vol. i.
LAVAUDEN, L. 1936. Les Francolins. Alauda 8: 301-315.
LILFORD, LORD. 1862. On the extinction in Europe of the Common Francolin (Francolinus

vulgaris). Ibis. (i). 4: 352-356.
LONG, R. C. 1960. The birds of the Port Herald District (i). Ostrich. 31: 85-104.

- 1961. Interesting specimens of birds in the Nyasaland Museum. Nyasa Mus. Ann. Rep.
and Bull. 1960-61 : 25-27.

MACDONALD, J. D. 1953. The races in South West Africa of the Orange River Francolin.
Bull. Brit. Orn. Cl. 73: 34-36.

1957- Contribution to the ornithology of western South Africa. London.

THE FRANCOLINS, A STUDY IN SPECIATION 201

MALBRANDT, R. & MACLATCHY, A. R. 1949. Faune de 1'Equateur Africain Fran9ais. (I).

Oiseaux. Enc. biol. Paris 35.

MARKS, M. B. 1959. Red-necked Francolin Pternistis afer in Bechuanaland. Ostrich 30: 44.
MCLACHLAN, C. R. & LIVERSIDGE, R. 1957- Roberts' birds of South Africa. Cape Town.
MOREAU, R. E. 1930. On the age of some races of birds. Ibis, (12), 6: 229-239.

in press. Vicissitudes of the African biota in the Late Pleistocene. Proc. Zool. Soc.

London.

PETERS, J. L. 1934. Check list of birds of the world. 2. Cambridge, U.S.A.

PRAED, C. W. M. & GRANT, C. H. B. 1952. Birds of eastern and north eastern Africa.

London.
RAND, A. L. 1950. On the name Francolinus sephaena spilogaster Salvadori. Auk 67

384-385-

- 1951. Birds from Liberia. Fieldiana: Zoology 32 (9): 561-653.
ROBERTS, A. 1936. Some unpublished field-notes made by Dr. (Sir) Andrew Smith. Ann.

Trans. Mus. 18: 271-323.
SERLE, W. 1957. A contribution to the ornithology of the eastern region of Nigeria. Ibis 99 :

371-418, 628-685.

SMITH, K. D. 1957. An annotated check list of the birds of Eritrea. Ibis 99: 1-337.
SMITHERS, R. H. N., IRWIN, M. P. S. & PATTERSON, M. L. 1957. A check list of the birds of

Southern Rhodesia. Cambridge.
TRAYLOR, M. A. ig6oa. Notes on the birds of Angola, non-passeres. Publ. cult. Co. Diam.

Ang. Lisboa 51: 129-186.

I96ob. Francolinus schlegelii Heuglin in Cameroon. Bull. Brit. Orn. Cl. 80: 86-88.

VAN SOMEREN, V. G. L. 1925. The birds of Kenya and Uganda. Part 2. Phasianidae.

Journ. E. Afr. & Ug. Nat. Hist. Soc. 23: 95-104.

VAN TYNE, J. & BERGER, A. J. 1959. Fundamentals of Ornithology. New York.
VERHEYEN, R. 1953. Exploration du Pare National de 1'Upemba, Mission G. F. de Witte.

Fasc. 19. Oiseaux.
VINCENT, J. 1934. The birds of northern Portuguese East Africa. Part III. Ibis, (13.)

4: 305-340.

WHITE, C. M. N. 1952. On the genus Pternistis. Ibis 94: 306-309.
WINTERBOTTOM, J. M. & CouRTENAY-LATiMER, M. 1 961. A list of the birds of Little

NAMAQUALAND. S. Afr. Av. Ser. Percy FitzPatrick Inst. Afr. Orn. 1: 1-46.

INDEX TO SPECIFIC AND SUBSPECIFIC NAMES OF FRANCOLINUS

Specific and subspecific names of francolins are listed under the nomenclature used in this
paper, those of species and subspecies here recognized appearing in roman type, others in italics.

No attention is paid to other generic names that have been given to francolins except in a
few cases where the merging of these names in Francolinus has created homonyms.

For any species that is here regarded as a member of a species group all the pages of the
section dealing with the whole group are listed, regardless of whether the species is actually
mentioned on each page.

adamauae, F. bicalcaratus, 144, 188 afra, Perdix (F. afer), 152, 155, 173, 177

adolffriederici, F. levaillantii, 178 africanus, F., no, in, 117, 120, 152-160,

adspersus, F., no, 113, 117, 120, 121, 140- 169, 173, 177, 192

148, 169, 176, 188 ahantensis, F., no, in, 116, 136-140, 167,

afer, F., no, in, 113, 116, 119, 121, 122, 168, 169, 172, 186

125-131, 132, 140, 145, 146, 168, 169, 175, ahantensis, F. a., [136], 186

177, 182 albogularis, F., no, 114, 117, 122, 160-165,

afer, F. a., 127, 175, 182 169, 170, 178, 194

202

B. P. HALL

albogularis, F. a., 162, 194
altumi, F. hildebrandti, 142, 146, 188
angolensis, Pternistis afer, 175
angolensis, F. coqui, 163, 175, 194
arabistanicus, F. francolinus, 124, 180
archeri, F. levaillantoid.es, 155, 156, 178, 192
asiae, F. francolinus, 124, 175, 180
atrifrons, F. castaneicollis, 112, 133, 135, 136,

184

ayesha, F. bicalcaratus, 141, 144, 147, 188
aylwinae, F. afer, 175
benguellensis , F. after, 175
benguellensis, F. levaillantii, 178

bicalcaratus, F., no, 117, 118, 120, 121, 123,
140-148, 161, 164, 165, 169, 171, 172, 176,
1 88

bicalcaratus, F. b., 144, 188

billypaynei, F. francolinus, 175

bogdanovi, F. francolinus, 175

bohmi, F. afer, 175

boineti, F., 175

bottegi, F. castaneicollis, 176

bradfieldi, F. hartlaubi, 145, 177, 188

buckleyi, F. albogularis, 162, 178, 194

camabatelae, F. afer, 175

camerunensis, F., no, 116, 118, 120, 131-136,

169, 184

campbelli, F. coqui, 178
capensis, F., no, 117, 140-148, 169, 188
castaneicollis, F., no, in, 116, 131-136, 169,

176, 184

castaneicollis, F. c., 133, 135, 176, 184
castaneiventer, F. afer, 128, 130, 131, 175, 182
caucasicus, F. francolinus, 1 75
cavei, F. clappertoni, 176
ceylonensis, F. pondicerianus, 167, 196
chapini, F. nobilis, 134, 184
chio, F. afer, 175
chobiensis, F. sephaena, 177
chobiensis, Pternistis swainsonii, 175
chyuluensis, F. squamatus, 176
clappertoni, F., no, 114, 117, 118, 119, 120,

121, 140-148, 161, 164, 165, 169, 170, 176,

188

clappertoni, F. c., 141, 144, 176, 188
clayi, F. levaillantii, 178
confusa, F. squamatus, 176
confusus, F. schlegelii, 179
cooperi, Pternistis, 128
coqui, F., no, in, 117, 118, 119, 121, 122,

145, 147, 160-165, 167, 169, 170, 171, 178,

194

coqui, F. c., 161, 162, 163, 164, 178, 194
cranchii, F. afer, 127, 168, 175, 182
crawshayi, F. levaillantii, 157, 192
cruzi, F., 176

crypticus, F. hartlaubi, 145, 188
cunenensis, F. levaillantoides, 178
cunen sis, Pternistis afer, 175

damarensis, F. swainsonii, 129, 182
delutescens, F. sephaena, 177
dewittei, F. albogularis, 162, 179, 194
doni, F. squamatus, 137, 139, 186
dowashanus, F. squamatus, 176
dybowskii, F. icterorhynchus, 144, 176, 188

elgonensis, F. psilolaemus, 153, 154, 192
ellenbecki, F. psilolaemus, 153, 177, 192
emini, F. icterorhynchus, 176
erckelii, F., no, in, 116, 120, 131-136, 141,

143, 145, 146, 169, 184
erckelii, F. e., 135, 184
eritrea, F. levaillantoides, 178

festinus, F . francolinus , 175

finschi, F., no, 117, 119, 122, 152-160, 169,

170, 192
francolinus, F., no, in, 116, 118, 119, 122,

123-125, 165, 166, 167, 168, 169, 175, 180
francolinus, F. 1, 124, 175, 180
fricki, F. psilolaemus, 177
friedmanni, F. levaillantoides, 155, 178

gambagae, F. albogularis, 178
gariepensis, F., 155, 173, 177
gedgii, F. clappertoni, 144, 176, 188
gilli, F. swainsonii, 129, 182
gofanus, F. castaneicollis, 176
grantii, F. sephaena, 150, 151, 152, 177, 190
grisescens, F. icterorhynchus, 176
griseostriatus, F., no, 116, 136-140, 167, 169
grotei, F. hildebrandti, 176
gularis, F., 109, no, 117, 167-168, 169, 196
gurae, F.jacksoni, 176

gutturalis, F. levaillantoides, 155, 156, 159,
160, 178, 192

harterti, F. afer, 127, 182

hartlaubi, F., no, 117, 120, 121, 140-148,

169, 177, 188
hartlaubi, F. h., 144, 188
harwoodi, F., no, 117, 118, 120, 140-148,

169, 1 88

helleri, F. hildebrandti, 176
henrici, F. francolinus, 124, 175, 180

THE FRANCOLINS, A STUDY IN SPECIATION

heuglini, F. clappertoni, 144, 188
hildebrandti, F., no, in, 117, 118, 140-148,

1 68, 169, 176, 1 88
hildebrandti, F. h., 142, 176, 188
hoeschianus, F. coqui, 163, 194
holtemulleri, F. leucoscepus, 175
hopkinsoni, F. ahantensis, 136, 186
hubbardi, F. coqui, 161, 162, 164, 194
humboldtii, F. afer, 127, 175, 182

icterorhynchus, F., no, 114, 117, 120, 122,
140-148, 163, 164, 165, 169, 171, 176, 188
icterorhynchus, F. i., 144, 176, 188
infuscatus, F. leucoscepus, 129, 175, 182
intercedens, F. afer, 127, 175, 182
intermedius, F., 124

interpositus, F. pondicerianus, 167, 179, 196
itigi, F. afer, 175

jacksoni, F., no, 116, 118, 120, 131-136, 169,

176, 184

jacksoni, F. j., 133, 176, 184
johnstoni, F. hildebrandti, 142, 176, 188
jubaeensis, F. sephaena, 177
jugularis, F. levaillantoides, 156, 178, 192

kaffanus, F. castaneicollis, 133, 134, 135, 176,

184

kalahari, F. adspersus, 176
kalaharica, F. levaillantoides, 156, 177, 178,

192

kapitensis, F. squamatus, 176
kasaicus, F. coqui, 163, 178, 194
keniensis, F. leucoscepus, 175
keniensis, F. squamatus, 175, 176
kikuyuensis, F. levaillantii, 157, 178, 192
kilimensis, F. leucoscepus, 175
kirkii, F. sephaena, 177
konigseggi, F. clappertoni, 144, 188
krebsi, F. afer, 175

langi, F. levaillantoides, 178

lathami, F., 109, no, 117, 119, 123, 165-166,

169, 196

lathami, F. 1., [166], 196
lehmanni, F. afer, 128, 131, 182
leucoparaeus, F. afer, 127, 131, 182
leucoscepus, F., no, 113, 116, 119, 122, 125-

131, 169, 175, 182
leucoscepus, F. 1., 129, 182
levaillantii, F., no, 117, 121, 122, 152-160,

168, 169, 178, 192
levaillantii, F. 1., 157, 192

203

levaillantoides, F., no, 113, 117, 121, 122,

152-160, 168, 169, 173, 177, 192
levaillantoides, F. 1., 156, 177, 178, 192
lindi, F. hildebrandti, 176
loangwae, F. afer, 127, 129, 182
lorti, F. levaillantoides, 155, 156, 178, 192
ludwigi, F. levaillantoides, 177
lundazi, F. swainsonii, 129, 175, 182
lynesi, F. coqui, 178

mababiensis , F. sephaena, 177
macarthuri, F. shelleyi, 154, 192
mackenziei, F. afer, 175
maharao, F. coqui, 161, 164, 165, 167, 194
manueli, F. afer, 715

maranensis, F. squamatus, 137, 176, 186
mecranensis, F. pondicerianus, 167, 196
meinertzhageni, F. albogularis, 162, 179, 194
melanogaster, F. afer, 127, 129, 131, 175, 182
melanotus, F. francolinus, 124, 180
molunduensis, F. bicalcaratus, 176
momboloensis, F. levaillantii, 178
muhamedbenabdullah, F. leucoscepus, 175
mulemae, F. levaillantii, 178

nahani, F., 109, no, 117, 120, 123, 166-167,

169, 196
natalensis, F., no, 111, 117, 121, 128, 131,

140-148, 168, 169, 176, 188
natalensis, F. n., 117, 142, 176, 177, 188
neavei, F. natalensis, 117, 142, 143, 146, 188
nigrosquamatus, F. clappertoni, 144, 188
nobilis, F., no, 116, 120, 131-136, 169, 176,

184

nobilis, F. n., [134], 176, 184
notatus, F. afer, 128, 131, 175, 182
nyanzae, F. afer, 175

ochrogaster, F. sephaena, 177

ochropectus, F., 107, no, in, 112, 116, 131-

136, 169, 184

ogilviegranti, F. bicalcaratus, 144, 176, 188
ogoensis, F. castaneicollis, 133, 184
oldowai, F. leucoscepus, 175
ovambensis, F. hartlaubi, 177

paganus, F. pondicerianus, 179
pallidior, F. levaillantoides, 156, 178, 192
palliditectus, F. afer, 175
pallidus, F. pictus, 124, 180
parkerae, F. francolinus, 175
patrizii, F. castaneicollis, 176
pentoni, F. erckelii, 132, 135, 145, 184

204

B. P. HALL

phayrei, F. pintadeanus, 124, 180

pictus, F., no, in, 116, 119, 120, 123-125,

167, 169, 1 80
pictus, F. p., 124, 1 80
pintadeanus, F., no, 116, 118, 123-125, 169,

175, 180

pintadeanus, F. p., [124], 175, 180
pollenorum, F. jacksoni, 133, 184
pondicerianus, F., 109, no, 117, 122, 167,

169, 179, 196

pondicerianus, F. p., 167, 196
prepositus, F. pondicerianus, 179
proximus, F. africanus, 177
psilolaemus, F., no, 117, 121, 122, 152-160,

169, 177, 192

psilolaemus, F. p., 153, 192
punctulata, F. afer, 175

rovuma, F. sephaena, no, 150, 177, 190

ruahdae, F. coqui, 162, 164, 178, 194

ruandae, F. coqui, 176, 178

ruandae, F. nobilis, 176

rufopictus, F., no, 113, 116, 125-131, 169,

170, 182

sarudnyi, F. francolinus, 175

schlegelii, F., no, 114, 117, 118, 160-165,

1 66, 169, 170, 179, 194
schoanus, F. sephaena, 177
schubotzi, F. lathami, 166, 196
schuetti, F. squamatus, 137, 138, 139, 176,

186
sephaena, F., no, in, 113, 117, 118, 121,

123, 148-152, 167, 169, 177, 190
sephaena, F. s., 150, 152, 177, 190
sequestris, F. shelleyi, 177
sharpii, F. clappertoni, 144, 147, 176, 188
shelleyi, F., no, in, 113, 115, 117, 118, 121,

122, 152-160, 168, 169, 173, 177, 192
shelleyi, F. s., in, 113, 122, 153, 154, 155,

159, 160, 168, 177, 192
somaliensis, F. sephaena, 177
spilogaster, F. sephaena, 150, 177, 190
spinetorum, F. coqui, 114, 161, 162, 165, 194
squamatus, F., no, in, 116, 118, 121, 136-

140, 1 66, 1 68, 169, 176, 1 86

squamatus, F. s., 137, 138, 139, 186
stantoni, F. levaillantoides, 155, 178
streptophorus, F., 110, 114, 117, 120, 148-

152, 167, 169, 170, 171, 172, 190
stresemanni, F. levaillantoides, 178
stuhlmanni, F. coqui, 178
swainsonii, F., no, 116, 121, 122, 125-131,

169, 175, 182

swainsonii, F. s., 127, 129, 130, 182
swierstrai, F., no, 116, 131-136, 169, 176,

184
swynnertoni, F. afer, 127, 128, 131, 175, 182

testis, F. clappertoni, 176
tetraoninus, F. squamatus, 176
thamnobium, F. natalensis, 176
theresae, F. psilolaemus, 153, 154, 192
thikae, F. coqui, 161, 162, 164, 194
thompsoni, F. sephaena, 177
thornei, F. bicalcaratus, 144, 188
titar, F. pondicerianus, 179
tokora, F. leucoscepus. 175
tornowi, F. afer, 175
trothae, F. shelleyi, 177
tschadensis, F., 141, 176, 188

ugandensis, F., 141, 176, 188

uluensis, F. shelleyi, in, 154, 155, 159, 160,

168, 192

usambarae, F. squamatus, 137, 139, 186
uzungwensis, F. squamatus, 137, 139, 186

vernayi, F. coqui, 163, 194

watsoni, F. pictus, 124, 180
wattii, F. levaillantoides, 178
wellsi, F. pintadeanus, 175
whitei, F. squamatus, 176
whytei, F. shelleyi, in, 113, 122, 154, 155,
157, 159, 160, 192

zambesiae, F. sephaena, 150, 151, 152, 177,
190

zappeyi, F. squamatus, 176
zuluensis, F, sephaena, 177

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

TAXONOMY AND BIOLOGY

OF THE GENUS LEBETJ7S

(TELEOSTEI-GOBIOIDEA)

P. J. MILLER

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 3

LONDON: 1963

TAXONOMY AND BIOLOGY OF THE GENUS
LEBETUS (TELEOSTEI-GOBIOIDEA)

P. J. MILLER

(Department of Zoology, The University, Glasgow, W.2)

Pp. 205-256 ; Plate ; 21 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 3

LONDON : 1963

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. 3, of the Zoological series.

Trustees of the British Museum 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued March 1963 Price Twenty One Shillings

TAXONOMY AND BIOLOGY OF THE GENUS
LEBETUS (TELEOSTEI-GOBIOIDEA)

By P. J. MILLER

ABSTRACT

The marine gobioid genus Lebetus Winther 1877 is redefined and previous division into two
species orca Collett 1874 and scorpioides Collett 1874 shown to be based on sexual dimorphism
in one species, by page priority L. orca. A redescription of the species is provided including
details of sensory papillae and osteology. L. orca is eastern Atlantic boreal in distribution,
occurring as a small predator on coarser grounds down to 375 m. Specialization of the male
genitalia and sexual dimorphism are described and the breeding season provisionally regarded
as from March to August. The skeleton of Lebetus is typically gobiid, and evidence is presented
for a closer relationship to the genus Buenia Iljin as represented by B. Jeffrey sii (Giinther) than
to the Mediterranean Odondebuenia De Buen and Cabotichthys Whitley with which Lebetus had
been previously grouped.

INTRODUCTION

THE genus Lebetus was founded by Winther (1877) to contain two species of gobies
dredged by G. O. Sars on the coast of Norway and described by Collett (1874) as
Gobius orca and G. scorpioides. These species have subsequently been regarded as
among the rarest of European Atlantic fishes and, until recent years, the numbers of
demersal individuals known to science have been relatively few. In the last fifteen
years, protracted dredging surveys off the Isle of Man (Jones, 1951 ; Hartnoll, 1961)
have yielded many examples. Their small size and occurrence on rough grounds makes
collecting difficult ; in Manx waters, most specimens have been obtained by means of
a scallop dredge lined with shrimp netting. Relative infrequency of capture is
probably due to inadequacies of the fishing gear rather than to actual rarity, and
occasional hauls providing up to five individuals suggest the presence of appreciable
numbers in the area. Other long term faunistic investigations within the range of the
genus have produced more or less sizeable collections of both demersal (Grieg, 1913 ;
Le Danois, 1913) and planktonic stages (Petersen, 1919 ; Johansen, 1925 ; Russell,
1926-1940).

My interest in these fishes was aroused firstly by the difficulty experienced in
separating the two species. After endeavouring to use the diagnoses of several authors,
it was found possible to achieve this only on the basis of coloration. Then, while
examining the testes of various gobies, it was noticed that no female L. orca were
present in a collection of twenty two dissected individuals, and only immature males
among twenty eight L. scorpioides. The possibility that the two so-called species
were in fact based on sexual dimorphism in a single species indicated the need for
taxonomic revision. Although such a view was put forward by Smitt (1900), this
was not accompanied by any evidence and appears to have been overlooked by later
authors. Dissection of the extensive material from the Irish Sea presented oppor-
tunities for settling the position of Lebetus among the gobioid families by osteological

ZOOL. 10, 3 i4

208 P. J. MILLER

study, and also for investigating various aspects of diet and reproduction. A pre-
liminary account of some of this work has already been published (Miller, 19616).

MATERIAL AND METHODS

The following specimens have been available for examination :
Isle of Man : 33 $<$, standard lengths 15 -75-30-0 mm., including BMNH 1961 . 10 . 17 .

i. 3 5 an( i one m PEM, and 27 ?<j>, 14-5-29-0 mm., inc. BMNH 1961 . 10 . 17 . 2, 4,

and one in PEM.
Other British localities : 3 $, 12-0-30-0 mm., BMNH 88 . 3 . 22 . 9-10 part and BMNH

1903.4.14.4-7 part, and 2 $$, 10-5 and 17-0 mm., BMNH 1903.4.14.4-7 part.
Scandinavia: 13 <$<$, 13-5-26-0 mm.,ZMO J3999 (type of Gobius orca), J40OO, J4OOI,

J4023 ; ZMB 771, 1887, 1893, 1966, 4174, 4175, 5294 ; ZMC 89, 91 ; and 4 ??,

21-75-29-0 mm., ZMO J402O, J4O2I (types of G. scorpioides) ; ZMB 536, 2009.
Abbreviations for museums are as follows :

BMNH British Museum (Natural History).

PEM Port Erin Marine Biological Station Museum.

ZMB Zoologisk Museum, Universitetet i Bergen.

ZMC Universitetets Zoologisk Museum, Copenhagen.

ZMO Zoologisk Museum, Universitetet i Oslo.

The specimens were viewed by means of low power binocular and dissecting micro-
scopes. With suitable illumination the sensory papillae were readily seen without
staining after careful removal of mucus from the head and body surfaces. The skeleton
was investigated by dissection of alizarin stained whole-mount preparations.

THE GENUS LEBETUS WINTHER 1877

Lebetus Winther, 1877: 49 (orthotype Gobius scorpioides Collett 1874); Smitt, 1900: 554;
Jordan, 1919 : 392 ; 1923 : 225 ; Duncker, 1928 : 140 ; Iljin, 1930 : 55 ; De Buen, I93oa :
123 ; 19306 : 5, 21 ; Koumans, 1931 : 43, 46 ; Whitley, 1931 : 155 ; De Buen, 1931 : 50, 54,
61 ; 1940 : 3 ; 1951 : 56, 57 ; Norman, MS : 413.

Lebistes Jordan, 1920 : 487 ; 1923 : 226 (non Filippi, 1862 : 69).

Butigobius Whitley, 1930 : 123.

Body moderately elongate, subcylindrical, with 25-29 ctenoid scales in lateral
series. Head, nape, and back naked to origin of second dorsal fin, lacking skin folds
or barbels. Postorbital length not more than half length of head. Anterior nostril a
short tube. Opercle without scales. Preopercle and lower jaw unarmed. Branchi-
ostegous membranes attached to sides of isthmus, but not fused across isthmus. Jaws
oblique, subequal, maxillae ensheathed and not elongate. Teeth of both jaws in
several rows, simple, erect. Tongue truncate to very weakly emarginate, free from
floor of mouth anteriorly. Pelvic fins united along entire length, without anterior
membrane (frenum) and with fourth branched ray somewhat produced. Pectoral
girdle lacking flaps on anterior edge ; uppermost rays of pectoral fin not free from
membrane. Dorsal fins separate, first dorsal with 6 rays, not pungent. Caudal fin
rounded, not much longer than head.

Cephalic canals absent. Sensory papillae prominent but reduced in number. Two
pairs of interorbital papillae (/>). Cheek with only suborbital row a, a short longi-

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 209

tudinal row c, and an intermediate papilla ; row d without posterior longitudinal
section. Anterior dorsal (occipital) series (excluding h) in two groups of a few papillae.
Opercular transverse row ot in two sections. Preoperculo-mandibular series in
several parts.

In compiling his work on the genera of fishes, Jordan (1919, 1920, 1923) includes,
together with Lebetus Winther 1877, the genus Lebistes Smitt 1899. Since Jordan
gives " L. scorpioides Smitt " as the orthotype of this genus, the latter, as noted by
Koumans (1931), is clearly an erroneous reference to Lebetus for which Smitt (actually
1900, not 1899) indicated Gobius scorpioides Collett as the type. The generic name
Lebistes is preoccupied by Lebistes Filippi 1862 among the cyprinodont teleosts and,
to replace this name within Jordan's list of gobiid genera, Whitley (1930) proposed
Butigobius without consulting Smitt 's paper. In a later account, Whitley (1931)
unravels this confused story.

THE SPECIES OF THE GENUS

The present section embodies a review of the various characters proposed by
earlier authors as being of value in the diagnosis of L. orca and L. scorpioides.

Coloration. The two species are reputed to differ in both body and fin coloration.
The body of Gobius orca was described by Collett (1874, 18750) as uniformly greyish
except for darker areas between the second dorsal and anal fins, and at the root of
the caudal fin, while that of G. scorpioides was said to possess four or five dark brown
vertical bands across the sides. After examination of a living scorpioides, Winther
(1877) remarked on the colourless caudal peduncle, and both Holt & Byrne (1903)
and Le Danois (1913) emphasized the importance of this pale band, sharply demar-
cated anteriorly, as a means for the identification of L. scorpioides. In the original
description, Collett (1874) mentioned that the second dorsal fin of G. orca had an
indistinct white band in its outer half whereas the same fin in G. scorpioides showed
only ill-defined banding. Alternate dark brown and snow white bands were found
across the second dorsal fin of a later specimen of G. orca (Collett, 1885). Holt &
Byrne (1903) regarded a superior black band to this fin as characteristic of G. orca
and in their text referred to a row of white spots immediately below the dark peri-
phery. These authors described and illustrated the second dorsal fin in G. scorpioides
as banded with red, the markings in some cases being flanked by dark edging.

With the collection of Lebetus amassed for this work, it has been possible to segre-
gate the individuals into two groups on the basis of differences in the coloration of the
second dorsal fin. In the first of these, termed the orca group, the second dorsal fin
has a thick black superior edge, below which occur white areas and wide oblique
bands, yellow or ochre in life but grey in preserved material. Within the scorpioides
group, the black edge to the second dorsal fin is present but much thinner and less
intense than in the orca examples. The rest of the fin bears merely four narrow double
lines of melanophores, enclosing in the living fish thin red or ferruginous striae, and
separated by hyaline fin membrane. Among living scorpioides, madder brown lateral
marks are pronounced on the body which, in orca fishes, is more uniform grey or.
yellowish grey although comparable markings may be quite prominent. A pale band
on the caudal peduncle is more clearly defined in the scorpioides group, especially

2io P. J. MILLER

when preserved, but living orca show this feature to a noticeable degree (cf. PI. I and
Text-fig. 3) and it is shown in Collett's illustrations of Gobius orca (i875a, 1885).

TABLE I. Relation of Coloration Groups to Sex and Maturity Stage

No. of males

Mature and No. of females

Group developing Immature all stages

(a) Dissected

orca 21 i

scorpioides . 6 22

(b) Not dissected
orca . . 16

scorpioides . 5 n

The relation of these two colour patterns to sex and to gonad maturation is indicated
in Table la, based on fifty Manx specimens whose sex and sexual development were
confirmed by dissection. The sex and approximate maturity stage in a further thirty
two British and Scandinavian examples have been determined by inspection of the
genital papilla (Text-fig. 17), and these results are given in Table Ib. It is apparent
that mature or developing male scorpioides do not occur, and that female orca are
unknown at any stage of maturation. This distribution strongly implies that the
orca coloration is the male livery of a species whose females and immature males bear
the markings of scorpioides. Incipient development of the orca pattern in the second
dorsal fin can, in fact, be noted among some of the immature males in the scorpioides
group, which usually exhibit a pronounced black spot on the distal part of the first
two interradial membranes of the second dorsal fin (see Le Danois, 1910, fig. 4).
The above hypothesis is further supported by the range in standard length found in
the three categories shown in Table II. The examples of orca are from 18-0-30-0 mm.

TABLE II. Standard Length Frequency in Coloration Groups, Expressed in 2 mm.

Length Groups

Standard length

(
Coloration group

orca .

cJ scorpioides

9 scorpioides

in length while male scorpioides do not exceed 23-0 mm. Taning (1940) has reported
juvenile Lebetus orca of 5-15 mm., but relied for the identification on meristic
characters whose validity is doubtful. Female scorpioides attain about the same
maximum length as that for orca.

The first dorsal fin in the orca individuals is a uniform yellowish grey edged with
white, but in scorpioides is usually banded with a dark spot on the interradial mem-
brane between the fifth and sixth rays. Confluence of these bands frequently occurs,
and the spot may not be very distinct. This scorpioides group pattern is stated by

12 14

2 3

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS an

Holt & Byrne (1903) to be found only in the breeding male of " Gobius scorpioides ".
Fage (1918) regarded the presence of a black spot on the first dorsal fin of a 12-5 mm.
example from the Dana collections as a sign of precocious development of supposed
male coloration. However, in the material under consideration, the spot is well
represented in adult females and among male Lebetus occurs only in the immature
where it is not uncommonly absent. The single female G. orca mentioned in the
literature was taken by Patience (1906) in the Firth of Clyde. This specimen has
not been traced. The diagnosis of G. orca given by Patience was not derived from his
own examples but seems to be a translation from Collett (1896) and could apply to
individuals of either the orca or scorpioides group.

TABLE III. Meristic Characters Cited for Lebetus orca and L. scorpioides
No. of fin rays

First

Second

Verte-

dorsal

Anal

Pectoral

Caudal

brae

L. orca 1

3/13/3

VII

IO-II

3/12-13/3

VII

I/9-io

1/8

12-13

VI-VII

I/9-io

1/8

x/13/x

VII

IO-II

3/12-13/3

VI-VII

9-1 1

9-10

VI-VII

I/9-io

1 8 (sic)

17-18

12-13

VI-VII

I/9-io

1/8-9

VI-VII

10-12

8-10

18-20

7-8/1 1/6-7

28-29

scorpioides 2 ' 3

20 (18)

6/12/6

1 8-20

6/12/6

c. 18

1/8

1/7

18-20

1/8

1/7

18-20

X/I2/X

VI
VI
VI
VI
VI

V-VI
VI

1/9

1/8-9

8-9

1/7-8

1/8-9

7-8

1/6

1/6-9

6-8

1/6-7

18-20

18-20
18-20

X/I2/X

No. of
Scales
in 1.1.

28
28
24
25
c. 24

25-28

25-26
24-28

28-30

c. 28-30

28-30

27-28

25-26

26-30

28-30

Authority

Collett, 1874
Collett, 18750
Collett, 1885
Lilljeborg, 1884
Smitt, 1892
Collett, 1896
Holt & Byrne,

1903

Duncker, 1928
De Buen, 19300, b
Taning, 1940

Collett, 1874
Collett, 18750
Winther, 1877
Lilljeborg, 1884
Smitt, 1892
Holt & Byrne, 1903
Le Danois, 1910
Le Danois, 1913
Fage, 1918
Petersen, 1919
Duncker, 1928
De Buen, I93oa,
b, 1932

1 Scales in 1.1. c. 26-c. 28 (Grieg, 1913). z Scales in 1.1. 26, T) l V (Collett, 1902). 3 Scales in 1.1.
28 (Grieg, 1913)-

Meristic characters. Radial formulae, vertebral and scale counts provided by
various authors for the two species are shown in Table III. Although Collett origin-
ally ascribed 6 elements to the first dorsal fin of both species, he later changed this to

212 P. J. MILLER

7 for G. orca. Lilljeborg (1884) and Smitt (1892) employed this supposed disparity in
keys to Scandinavian gobies. L. orca is also regarded as having rather more articu-
lated rays in the second dorsal and anal fins than L. scorpioides, and slight differences
are indicated in pectoral and caudal fin ray counts. The number of scales along the
lateral midline is reputed to be greater in L. scorpioides than in L. orca.

For the orca and scorpioides groups of the present material the meristic values in
Table IV have been obtained. Methods of counting are described in the systematic
section below. To minimise the effect of regional variation, results for British and
Scandinavian collections have been kept separate. Mean values are given for the
more numerous British material.

TABLE IV. Meristic Characters of Coloration Groups : Number of Observations
(Highest in Bold Type) Against Value. M Mean Value for British Material

Origin of specimens

Coloration group

orca
scorpioides

orca
scor

orca
scorpioides

British Isles

5 6
26
34

First dorsal fin rays

Scandinavia

6 7

7 I
6

Second dorsal fin articulated rays

6 7 8 9 10 ii M 8 9 10 ii

i 18 7 9-23 531

i 2 4 21 6 8-85 2 4 i

Anal fin articulated rays
5 6 7 8 9 M. 56789

i 19 6 7 -19 54

32 22 7 6-97 i 4i

Pectoral fin rays

15 16 17 18 19 20 21 M 17 18 19 20 21

2 15 20 15 18-92 742

i 4 7 9 30 10 i 18-56 4 2

Caudal fin branched rays

8 9 10 ii M 8 9 10 ii

24 i 9 -04 5

2 31 i 8-97 2

Scales in lateral series

25 26 27 28 29 M 25 26 27 28 29

2 18 10 6 2 26-68 i 3 2 i

i 17 17 6 i 26-74 i i 2

Vertebrae including urostyle (Manx specimens only)
26 27 28 29 M

2 13 i 27-94
i 3 17 27-76

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

213

Except for pectoral ray and scale counts in the Scandinavian fishes, where the
number of observations is small, the most common value for each meristic character
is the same in the two coloration groups and mean values are only slightly different.
The wider range and somewhat lower means for fin ray counts in scorpioides are due

17-

< 15-

13-

<D 19

<J '*

o 11-

10-

E
A 3-

~D
O

Head Length (mm)

20
Standard Length (mm)

I
30

FIG. i . Relation between standard length or head length and certain body measurements
in the two coloration groups. A, body depth at origin of pelvic fin ; B, depth of caudal
peduncle before origin of caudal fin ; , orca ; O, scorpioides ; O, coincident points.

to the occurrence in this group of all the youngest and smallest individuals, among
which such variation, as may be expected, is more pronounced (Barlow, 19610) .
The specimen of G. orca (ZMO 54023) which prompted Collett (1885) to believe in a
difference in first dorsal count between the two species has been examined and is
the only individual among seventy-four examples of Lebetus to have seven first
dorsal rays. Another fish, identified by Collett (1902) as G. scorpioides, was des-

214

P. J. MILLER

cribed by this author as having only five first dorsal rays. This specimen (ZMB 536)
has also been studied, and was found to possess an abnormal first dorsal fin (not in-
cluded in Table IV) with only four rays, the first three of which are separated by an
unusually large interspace from the last ray.

Body proportions. From their first description, L. orca has been held to be consis-
tently slimmer than L. scorpioides. Collett (1874, 18750) stated that body depth in
the former was contained about seven times in the standard length, but only six
times in G. scorpioides. Holt & Byrne (1903) noted that body depth was equal to one
fifth and one sixth of the total length of G. scorpioides and G. orca respectively, and
referred to this apparent difference in their key to the species. Other body proportions
have been used in attempts at distinguishing these fishes. Collett (18750), Lilljeborg
(1884), Smitt (1892), and Holt & Byrne (1903) all regarded the eye to be relatively
larger in the head in G. orca than in G. scorpioides. A supposed difference in the
position of the anus with reference to the middle of the body was utilized in a key by
Smitt (1892).

The significance of these characters has been tested in the present material and the
results are given in Table V and Text-fig. I. These show that the two coloration
groups cannot be distinguished by such criteria.

TABLE V. Mean and Range of Values for Certain Body Proportions in the Two

Coloration Groups

Coloration group

scorpioides

Body proportions
(a) In standard length
Depth at origin of pelvic fin
Depth at origin of anal fin .
Depth (least) of caudal pe-
duncle before caudal fin
origin

Distance from snout to anus
(6) In head length
Diameter of eye

No. of
Obs.

21
23
23

Mean

5-64

6-50

10-00

1-87

Range

5-10-6-65
5-90-7-30
9-20-11 -o

i 75-2 oo

No. of
Obs.

28
26
29

3-02 2-80-3-20

Mean

5-65
6-55

IO-22

I-8 5
2-98

Range

10-6-30
.70-7-50
50-12-30

.75-1-90
80-3 40

Development of the dorsal fins. Collett (1874, 18750) reported that the first dorsal
fin was considerably higher than the body in G. orca, but only slightly so in G. scor-
pioides. It was also stated that the first rays of the fin were produced and free from
the interradial membrane in G. orca. Duncker (1928) treated the latter as a feature of
diagnostic value, and believed that L. orca differed from L. scorpioides in having the
first dorsal fin higher than the second dorsal fin.

In the present work, fin ray length has been used as an index of fin height because
the latter is difficult to measure accurately. With adults in both the orca and scor-
pioides groups, the second and third rays of the first dorsal fin were longer than any
in the second dorsal fin, and greater than the body depth at the origin of the pelvic

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

215

fin. In length of the second ray, there was little difference between the groups
(Text-fig. 2A), while for the third ray and the other rays of the first dorsal fin, as
well as in the second dorsal fin, there is a more or less pronounced divergence, orca
specimens having the longer rays. This trend is very noticeable in the last three rays
of the first dorsal fin (Text-fig. 2B) and, together with corresponding growth of the
interradial and posterior membranes, gives the fin a much greater spread in the orca

8-
o o

. . o e oo

(B)

0^3 O* O O

n ooo o

1 o o 6-

oo o

4 ~

ooo

oo oo o o

e
o
o

o 3-
o o o

o o oo oo
o oo o ooo
o ooo

o o

i i i i i i i i i i i i i i i i i i i i i
15 20 25 30

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
15 20 25 30

Standard Length (mm)

FIG. 2. Relation between standard length and lengths of certain first dorsal fin rays in
the two coloration groups. (A) Second ray ; (B) Fifth ray. Other symbols as Fig. i.

individuals. In the best preserved specimens of both groups, the tips of the dorsal
fin rays do not project beyond the fin membrane.

Other morphological characters. Comparison of other body proportions and ana-
tomical features, such as patterns of sensory papillae over the head and body, does
not reveal any differences between the orca and scorpioides groups.

Distribution. The two species have been recorded throughout the geographical
range of the genus, which is eastern Atlantic boreal in occurrence. Thus L. orca and
L. scorpioides are both reported from Iceland (Saemundsson, 1926 ; Taning, 1940)
and the western coast of Norway (records summarized by Grieg, 1913). In the south,
they are known from the Gulf of Gascony (Collett, 1896) and Roscoff (Le Danois,

ZOOL. io, 3 I4

216 P. J. MILLER

1910) respectively. It is only in Danish waters and in the western Channel, where
most records are of juvenile specimens, that many individuals of one species (L.
scorpioides} have been obtained and none of the other.

A slight divergence in habitat preference was implied by Duncker (1928) who men-
tioned that L. orca occurred on " fine to coarse sandy ground " while L. scorpioides
inhabited " mostly hard ground (Lithothamnion, shelly sand, gravel) ". Taning
(1940) suggested that L. orca was a more offshore form than L. scorpioides, but Le
Danois (1948) listed both gobies as occuring in the " zone subcotiere " (4o-ioom.)
and deeper parts (120-200 m.) of the continental shelf. While investigating the
bottom fauna off the south of the Isle of Man, Jones (1951) encountered both
Lebetus species on the coarser offshore grounds. The records of L. orca were restricted
to an area of several square miles south west of the Calf Island, where the bottom is
overlaid by a deposit of dead shells, chiefly Glycimeris. Both species were obtained
by Grieg (1913) from similar bottoms in Hardangerfjord, Norway. The positions
of capture for the Manx specimens collected by Jones and later workers have been
plotted in Text-fig. 15, where the number of orca and scorpioides specimens are indi-
cated by symbol at each locality from which fishes have been examined in the present
work. No ecological separation is demonstrable between the two groups and, at locali-
ties where adequate material has been obtained, both orca and scorpioides appear
to be represented in comparable numbers.

The bathymetric range off the Isle of Man is 18-64 m - f r the orca group, and 0.18-64
m. for scorpioides. Previous authors have cited depth ranges of 20-375 m. and 3-5-135
m. for L. orca and L. scorpioides respectively (Holt & Byrne, 1903 ; Grieg, 1913).
The apparently greater penetration of deeper waters by the former species may well
be due merely to lack of collecting. Pronounced contrast in bathymetric distribution
between the two species remains to be established.

Conclusion. The above paragraphs indicate that the two groups to which Lebetus
specimens may be assigned are obviously different only in coloration and hi growth
of the dorsal fins, especially the first. The orca group consists entirely of males, almost
all maturing or mature, and the scorpioides of females and immature males. The
type specimens of Gobius orca (ZMO J3999) and G. scorpioides (ZMO J4O2O, J4O2I)
have been available for study, and are a male, maturing if not mature, and two females
respectively. Sexual dimorphism in colour and fin ray length is a well known
phenomenon among many teleosts, and has recently been discussed by Egami (1960)
for Japanese gobies. It appears safe to conclude, therefore, that the two names apply
to one and the same species, for which the name orca Collett 1874, p. 446, ranks by
page priority as the senior synonym to scorpioides Collett 1874, p. 447.

Lebetus orca (COLLETT 1874)
(Text-figs. 3 and 4)

Gobius orca Collett, 1874 : 446 ; 18750 : 172, pi. Ill, figs. 1-3 ; 18756 : 57 ; Winther, 1877:
49; Lilljeborg, 1884: 616 ; Smitt, 1892: 259, fig. 66 ; 1900: 554; Collett, 1902: 57 ; Holt
& Byrne, 1903 : 57, fig. 5 ; Patience, 1906 : 74 ; Grieg, 1913 : 41 ; Jenkins, 1925 : 88,
97 ; Elmhirst, 1926 : 154 ; Koumans, 1931 : 47 ; Whitley, 1931 : .155 ; Lonnberg &

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 217

Gustafson, 1935 : 246, fig. 2 (photograph) ; Norman, 1935 : 58 ; Jenkins, 1936 : 88, 97,
388 ; Ehrenbaum, 1936 : 168, fig. 147 ; Andersson, 1942 : 85 ; Jones, 1949 : 28.
Gobius scorpioides Collett, 1874 : 447 ; 18750 : ^75, pi- HI, figs. 4-6; 18756 : 58 ; Lilljeborg,
1884 : 620 ; Smitt, 1892 : 260, fig. 67 ; Holt & Byrne, 1898 : 337 ; Grieg, 1899': 7 ; Smitt,'
1900 : 554 ; Collett, 1902 : 59 ; Holt & Byrne, 1903 : 56, pi. II, figs. 2 and 3 : Aflalo, 1904 :
158 ; Holt & Byrne, 1910 : 7 ; Crawshay, 1912 : 300, 385 ; Farran, 1912 ; 2, 3 ; Otterstram,
1912 : 162 ; Grieg, 1913 : 43 ; Clark, 1914 : 372 ; Petersen, 1917 : 10, 13, 14 ; Page, 1918 :
63, 71, 101, figs. 54-56 ; Jordan, 1919 : 392 ; Jenkins, 1925 : 88, 96, pi. 30, figs. 2 and 3 ;
Johansen, 1925 : 5, 6, 10, fig. 6; Saemundsson, 1926: 178; 1927: 19; Koumans, 1931 :
43, 46 ; Whitley, 1931 : 155, 156 ; Lonnberg & Gustafson, 1935 : 247 ; Norman, 1935 : 58 ;
Jenkins, 1936 : 88, 96, 388, pi. 30, figs. 2 and 3 ; Ehrenbaum, 1936 : 168 ; Andersson, 1942 :
85 ; Jones, 1949 : 28 ; Went, 1957 : J 8-

1 cm

FIG. 3. Lebetus orca (Collett), mature male.

Lebetus scorpioides Winther, 1877 : 49 ; Malm, 1877 : 651 ; Winther, 1879 : 18 ; Petersen,
1919 : 48, 49, 55, 59, 61, pi. I, figs. 27-29, pi. Ill, fig. 4 ; Clark, 1920 : 226, 232 ; Russell,
19260 : 134, 152-153, 155-157 ; 19266 : 397, 411 ; Duncker, 1928 : 141, fig. 13 ; Ehrenbaum
et al., 1929 : 155 ; Iljin, 1930 : 55, fig. 43 ; De Buen, 19300 : 125 ; 19306 : 6, 21, 27, figs. 9 and
10 ; Russell, 19300 : 661, 672 ; 19306 : 708, 713, 718 ; De Buen, 1931 : 50 ; Marine Biological
Association, 1931 : 324 ; De Buen, 1932 : pi. 300 ; Russell, 1935 : 153, 154 ; 1936 : 604 ;
J 939 : 386 ; Bal, 19400 : 82, 86 ; 19406 : 16 ; Russell, 1940 : 270 ; Taning, 1940 : 279 ;
Brunn & Pfaff, 1950 : 50 ; Jones, 1950 : 31 ; 1951 : 137, 144 ; Cantacuzene, 1956 : 41 ;
Marine Biological Association, 1957 : 4 O1 -

Lebetus orca Collett, 1885 : 61, pi. I, figs, i and 2 ; 1896 : 41 ; Duncker, 1928 : 142, fig. 14 ;
De Buen, 19300 : 125 ; 19306 : 21, 23, 27 ; 1931 : 50 ; Taning, 1940 : 277, 282 ; Saemundsson,
1949 : 38, 136 ; Bruun & Pfaff, 1950 : 50 ; Jones, 1950 : 31 ; 1951 : 137, 144.

Gobius Jeffrey sii Gunther, 1888 : 210 (partim), pi. Ill, fig. 3 (non Giinther, 1867 : 290, pi. V, fig. c).

218

P. J. MILLER

Gobius (Lebetus) scorpioides Smitt, 1900 : 554 ; Le Danois, 1910 : 168, figs. 4 and 5 ; 1948 : 279 ;

953 : 136 and facing pi., 184.

Gobius (Lebetus) scorpioides guilleti Le Danois, 1913 : 91, 160, 161, 169, 186, 189, figs. 159-161.
Gobius (Lebetus) orca Le Danois, 1948 : 121, 279.
Gobius sp. Schmidt, 1904 : 53 ; Saemundsson, 1908 : 32.

TYPE SPECIMENS. The holotype of Gobius orca (ZMO J3999) is a male from Espe-
vaer, Hardangerfjord, Norway, dredged in 145-180 m. during July 1873. The two
syntypes of Gobius scorpioides (ZMO J4020, J4O2I) are females from Hvittingso,
Stavangerfjord, and Lyngholmen, Hardangerfjord, dredged in 37 and no m., during
July 1872 and August 1873 respectively. In condition, these specimens are now some-
what shrunken, and the caudal fin in each one has been cut off by an earlier investi-

1 cm

FIG. 4. Lebetus orca (Collett), mature female. Pectoral fin removed to show body markings.

gator. Present lengths, original lengths (Collett, 1874, 18750) and meristic values are
given in Table VI.

TABLE VI. Present Lengths, Original Lengths (Collett, 1874, 18750, b) and Meristic
Values for Type Specimens of Gobius orca and G. scorpioides. d = Damaged

Lengths (mm.)

Meristic values

Original standard Present

First

Second

Pect-

Scales

Sex

ZMO No. .

+ Caudal fin

standard

Dorsal

dorsal

Anal

oral

in 1.1.

J 3999

26-0 + 6-0

25-0

I/IG

1/8

26-27

J 4020

total 28-0

21-5

1/9

1/7

J 4 21

30-0 + 7-0

29-0

1/9

1/7

27-28

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 219

GENERAL DESCRIPTION. Body subcylindrical, laterally compressed towards
caudal fin ; in standard length, depth at origin of pelvic fin 5-1-6-65 (mean 5-65), at
origin of anal fin 5-7-7-5 (6-53), of caudal peduncle before origin of caudal fin 8-5-12-3
(10-12). Head rounded ; in standard length, horizontal length (snout to upper origin
of opercle) 3-2-4-1 (371), maximum width (between upper origins of opercles)
5'25-5 '95 (5 '68). Horizontal distance from tip of snout to origin of first dorsal fin
2.7-3-05 (2-9), to origin of second dorsal fin 1-7-1-9 (1-81), to anus 1-75-1-95 (1-86),
to origin of anal fin 1-6-1-75 (1-69), to origin of pelvic fin 3-2-4-0 (3-47), all in standard
length. Caudal peduncle (end of anal fin base to origin of caudal fin) horizontal
length 4-15-5-1 (4-71) in standard length, depth 1-8-2-35 ( 2 *7) m own length. In
head length, snout 3-8-5-6 (4-7), eye 2-8-3-4 (3-0), postorbital length 2-0-2-7 (2*34),
cheek 3-8-4-85 (4-34) ; eye and cheek direct, rest horizontal measurement. Inter-
orbital minimum width (bony) 4-3-7-4 (5-5) in eye length. Snout less than diameter
of eye, with moderately sloping profile ; nostrils adjacent, anterior on short tube,
not projecting over upper lip and with entire distal rim lacking local prolongation ;
posterior nostril also tubular but shorter. Eyes large, close together. Postorbital
region never exceeding half head length ; profile more or less flat, without dermal
ridges. Upper jaw relatively wide, about equal to preorbital area. Mouth oblique,
lower jaw at most only slightly in advance of upper ; maxilla ends posteriorly under
anterior quarter of eye. Lower jaw unarmed, without barbels. Cheek smooth, without
dermal ridges. Preopercle unarmed. Branchiostegous membrane attached to entire
lateral margin of isthmus from immediately anterior to lower edge of pectoral lobe and
without transverse fusion across isthmus (Text-fig. 5). Teeth simple, caniniform, erect
in both jaws. Dentition of jaws an outer row of larger teeth with narrow band of two
to four rows of smaller teeth. Pharyngeal teeth relatively large, recurved, in two
superior and one inferior median patch (see description of skeleton). Tongue more or
less truncate, with anterior edge rounded laterally ; at most pronounced, a very
slight median emargination (Text-fig. 6s). Gill rakers small protuberances, not
spinous, about 6-8 on first arch. Sagittal otolith rectangular, anterior and posterior
edges slightly concave, inferior edge slightly convex, superior edge with angular
projection near posterior end ; outer surface convex, more flattened in middle, inner
surface convex, with furrow in middle and shallower groove to posterior edge (Text-
fig. 6A).

FINS. In the following counts, spinous elements are represented by large roman,
articulated rays by arabic numerals. The terminal bifid ray in the second dorsal
and anal fins is counted as one element. Only in the caudal and pelvic fins are articu-
lated rays branched. The range of values is given in parentheses after the most
common number, frequency distribution being shown in Table VII.

First dorsal fin VI (VI-VII) ; second dorsal fin 1/9 (6-n) ; anal fin 1/7 (5-8) ;
caudal fin (branched rays) 9 (8-10) ; pectoral fin 19 (15-21), pelvic fin 1/5 -f- 1/5.
In standard length, first dorsal base 7-95-8-95 (8-24), second dorsal base 3-75-4-55
(4-24), anal base 4-85-5-8 (5-4), longest caudal ray 3-2-4-45 (for standard lengths up
to 20-0 mm., mean = 3-55 ; more than 20-0 mm., mean 3-92), longest pectoral
ray 3-7-4-45 (4-0), pelvic origin to tip of fifth pelvic articulated ray 3-85-4-55 (S.L.
up to 25 mm., mean = 4-05 ; more than 25-0 mm., mean = 4-2). First dorsal fin

P. J. MILLER

TABLE VII. M eristic Characters of Lebetus orca : Number of Observations

against Value

Locality

British Isles
Scandinavia

Isle of Man

First dorsal fin rays
5 6 7
60

Second dorsal fin articulated rays
6 7 8 9 10 ii

i 2 5 39 13

2941

Anal fin articulated rays
567*9
3 3 4i 13
i 9 5

Pectoral fin rays

75 16 I^ 18 19 20 21
i 4 9 24 50 25 i

784

Caudal fin branched rays
8 9 10 ii
2 55 2
7

Scales in lateral series
25 26 27 28 29
3 35 27 12 3
1433

Vertebrae including urostyle
26 2j 28 29
i 5 30 i

TABLE VIII. Dorsal Fin Ray Lengths Expressed in Standard Length for Mature and

Immature Lebetus orca

Mature <Jo*
Mature ?$ .

Immature 0*6*
Immature $$

First dorsal fin

Second dorsal fin

III

3-85-

3'4-

3'35-

3-20-

3-25-

4-15-

4'35-

4-10-

5-20

4-35

3-95

4-70

5-90

5-80

4-90

4-95-

3-20-

3-95-

5'40-

6-35-

9'5-

6-40-

5-75-

6-50

4-50

5-30

7'25

9'35

I3-50

8-25

7-05

(21-0)

4-65-

3-65-

4-00-

4'95-

5-30-

7-25-

6 05-

5-90-

6-80

6-05

6-80

9-10

13-60

9-20

7'25

6-05-

4-15-

4'45-

5'4-

7-00-

I I OO-

6-80-

5-40-

9'35

6-40

5-80

7-15

8-50

14-40

8-50

7-00

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 221

commences behind vertical of pelvic origin, last ray arising well before vertical of anus.
In standard length, dorsal fin ray lengths as in Table VIII (see also Text-figs. 19 and
20). None of dorsal rays prolonged beyond interradial membrane. In most males, no
dorsal interspace, posterior membrane of first dorsal reaching origin of second dorsal
fin, sometimes extending to one side of first element of latter. In females and some
immature males, a short dorsal interspace. Second dorsal fin begins over or just behind
vertical of anus, finishes opposite posterior end of anal fin. Anal fin commences below
first elements of second dorsal fin, and ends below last element of latter. In standard
length, spinous ray 875-10-4 (males), 10-45-12-85 (females) ; first articulated ray

5 mm

FIG. 5. Ventral view of female L. orca. F, fold of attached branchiostegous membrane ;
u, origin of branchiostegous attachment to isthmus ; P, pectoral fin ; A, anal fin.

6-65-8-95 (males), 8-2-10-7 (females) ; penultimate ray 7-1-9-7 (males), 9-2-12-45
(females). Caudal fin with rounded margin, convex posteriorly. Pectoral fin extends
back to opposite anus ; uppermost pectoral rays not separated from fin membrane ;
no dermal flaps on anterior edge of pectoral girdle. Pelvic fins not adnate, completely
united along their length but with anterior membrane between spinous rays entirely
absent ; fourth articulated ray somewhat prolonged, terminating lateral to anus or
genital papilla, but fifth ray ending immediately before anus (Text-fig. 5).

SCALES. Behind vertical of anus, body covered with large imbricating ctenoid
scales ; anteriorly, ctenoid scales present in narrow belt on lateral midline forward

222

P. J. MILLER

to axilla. Lateral series 26-27 (25-29), transverse series (anterior end of second dorsal
backwards to anal fin) 7 (see Table VII). Scales absent from head, predorsal region,
upper part of body lateral to first dorsal and origin of second dorsal fin, pectoral lobe,
axilla, isthmus, breast, and abdomen. Midline scales of caudal peduncle hexagonal,

(A)

FIG. 6. L. orca : (A) Sagittal otolith, median view, (B) Tongue, showing extremes of
variation in shape of anterior edge, (c) Scale from caudal peduncle (sclerites omitted),
a, anterior ; p, posterior ; s, superior ; i, inferior. Scale with each figure i mm.

somewhat elongate, with about 20 ctenii (Text-fig. 6c) ; scales at base of caudal fin
without elongate lateral ctenii.

VERTEBRAE. Usually 28 (26-29), including urostyle (Table VII).

COLORATION. Basic pattern of more or less well developed vertical bars ; first
and widest below anterior part of first dorsal fin, second below posterior end of first

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 223

dorsal, third below anterior end of second dorsal, and fourth a curved mark across
caudal peduncle posterior to termination of second dorsal and anal fins. Bars tend
to be connected along lateral midline by denser pigment but, nearer dorsal and ventral
edges of body, are separated by paler areas. Caudal peduncle from fourth bar to
origin of caudal fin almost colourless with some pale yellowish or ferruginous mottling
in life ; a vertical dark mark across origin of caudal fin.

(a) Mature male (Text-fig. 3 and PL I) . Body yellowish or more dusky grey ;
bars greyish, somewhat less well marked than in female, (especially after preservation,
c.f. Text-fig. 3 and PL I) but posterior edge of caudal peduncle bar clearly defined
in life and caudal peduncle more or less colourless. Head yellowish to reddish grey
above ; underside of head, cheeks, isthmus, and breast suffused with reddish orange.
First dorsal fin dusky yellow, edged with white. Second dorsal fin with intense black
edge, especially wide anteriorly on distal part of first two interradial membranes, and
broad oblique yellow bands separated from edge and one another by white bands,
which, excepting posterior band, usually become narrower and greyish towards base
of fin (wide bands to base of second dorsal shown in PL I are exceptional). Anal
fin dusky, without white edge. Pectoral and pelvic fins dark ; pectoral suffused with
orange red proximally, being most intense over bases of lower pectoral rays. Caudal
fin usually clear, or with traces of yellow or red.

(b) Mature female (Text-fig. 4). Body pale brown ; vertical bars conspicuous
purplish brown with anterior edge of first and posterior edge of caudal peduncle bar
sharply defined ; fusion and irregularity of intermediate bars may produce effect of
coarse mottling, and sometimes a pale transverse band lies across origin of first
dorsal fin between level of pectoral bases and anterior edge of first vertical bar. Head
pale brown, reddish or ferruginous above ; reddish brown vertical marks across cheek
and immediately behind preopercle ; red striae on underside of head across mental
region, at angle of jaw, and below eye. First dorsal fin with oblique yellow to orange red
bands each bordered with black and tending to coalesce across middle interradial
membranes ; uppermost band wider, darkening posteriorly to form more or less well
defined spot on distal part of last interradial membrane ; edge of fin membrane may
be whitened especially adjacent to spot. Second dorsal fin with narrow dark
edge, and four narrow oblique bands of orange red, each bordered with black,
across hyaline fin. Caudal, pectoral, and pelvic fins usually hyaline ; a reddish
brown mark on upper part of pectoral lobe. Anal fin pigmented on proximal areas
of posterior interradial membranes.

(c) Immature fishes. Both sexes resemble mature female. First dorsal fin of males
usually more uniform in colour but may show bands and spot ; second dorsal fin
with conspicuous dark spot at distal end of first and second interradial membranes,
as well as rather wider oblique bands and more intense dark edge. In youngest
individuals, pelvic and anal fins black, and dark vertical bar across pectoral fin ;
underside of head, isthmus, and breast covered with melanophores, which later
disappear.

LATERAL LINE SYSTEM. Terminology and lettering of the sensory papillae in the
present text and diagrams follow the system invented by Sanzo (1911). As part of
the general reduction in number of papillae, in certain cases only one or at most a

ZOOL. 10, 3 *4

224

P. J. MILLER

few papillae occur in areas where a well denned row or rows are found in the genera
described by Sanzo. Under these circumstances, it is not possible to be sure of the
exact homologies of the papillae in Lebetus so that in some instances composite
lettering has been adopted.

(a) Cephalic canals. These are absent.

(A)

(B)

FIG. 7. Sensory papillae of head in L. orca. (A) Lateral, (B) Dorsal, (c) Ventral views.
AN, anterior nostril ; PN, posterior nostril ; D, first dorsal fin ; p, pectoral fin ; v,
pelvic fin. Other abbreviations as in text.

(b) Sensory papillae. Relatively large, and easily seen. Numbers of papillae are
given in parentheses.

(i) Preorbital (Text-figs. 7A, 78). Median series in two parts, anterior s (i) and
posteriorly rs (2, inner sometimes absent) ; another papilla rarely present between

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

225

rs andp 1 of interorbital series. Lateral series c 1 (5) and c 2 (i), the latter rarely present,
(ii) Suborbital (Text-fig. 7A). Infraorbital row a (5) continued as a 1 (3) into oculo-
scapular groove. Longitudinal row c (2-3, rarely 4). Between rows a and c a single
papilla, probably referable to c. Row b absent. Row d represented by d 1 (2), without
horizontal posterior section.

pter

soc

vert 1

(A)

1 mm

(B)

pro

FIG. 8. Neurocranium of L. orca. (A) Dorsal, (B) Ventral views. Abbreviations as in text.

(iii) Preoperculo-mandibular (Text-figs. 7A, 70). External row e in two parts,
posterior e 1 (3-4, rarely 5), and anterior e z (2-3). Internal row i in three parts ;
superior i 1 (2), posterior i 2 (7-8), and anterior i 3 (2-3). A single papilla ei usually
present halfway between angle of jaw and mental region. Mental row eif (2-4).

(iv) Oculoscapular (Text-fig. 7A). Longitudinal row u of oculoscapular groove in
two parts, anterior u 1 (i), and posterior w 2 (3). One papilla z anterior to commence-

226 P. J. MILLER

ment of preopercular groove, another x over u 2 , and a third xy above border of
opercle. Axillary series las in three vertical rows (2 ; 2 ; and 2-3, rarely 4,
respectively).

(v) Opercular (Text-fig. 7 A). Transverse row ot in two parts, upper ot 1 (4,
very rarely 5) and lower ot 2 (2) ; superior longitudinal row os (2) ; and inferior longi-
tudinal row oi (i).

(vi) Anterior dorsal (occipital) (Text-fig. 73). Three groups, anterior n (3, rarely 4),
middle gm (3), and posterior h (2).

(vii) Interorbital (Text-fig. 76). Two papillae, anterior p l and posterior p 2 ;
former occasionally absent.

(viii) Trunk (not illustrated except lv 1 ). Three series ; superior with Id 1 (1-2)
flanking first dorsal fin near origin, Id 2 (2) near posterior membrane of first dorsal
fin, and Id 3 (3) on caudal peduncle near origin of caudal fin ; median with about nine
or ten short transverse rows Urn 1 to Itm 9 or 10 (first usually 4, rest 3 each) along midline
to origin of caudal fin ; inferior with lv 1 (4-8) between origins of pelvic and pectoral
fins (Text-fig. 70), lv 2 of one, rarely two, lateroventral rows on abdomen (3-4 each),
and lv 3 (2) just before anus.

(ix) Caudal (not illustrated). Three papillae around posterior edge of last scale of
midline which overlaps base of caudal fin, and up to 12 papillae, from middle papilla
of former group, in line on interradial membrane to near posterior edge of fin. Some-
times a parallel series on membrane below.

SKELETON. This is well ossified.

(a) Neurocranium (Text-figs. 8, 9). (i) Proportions. Postorbital region broad,
slightly wider than long, depressed ; interorbital septum strongly compressed
especially in dorsal aspect ; preorbital region small.

(ii) Occipital bones. Basioccipital (bo) with broad forks truncate anteriorly,
articulating posteriorly with centrum of first vertebra (vert. i). Exoccipital (eo)
carrying posteriorly stout articular process opposed to upper surface of prezygapo-
physis of first vertebra, extending ventrally along side of skull with knob near distal
end to which inferior limb (pti) of posttemporal is attached by strong ligament, and
dorsally in contact with fellow behind supraoccipital. Supraoccital (so) relatively
large, diamond shaped, bearing prominent crest (soc), which is triangular with emargi-
nate posterior edge and pointed apex and continues backwards on to anterior part of
exoccipital area.

(iii) Otic bones. Sphenotic (sph) at anterior corner of neurocranium produced into
horizontal flange for articulating with hyomandibular, and with wide posterior
extension separated from epiotic ; a dermosphenotic element (dsph) forms part of
posterior border of orbit, firmly attached to sphenotic proper. Pterotic (pter) forming
posterior corner of neurocranium with articular surface on ventral face for hyomandi-
bular. Epiotic (ep) large, in contact with superior limb (pts) of posttemporal,
approaching fellow in dorsal midline but separated by posterior end of supraoccipital.
Prootic (pro) broad, meeting basioccipital. No opisthotic, but on ventral surface of
neurocranium, a cartilaginous area (c) bounded by prootic, pterotic, exoccipital, and
to a slight extent by basioccipital.

(iv) Preorbital neurocranium. Mesethmoid (me) with cartilaginous body, ossified

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

227

as posterior vertical lamina at anterior corner of interorbital space surmounted by
triangular ossification, from whose anterior end a thin, dorsally concave plate extends
forwards over cartilage towards vomer, and with lateral ossified area adjacent to
lateral ethmoid. Vomer (v) consisting of well ossified transverse bar across anterior
end of ethmoid cartilage with thin scale of bone extending backwards over ventral
surface of ethmoidal region and end of parasphenoid. Lateral ethmoid (le) attached
to side of ethmoid cartilage ; proximal part of upper surface and adjacent meseth-
moid forming articular surface for ethmoidal head of palatine, and, laterally, alate
in border of orbit. Prefrontal (pf) falciform, attached to anterior face of lateral
ethmoid wing and distally related to maxilla.

(v) Other dermal bones. Frontal (/) large, overlapping sphenotic, pterotic, epiotic,
and supraoccipital, ending anteriorly below posterior edge of mesethmoid. No
parietal. Parasphenoid (p) large triangular plate covering edges of basioccipital

I mm I

FIG. 9. Preorbital cranium of L. orca. (A) Dorsal, (B) Lateral views. Abbreviations as in
text. Note that only outer row of teeth is shown on premaxilla and dentary.

and prootic with anterior process terminating below mesethmoid and above vomer.
Between orbits, frontals and parasphenoid much compressed, forming upper and
lower boundaries of interorbital septum, which is otherwise unossified save for
mesethmoid lamina. Nasals and suborbitals absent.

(b) Branchiocranium (Text-figs. 9, 10, n). (i) Suspensorium. Hyomandibular (h)
broad, with two heads ; anterior head articulating with lateral flange of sphenotic,
posterior head with anterior end of ventral face of pterotic ; posteriorly, strong stay
on inner face articulating with operculum ; posteroventrally, sword-shaped process
extends over upper end of preoperculum. Symplectic (sy] a slender bone running
along lower anterior corner of hyomandibular to inner surface of quadrate almost to
angle of jaw. Metapterygoid (mpt) elongate, along anterior edge of hyomandibular
and symplectic but not reaching quadrate. Quadrate (q) with anterior triangular
part bearing articulation for lower jaw at inferior corner ; a broad curved blade
produced backwards from posterior edge and applied to inner face of preoperculum.

228

br4

br3

art

(B)

FIG. 10. Suspensorium of L. orca. (A) Lateral, (B) Median views, (c) Interhyal and
epihyal bones, median aspect. Abbreviations as in text.

Preoperculum (pop) an integral part of suspensorium with strongly ridged posterior
edge ; upper part of wide anterior lamina firmly united with hyomandibular over
wide area. Between preoperculum, symplectic, and quadrate exists a wide oval non-
osseous foramen.

(ii) Palatopterygoid arch. Consists of two bones, pterygoid (ptg) a posterior splint
applied to upper anterior corner of quadrate and lying along lower part of palatine,
and palatine (pal) larger, anterior end with two heads, posterior head (ppal) articu-
lating with ethmoid region, anterior (apal) with posterior face of proximal end of
maxilla.

(iii) Upper jaw. Maxilla (mx) bearing outer ridge ; distal end blunt with rounded

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

229

posterior margin ; proximal end with anterior lamina below posterior extension of
premaxilla, and accommodating behind the anterior head of palatine. Premaxilla
(pmx) pointed distally ; at proximal end, posterior tongue overlies process of maxilla ;
medially, a long laterally compressed rod ascends over mesethmoid and together
with fellow is embedded in oval plate of cartilage (not illustrated) related to meseth-
moid groove. Premaxilla toothed along entire anterior edge.

(iv) Lower jaw. Articular (art) with large ventral plate. Angular (an) small, at
posterior corner of articular. Dentary (d) toothed along upper edge anterior to
relatively small blunt coronoid process.

(v) Hyoid arch. Consists of relatively large interhyal (ih), broad epihyal (eh), and
large ceratohyal (ch) abruptly narrowing halfway along its length. Hypohyal (hh)

(A)

1 mm

FIG. ii. L. orca : (A) Branchial skeleton, dorsal view. (B) Glossohyal, dorsal view, (c)
Urohyal, lateral view, a, anterior ; p, posterior. Other abbreviations as in text.

firmly applied to anterior end of ceratohyal. Glossohyal (gh) fan shaped with straight
anterior edge (Text-fig. IIB). Urohyal (uh) a vertical lamina with strongly emarginate
posterior edge (Text-fig. lie).

(vi) Branchial arches. Pharyngobranchials (pb i, 2) represented by two dentigerous
plates, the more posterior much larger and overlapping anterior one, opposite medial
end of last three epibranchials. Epibranchials (eb) four, first one forked. Cerato-
branchials (cb) four. Ventrally, first two hypobranchials (hb 1, 2) distinct ; a small round
ossification (hb 3) in position of third, and a median cartilaginous plate (hbc) related
to lower ends of fourth ceratobranchials. Fifth branchial arch represented by ventro-
lateral subtriangular dentigerous plate (dp) approaching fellow but separated in
ventral midline. Ossified basibranchials absent.

(vii) Opercular series. Operculum (op) and suboperculum (sop) somewhat vertically
elongate. Interoperculum (iop) long, anteriorly connected to angular by strong

230

P. J. MILLER

ligament, posteriorly to outer face of epihyal near articulation of latter with interhyal.
Branchiostegous rays (bv 1-5) five, uppermost and broadest attached to epihyal,
next three to broad part of ceratohyal, and lowest and shortest to narrow section of
ceratohyal.

(c) Pectoral Girdle (Text-fig. IZA). Cleithrum (cl) large, slightly bifid at upper end,
in contact with fellow in ventral midline ; at level of hypocoracoid, broad triangular
process (Text-fig. I2B, ctp) meets pelvic element. Cleithrum joined by small supra-
cleithrum (scl) to posttemporal (pt], which has two well developed anterior arms, with
upper, stronger, to epiotic (pts), and lower, more slender, to exoccipital (pti). No
hypercoracoid. Hypocoracoid (he) small, at lower end of vertical part of cleithrum.

(A)

(B)

1 mm

FIG. 12. L. orca : (A) Pectoral girdle, lateral view. (B) Pelvic girdle, ventral view.

Abbreviations as in text.

Radials (r 1-4) large plates set in cartilage lamina attached along median side of
posterior edge of cleithrum, with lower corner of most inferior radial joined by
cartilage to upper part of posterior edge of hypocoracoid.

(d) Pelvic girdle (Text-fig. IZB). Each innominate bone consists of stout lateral
section curving medially to form with fellow the posterior lip of girdle on which
pelvic rays articulate except at median blunt protuberance, and a wide lamina
arching dorsally and medially to join that from opposite side along median longi-
tudinal suture. At posterior apex of girdle, a rod of bone (rd) with expanded base
projects forwards ventral to lamina. Broad anterior ends of innominate bone related
along lateral edge to median face of triangular process (ctp) from cleithrum (cl) of
pectoral girdle.

(e) Axial skeleton. Vertebrae usually 27 (excluding urostyle) ; first five centra
progressively shorter towards occiput, with first two shorter than wide. Third to
thirteenth inclusive carry pleural ribs ; twelfth to last complete vertebra with closed

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

231

haemal arches ; first sixteen bear epipleurals. Last two epipleurals only ossified
distally ; in preceding two, ossified in two parts.

(/) Caudal skeleton (Text-fig. 13). Two large hypurals (hyp], upper fused to urostyle;
along upper and lower edges of hypural fan lie splintlike epaxial (epr) and hypaxial
(hyr) caudal radials. Lying freely between upper hypural and expanded neural arch
of last vertebra occurs large roughly triangular epural plate (epu}. Ventrally this
may bear vestige of neural arch (vna), another part of which (vnb) occurs on anterior
half of urostyle (ur). Neural (nsp} and haemal spines (hsp) of last three or four
vertebrae are progressively expanded towards urostyle ; on fusion of neural arch
with postzygapophysis, root of arch comes to lie along whole length of centrum.
Similar spread of haemal arches. Minor abnormalities found in caudal skeleton

vna

epu

epr

nsp

(C)

FIG. 13. Caudal skeleton of L. orca. (A) Normal. (B) Without vestigial neural arch.
(c) With complex vertebra. Abbreviations as in text.

include loss of vestigial neural arch (vna, vnb}, and a complex vertebra (cv), shown in
Text-figs. 136, c.

BIOLOGY

GEOGRAPHICAL DISTRIBUTION (Text-fig. 14). The genus Lebetus seems to be
restricted to the European Atlantic boreal region, as defined by Ekman (1953).
Outside the broad limits of this area, there are a few records of larval fishes from the
Straits of Gibraltar, doubtfully assigned to Gobius scorpioides by Fage (1918).
As yet, only young stages are recorded from southwest Iceland and the Faeroes ;
mature individuals no doubt remain to be discovered by systematic dredging in these
waters. Lebetus does not appear to extend from the Belt Sea into the Baltic, and is
also unknown from the southern North Sea and the eastern Channel. In the British

232

P. J. MILLER

fauna, this goby was first known from a single example (BMNH 88. 3.22.9-10 part)
obtained in Kilbrennan Sound and erroneously believed by Giinther (1888) to repre-
sent the mature male of his Gobius jeffreysii. Collett (1896) drew attention to this
mistake, and in 1898 Holt & Byrne recorded a ripe female (as G. scorpioides} from
Falmouth Harbour. A full list of records covering the entire area of distribution is
provided in the Appendix.

65 N 3OW 25 2O 15 IO

IO 15 2O 25 3O

- 45

4O -

FIG. 14. Geographical distribution of L. orca. Locality and not number of records is
indicated. Open circles denote localities from which there are only records specified as
planktonic.

ECOLOGICAL DISTRIBUTION. Lebetus orca has been taken over a wide bathymetric
range, from about 2 m. (Holt & Byrne, 1903) to 375 m. (Lilljeborg, 1884 ; Collett,
1902), chiefly from coarse grounds but also on mud (Patience, 1906). Le Danois
(1913) regarded Gobius scorpioides guilleti as a characteristic inhabitant of the
" maerl " deposits of calcareous algae (Lithothamnion spp, Lithophyllum] in the bays
and estuaries near Roscoff although, to judge from the listed fauna and flora, these
estuaries could hardly have shown a marked reduction in salinity. This type of

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 233

bottom was classified by Le Danois among the sandy and muddy sand areas of the
" littoral " zone, extending down to 40 m. Together with representative animals,
including Lebetus, a stretch of maerl is imaginatively depicted in water-colours by
Le Danois (1953).

In Manx waters (Text-fig. 15), where the bottom fauna and deposits have been
reviewed by Jones (1951) and Southward (1957), the species has been found on
grounds ranging from muddy sand to stones and coarse gravel, always with dead
shells. Calcareous algae (e.g. Lithothamniori) are common on stones in this area ;
near the shore, larger weeds may be present, and other localities where Lebetus has
been collected are notable for the abundance of the polyzoan Flustra, various

FIG. 15. Distributionof L. orcaoft the southern end of the Isle of Man. Position of areas
A and B shown in c. Symbols indicate number of individuals in orca and scorpioides
coloration groups respectively thus : , i orca ; D, i, and A, 2 scorpioides ; O, i
orca and i scorpioides; station i, 2 orca and 3 scorpioides; station 2, 10 and n ;
station 3, 4 and 3 ; , records by Jones (1950) not examined by the author. A, stones,

gravel, shell ; B, fine sand ; c, muddy sand ; D, mud ; M, Modiolus epifauna ; ,

approximate limits of the various deposits; - . . -,20 fm. (36-6 m.) ; -...-,
30 fm. (54-9 m.) ; - . . . . -, 40 fm. (73-2 m.). Scale in nautical miles. Br. Bradda
Head ; Ca, Calf Island ; CR, Chicken Rock ; Do, Douglas ; PE, Port Erin ; PStM,
Port St. Mary ; SH, Spanish Head.

ophiuroids, or the lamellibranch Modiolus modiolus (L.). Off the Isle of Man, Lebetus
has been found at depths from 18 to 51 m.; trawling in shallower water over the
sandy bottom of Port Erin Bay and runner-dredging in deeper water on mud and
finer muddy sand to the west of the island have both failed to reveal the occurrence
of this species, even when the collecting gear included a bag of fine-meshed material.

234 p - J- MILLER

Le Danois (1913) suggested that the rosy or violet coloration of Lebetus was a cryptic
adaptation to a background of Lithothamnion, but the species has been obtained in
the Manx area from localities where calcareous algae are missing from the dredge haul.
In general, Lebetus is probably common on the coarser grounds around the southern
end of the Isle of Man and may extend on to softer deposits where ample cover in the
form of dead shells is available. Other small teleosts obtained in the same area
include the rockling Ciliata septentrionalis (Collett), the dragonets Callionymus
maculatus Rafinesque and C. reticulatus C. & V., the blennioids Blennius ocellaris L.,
Pholis gunnellus (L.) and Chirolophis ascanii Walbaum, the gobies Buenia jeffreysii
(Giinther) and Pomatoschistus pictus (Malm), the cottid Acanthocottus lilljeborgi
(Collett), and the clingfish Diplecogaster bimaculata (Bonnaterre) .

DIET. A total of 48 guts were examined, all from the Manx region. Nine of these
were completely empty, and a further eight contained only a trace of food. The
method adopted in assessing diet was a modification of the points system described
by Hynes (1950). The number of points awarded to a stomach depended on fullness,
20 being scored by a full stomach, and this was then divided between the food items
present in relation to their volume as estimated by eye. A proportionate number of
additional points was allotted to food in the remainder of the gut. Half points were
awarded to items occurring in very small quantities ; in practice this convention
enables the more insignificant items to be considered with the rest and does not
result in any obvious artificial exaggeration of the importance of these constituents.
Results are shown in Table IX.

Rather more than half of the gut contents was made up of Crustacea. Three species
of decapods together comprised the most important part of this ; amphipods, chiefly
Gammaridea, formed another major component but only Amphilochus manudens
Bate was identified among this group. None of the ostracods or harpacticoid copepods
were named although these comprised an appreciable fraction of the diet. Polychaetes
and lamellibranchs were the other notable food items. The vast majority of polychaete
material consisted of the polynoid Harmathoe, which was by far the most abundant
animal of any group identified in the food. About half the lamellibranch remains
were unidentifiable ; in the rest, the occurrence of an intact example of Musculus
marmoratus (Forbes), 2-5 mm. long, in a fish of only 25 mm. may be noted. These
findings suggest that Lebetus, like many other gobies, is exclusively predatory, feeding
on small demersal animals. The species named are all typical of the coarser grounds
inhabited by Lebetus. The presence of a serpulid stalk and cirripede thoracic limbs in
two guts indicates that rather larger animals may be attacked, probably when the
goby is attracted by movement of some appendage.

ENDOPARASITES. The parasites recorded in Table X were found during removal
and inspection of guts from the Manx specimens of Lebetyts. None appears to be
common. Ectoparasites were not seen on any of the fishes studied.

REPRODUCTION, (a) Reproductive organs. The male genitalia of Lebetus (Text-fig.
16) display several peculiar features. The mature testes are relatively small, oval to
moderately elongate, somewhat flattened bodies lying against the roof of the ab-
dominal cavity, up to 1-5 mm. long by o-ig-o^S mm. wide. Each testis has about
20-30 transverse lobules, enclosed by a definite tunic and emptying into a median

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

235

TABLE IX. Gut Contents, Number of Points Awarded, and Percentages of Total Points
Awarded to Identifiable Gut Contents for More Important Constituents

% of total points

No. of awarded to identifiable

points gut contents

POLYCHAETA

Harmathoe sp. (impar ?) . . . . 58

Pholoe minuta (Fabr.

r-)\

Phyllodocid indet. j

Serpulid opercle and stalk
CRUSTACEA total

Crustacean debris

Ostracods ....

Harpacticoid Copepods

Amphipods indet.

Amphilochits manudens Bate

Paratanais batei G. O. Sars

Cirripede appendage

Caprellid indet.

Decapod appendages

Hippolytid (Spirontocaris ?) .

Galathea sp.

Pagurus bernhardus (L.)
PYCNOGONIDA

Ammothea sp.

Pallene sp. .
MOLLUSCA

Shell fragments .

Gastropods indet.

Lamellibranch indet.

Musculus marmoratus (Forbes)

Chlamys sp., newly settled

Hiatella arctica (L.)

ECHINODERMATA

Ophiuroid indet.
SAND GRAINS
UNIDENTIFIABLE MATERIAL

134

32*

5
i

17
3*

12 J

7
i

25-3

55-4

4'4
15-0

21-8

12-6

TABLE X. Endoparasites Recorded from 48 Manx Examples o/Lebetus orca

Parasite

TREMATODA DIGENEA
Plagioporus sp.

Derogenes sp., unencysted

metacercaria

CESTODA PSEUDOPHYLLIDEA
Bothriocephalus larvae .

NEMATODA ASCAROIDEA
Anisakid larvae

No. of fishes
infected

Site of infection

intestine (3),
rectum (2)
stomach

intestine (i),
rectum (i)

abdominal cavity

236

P. J. MILLER

longitudinal sperm duct. There is no pronounced band of the so-called interstitial
tissue which occurs in the testes of some other gobiids (Eggert, 1931 ; Vivien, 1939 ;
Coujard, 1941). Posterior to the testes, each sperm duct pursues a long and indepen-
dent course to the seminal vesicle of that side. This condition appears to be derived
from one in which the testis was longer with spermatogenic lobules along the entire
length of the duct, since vestigial lobules are sometimes visible on the free duct in
immature fishes (Text-fig. 160). The seminal vesicles, small in the immature male
(Text-fig. i6c), increase greatly in size with maturation, the anterior tips in some
cases reaching as far forward as the posterior ends of the testes. They are transparent
in life, conspicuously lobulate, without specialized regions, and empty into the united
sperm ducts, which also receive the secretion from a small accessory glandular mass

i 1 mm I

A,B,C

ag-

1 mm

(A) (B) (C) (D)

FIG. 1 6. Male genitalia of L. orca. Mature (A) dorsal view, (B) ventral view, and (c)
immature, ventral view. (D) Sperm duct with vestigial lobules in immature fish,
t, testis ; sd, sperm duct ; sv, seminal vesicle ; ag, accessory gland ; ed, ejaculatory
duct ; vl., vestigial lobules.

at the posterior ends of the seminal vesicles. The urogenital papilla of the mature
male (Text-fig. iyA) terminates in a circlet of long vascularized processes. During
fertilization these may perhaps embrace the tip of the egg and achieve an economy of
sperm correlated with the small size of the testes. The surface of the male urogenital
papilla carries a scattering of melanophores in adult specimens. In the female, the
ovaries present the usual cystovarian appearance and are paired. Ripe oocytes reach
a maximum diameter of probably about 0-6 mm. (Holt & Byrne, 1898 ; personal
observation). The urogenital papilla (Text-fig. 176) is stouter than that of the male,
with the terminal villose processes shorter and broader, and, as in the immature male
(Text-fig. 170), lacks surface pigmentation.

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

237

(b) Breeding season. Holt & Byrne (1898) recorded a ripe female from Falmouth
Harbour in mid- July, and, in the nearby Plymouth area, Russell (i93o&-i94o)
found postlarval Lebetus from May to October and also in December, with August as
the month of greatest abundance. Bal (19400, b) obtained postlarvae in late June and
July off Port Erin, Isle of Man. At Roscoff, Cantacuzene (1956) stated that the spawn-
ing season was from April to September. In the Manx material, females with swollen
ovaries containing opaque yellow oocytes were present from mid-February to mid-
June. Because of reduction in the amount of dredging performed, only one specimen
a male was obtained in July and none at all in August. In early September a
female was taken with ovaries showing possible traces of oocyte resorption, a process
characterizing the termination of the breeding season in other gobies (Prasad, 1959 ;

(B)

(C)

1 mm

FIG. 17. Urogenital papillae of L. orca. (A) Mature male. (B) Mature female, (c) Immature
male, a, anus ; af, origin of anal fin.

Miller, 19610) . Fully mature males were obtained from February to July. In all
probability the breeding season commences in March and may extend to August.
Within such a period it seems possible that a number of broods may be produced by
one female, but attempts to investigate this from size frequency distribution of
oocytes have been unsuccessful because of failure to obtain a fully ripe ovary. The
fertilized egg and early developmental stages of this species are unknown.

(c) Fecundity. The ovaries of four females of standard lengths 24-0, 25-0, 25-5, and
28-0 mm. contained respectively 185, 140, 270, and 194 ripening oocytes. The rela-
tively low number of these is in accordance with the small size of this goby and are
comparable to values made available by Dotu for some small Japanese species. Thus
individuals of Paleatogobius c/wW00 Takagi, 23 to 38 mm. in standard length, possessed
158-430 " ovarian eggs ", Luciogobius saikaiensis Dotu, 25-32 mm., 66-148, and
Eutaeniichthys gitti Jordan & Snyder, 35-40 mm., 86-310 (Dotu, 1955, 19570 ; Dotu
& Mito, 1958). In Gobiosoma robustum Ginsburg from Tampa Bay, Florida, Springer

238

P. J. MILLER

& McErlean (1961) recorded 56-462 ' large eggs ' in females from 15-28 mm. How-
ever, larger numbers were found in Acentrogobius masago (Tomiyama) (14-26 mm.,
264-961) and Tridentiger undicervicus Tomiyama (24-38 mm., 1,812-5,770), while
Gobius lidwilli McCulloch, a species smaller than L. orca, contained from 154-311
eggs at 15-16 mm. (Dotu, 19576, 19580, &). In the Philippine Mistichthys luzonensis
Smith, where adult females attain lengths of 11-14-5 mm., Te Winkel (1935) counted
only 20-40 mature ova per fish.

(d) Sexual dimorphism. This is expressed in coloration, and size of the dorsal and
anal fins. There appears to be no divergence in body proportions between the sexes

(A)

1 2

(B)

2 3

FIG. 18. Homologous areas in coloration of dorsal fins in mature male (A) and female
(B) of L. orca. H, hyaline fin membrane ; G, guanophores ; E, erythrophores ; M,
melanophores ; 1-4, comparable bands of erythrophores in second dorsal fin. Increasing
admixture of melanophores with erythrophores shown by closer cross-hatching.

and both reach about the same maximum standard length. Detailed accounts of the
coloration of mature individuals in each sex have been given above. Differences are
most pronounced in the two dorsal fins, but pelvic, pectoral, and anal fins are all
darker in the male than in the female. The components of the dorsal fin patterns are
basically the same in the two sexes, and the considerable disparity between adult
males and females results more from differential emphasis and spread of existing
parts than from the appearance of new elements as maturity is approached. Homo-
logous parts of the dorsal fin coloration are shown diagrammatically in Text-fig. 18.
On reaching the size of maturity, dimorphism in fin size is brought about by

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 239

changes in the growth of individual rays and adjacent membrane (Text-figs. 19, 20).
Growth of the second ray (II) of the first dorsal fin is about the same in the two
sexes, but differences in length become more apparent in the other first and second
dorsal rays investigated, especially the posterior ones of the first dorsal fin, these rays
attaining a greater length in males than in females. In the male, the first dorsal
membranes have considerable elasticity, and an enormous spread is achieved at full
extension of the fin (PI. I). Growth of the anal fin was not studied in detail, but a
divergence in size at sexual maturity is indicated in the systematic description above.

Sexual dimorphism in teleosts has recently been reviewed by Bertin (1958). The
phenomenon is of widespread occurrence among the gobies and has been investigated
in Japanese species by Egami (1960). This worker lists a number of species in which
the first dorsal rays are more elongate in the male, but does not mention any instances
of sexual differences in coloration of the dorsal fins. In searching for dimorphism in
ray length, Egami compares the lengths of only the longest ray in each fin. With the
first dorsal fin of Lebetus, where such dimorphism in fin size is most conspicuous, this
method of comparison is inapplicable since the longest ray in this fin reaches a similar
length in both sexes and by itself would fail to indicate the great disparity between
male and female in growth of the shorter first dorsal rays. The secondary sexual
characters in Lebetus are probably involved in the release of certain sequences in
reproductive behaviour. The enlarged dorsal fins of the male, with the contrast of
bright colours disposed in an arrangement peculiar to the species, have all the
characters of sign stimuli. Spawning activities in several European gobiids, where
sexual dimorphism in coloration and shape of the dorsal fins may occur, are known
to involve display of these fins by the male in fighting and courtship (Kinzer, 1960).

LIFE HISTORY AND GROWTH. Scale readings from Manx Lebetus have been inter-
preted in the light of unpublished data concerning scale growth in the goby Pomato-
schistus microps (Kr0yer). Results indicate that Lebetus may live for at least two years.
In some individuals, sexual maturity may be attained at the age of about one year,
while in others this stage is not reached until after the second winter of life. Insuf-
ficient material exists for studies on growth rate ; the largest specimen collected, a
male, was 30-0 mm. in standard length (total 37-5 mm.) and in the second autumn
of life. The largest female, 29-0 + 7-0 mm., was probably nearing the end of its
second year. The maximum size recorded for the genus is 39-0 mm. (Lonnberg &
Gustafson, 1935). Among the Manx collection, the smallest mature male and female
individuals were 22-0 and 23-75 mm. in standard length, but Holt & Byrne (1898)
obtained a ripe female of only 21-0 mm. total length. Further investigation of the
life history of this species must depend on a more satisfactory means of sampling,
since the scallop dredge, even when lined with fine meshed netting, appears to be
highly selective of older and larger fishes.

DISCUSSION

The percomorph suborder Gobioidea was defined in osteological terms by Regan
(1911). Since then, the skeletons of individual species have been treated by Gregory
(1933), Te Winkel (1935), Lele & Kulkarni (1938, 1939), Petit (1941), Smith (1951),
and Matsubara & Iwai (1959), while Gosline (1955) has recently compared those of

240

P. J. MILLER

FIRST DORSAL

o oo

t *

0^3

*> OOO

too

o o

0> 0>

O* COD O

o * o o o

0*0000

o
^ o o

* O

^ 2-

.c 1

"5)

c
_l

FIRST DORSAL

.E 8-

'" I

IV J

. \I

o> o

O O 00

t o

o o o o

o o o

000

O O_p O OOo

o oo o

o o

00 O

15 20 25 30 15 20 25 30

Standard Length (mm)

FIG. 19. Relation between standard length and lengths of first four rays of first dorsal
fin (I-IV) in L. orca. , male ; O, female ; O, coincident points.

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS

241

8-
7-

6-
5-
4-
3-
2-

FIRST DORSAL

o o o

CO OO O O
i O OO OO

o oo o o o o
o o

FIRST DORSAL
VI

00 O O

oo coco o oo

o
o ooo o o

o o

o
ac.

.E 8-

7-
6-

5-
4-
3-
2-

SECOND DORSAL

0> CD
O
O CB O

OO O O

o o

O oq
o o e o

O OCX OOO

1 | - 1 | - 1

1 I - 1 I

1 I 1 - 1 1 1

25 30

I I I I I I
15 20

25 30

Standard Length (mm)

FIG. 20. Relation between standard length and length of last two rays of first dorsal fin
(V, VI) and first two rays of second dorsal fin (I, i) in L. orca. Symbols as in Fig. 19.

242 P. J. MILLER

eight diverse genera of gobies. The group is characterized by several peculiarities.
Among these are the large non-osseous area between the symplectic, quadrate, and
preoperculum ; the arrangement of the branchiostegals on the ceratohyal ; the absence
of the parietal bones ; and the presence of a splintlike radial above and below the
hypurals in the caudal skeleton. In all these features, the genus Lebetus appears as a
typical gobioid fish, but there has been some doubt about the position of Lebetus among
the families of this suborder. Of these about six are now recognized (Koumans, 1953 ;
Gosline, 1955 ; Norman, MS.). The Rhyacichthyidae, Taenioididae, Kraemeriidae, and
Microdesmidae are highly aberrant gobioids and need not be further dealt with in this
discussion. The two largest and most important families are the Eleotridae and
Gobiidae. Skeletal differences between these have been summarized by Regan (1911),
and the diagnostic value of certain features reviewed by Gosline (1955). Nevertheless,
in referring gobies to one or other of these families, it has been customary to use the
external form of the pelvic fins as the chief guide in the usual absence of osteological
data. According to Regan (1911) and other authors, the Eleotridae are characterized
by the pelvic fins being separate, and the Gobiidae, when these fins are developed, by
their union to produce a simple cup-shaped disc. However, various stages in the
union or separation of the pelvic fins have been recognized among gobioid genera and
the importance of this feature as a criterion of family position has recently been
questioned, especially by Bohlke & Robins (1960^, b) who have found eleotrid, gobiid,
and intermediate types of pelvic fins in a single genus (Coryphopterus Gill). In such
cases as these, determination of systematic position within the suborder requires
investigation of the skeleton.

When erecting the genus Lebetus, Winther (1877) emphasized that the anterior
membrane of the pelvic disc was missing, although the two fins were joined along their
entire length in the midline. At first accepting (1885) the new genus, Collett (1896)
was later of the opinion that intermediate stages in reduction of the anterior mem-
brane could be found between the fully developed condition in Gobius and complete
loss as reported for Lebetus, and doubted the validity of separating the latter from
Gobius. On the other hand, Duncker (1928) found that the two pelvic fins were
completely separated in Lebetus but stressed the desirability for re-examination of
these fins in better preserved material in order to decide whether Lebetus belonged
to the Eleotridae or the Gobiidae. De Buen (19306) associated Lebetus with two
Mediterranean genera, one of which had previously been regarded as an eleotrid
because of its almost completely separated pelvic fins, and placed this " Lebetus "
group in the subfamily Gobiinae of the Gobiidae. Koumans (1931) similarly accepted
Lebetus as a gobiid genus. In the present work, it has been confirmed that, in the least
damaged examples, the pelvic fins of Lebetus are completely united between their
origins and the distal extremity of the fifth ray of each fin, but are entirely without
trace of an anterior membrane to complete the disc. To settle the question of family
position with greater certainty, it became obvious that examination of the skeleton
was necessary, and the results of this study reveal that Lebetus has the characters of
a typical member of the Gobiidae. These include a T-shaped head to the palatine,
loss of the mesopterygoid from the palatopterygoid arch, and absence of a hyper-
coracoid in the pectoral girdle, where the laminar radials are inserted on the cleithrum

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 243

and the lowest also related to the hypocoracoid. In the possession of a spatulate
glossohyal, five branchiostegous rays, a reduced metapterygoid, and loss of the
opisthotic, Lebetus also reaches a higher level of specialization than that described
among the eleotrids (Regan, 1911 ; Takagi, 1950 ; Gosline, 1955).

Within the Gobiidae, the affinities of Lebetus undoubtedly lie with the subfamily
Gobiinae, since the genus exhibits none of the more extreme modifications of the
Sicydinae, Periophthalminae, Apocrypteinae, Tridentigerinae, Gobiodontinae, or
Benthophilinae (Koumans, 1953 ; Norman, MS.). In the absence of an anterior
membrane to the pelvic disc, Lebetus resembles a number of small Indo-Pacific
genera i.e. Herreogobius Koumans, Quisquilius Jordan & Everman, Fusigobius
Whitley, Amblyogobius Bleeker, and Zonogobius Bleeker (Koumans, 1953 ; Gosline,
1959) . With head and anterior part of body naked, elongate rays in the first dorsal
fin, 22-30 scales in lateral series, and radial formula of D 2 1/8-9, A I/7~8 (Koumans,
I 953)> Zonogobius appears to be closest of these to Lebetus. Differences comprise
the larger gill openings, and more laterally compressed head in Zonogobius. However,
it is impossible to trace relationships between European and exotic genera, when the
arrangement of the lateral line system in so many of the latter, including Zonogobius,
remains to be adequately described and figured. As shown in the normal illustration
of Zonogobius corallinus sp. nov. (Mozambique) by Smith (1959, Fig. 29), the dis-
tribution of sensory papillae on cheek and postorbital regions differs from that of
Lebetus in the complete longitudinal rows and more numerous papillae. Koumans
(1931) associated Lebetus with the genus Coryphopterus Gill ; as defined by Koumans,
the latter included European Pomatoschistus species together with the genotype, C.
glaucofraenum Gill, from the tropical Western Atlantic. In a revision of Coryphop-
terus sensu stricto, Bohlke & Robins (19606) provide details of the cephalic lateral
line system and other characters which do not point to any close affinity with Lebetus,
in spite of the fact that two of their new Western Atlantic species (C. alloides and C.
dicrus) are without the anterior pelvic membrane.

In the European fauna there are two Mediterranean genera of gobies lacking
anterior membranes to the pelvic fins. These are Odondebuenia and Cabotia, both
introduced by De Buen (19306). Since Cabotia is preoccupied among the Lepidoptera,
De Buen (i5th June, 1940) suggested Fagea as a replacement but was preceded in this
by Whitley (May, 1940) who introduced Cabotichthys. The genus Odondebuenia
contains two species which were originally (1907) placed among the Eleotridae as
Eleotris balearicus Pellegrin & Fage and E. pruvoti Fage. Odondebuenia is character-
ized by pelvic fins united only at their bases, naked nape and throat, unusual gill
rakers (Fage, 1907, figs. 3 and 9), modified scales at the origin of the caudal fin, and
meristic characters of D 2 1/9-11, A I/8-io, and Sc. 1.1. 24-32. Both species of
Odondebuenia show sexual dimorphism in length of first dorsal fin rays (Fage, 1918 ;
De Buen, 19306). The second of these Mediterranean genera, Cabotichthys, has the
pelvics joined together for about half their length, nape and throat fully scaled, scales
of the caudal peduncle not modified, and fin ray and scale counts of D 2 1/14, A 1/13,
and Sc. 1.1.50. The genus is founded on one species, C. schmidti (De Buen) known
only from the single type specimen (De Buen, 19306). De Buen (1931) included
Odondebuenia, Cabotichthys and Lebetus in his " Lebetus " group of genera, the principal

244 P. J- MILLER

character of which is the absence of the anterior pelvic membrane. Similarities in habi-
tat also link these fishes, all three being obtained on coarse grounds where calcareous
algae are noticeable constituents of the bottom deposits, and none exceed 50 mm. in
length.

On the face of the diagnoses given above, Lebetus would appear to be fairly closely
related to Odondebuenia except for the occurrence in the latter of modified caudal
scales and gill rakers, and the greater separation of the pelvic fins, while Cabotichthys
stands quite distant from both in several respects. The lateral line system in Odonde-
buenia and Cabotichthys has fortunately been investigated by De Buen (19306, 1940)
and that of Lebetus is described above. Comparison of these accounts for the three
genera reveals the heterogenous nature of their grouping. The patterns of sensory
papillae and extent of the cephalic canals in Cabotichthys point to an intimate connec-
tion with the genus Gobius, as defined by De Buen (19300, 1931), and, apart from the
reputed form of the pelvics, other characters of this fish do not warrant any generic
separation from Gobius. My own experience with dredged and trawled examples of
Lebetus and other gobies is that the pelvic membranes are very susceptible to damage,
and the pelvics of Cabotichthys, as figured by De Buen (19306, fig. 7) are reminiscent
of such a condition. It may well prove that C. schmidti is no more than a damaged
specimen of a Gobius species.

The state of the pelvics in the Odondebuenia species seems well established from a
number of specimens. In the arrangement of sensory papillae, this genus differs
markedly from Lebetus in the greater number of papillae, and their distribution in
well marked transverse and longitudinal rows on the cheek approaches that in Gobius,
which Odondebuenia also resembles in the possession of a similar cephalic canal
system. The morphological attributes common to both Odondebuenia and Lebetus
are not of a very highly specialized nature and do not outweigh the considerable
disparity in patterns of sensory papillae. Modification of scales at the base of the
caudal fin involves pronounced elongation of all the ctenii in 0. pruvoti (Fage, 1907,
fig. 7) or merely the lateral ctenii in 0. balearica (Fage, 1907, fig. n ; De Buen, 19306,
fig. 2). Although not seen in Lebetus, it is of interest to note the occurrence of this
peculiarity in the new gobioid genus Various, recently described from the West
Indies by Robins & Bohlke (1961) and also with separated pelvic fins. These authors
commented on the existence of comparable scales in the Californian eleotrid Chriolepis
Gilbert and the gobiid Garmannia Jordan, which is represented on both sides of
Central America and in the Caribbean. It is not proposed at this point to deal
further with the relationships of Odondebuenia, except to call attention to the need
for an osteological investigation of this genus.

Loss of the anterior membrane from the pelvic disc has probably occurred on a
number of different evolutionary lines among the Gobiidae and, as in the case of
Odondebuenia and Lebetus, need not be a sure indication of close phyletic association.
In geographical distribution Lebetus appears to be confined to the eastern Atlantic
boreal region and may well have evolved there. It is in this area, therefore, that the
closest relatives of the genus may be sought. Eight other gobioid genera are recorded
from the eastern North Atlantic (De Buen, 1931) and all of these have a pelvic disc
complete with anterior membrane. The two pelagic forms Aphya Risso and Crystal-

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 245

logobius Gill may be excluded from consideration on the grounds of extreme specializa-
tion. Among the demersal gobies, the genera Deltentostem Gill, Pomatoschistus Gill,
Chaparrudo De Buen, and Gobius L. exhibit various features of anatomy and lateral
line organization (see De Buen, 1930^, 1931) which do not suggest any near relation-
ship with Lebetus. The remaining genera, Lesueurigobius Whitley and Buenia Iljin,
deserve more attention. The first of these, represented by Lesueurigobius friesii
friesii (Collett), has a number of characters in common with Lebetus. Thus both
genera are entirely without cephalic canals and in disposition of sensory papillae
show a posterior extension of row a behind the eye, papillae in the oculoscapular
furrow (row u), interorbital papillae, a high posterior termination for row i, and reduc-
tion of rows m, n, and o (Sanzo, 1911, pi. 9, figs. 4, 5, as Gobius macrolepis ; De Buen,
1923, figs. 21, 22). However, the great abundance of papillae in Lesueurigobius
contrasts with their relative scarcity in Lebetus, where several rows (g, x, z, c 2 , b, d)
found in Lesueurigobius are little if at all developed, and others (i, e, ot, c-cp] inter-
rupted. Except for reduction in rows m, n, and o, the resemblances listed may be
accounted for by independent action of a similar evolutionary process in the past
history of each genus. This was the replacement of existing cephalic canals by rows
of sensory papillae. Considerable anatomical differences in size, habit, squamation,
radial formulae, and coloration (Holt & Byrne, 1903 ; Duncker, 1928) would also
suggest similarities are due to convergence, and that any affinity is relatively distant.
No closer relationship can be demonstrated with Lesueurigobius sanzoi (De Buen) or
L. lesueuri (Risso) from the Mediterranean (De Buen, 1923).

The only genus now remaining is Buenia, the Atlantic species of which is B. Jeffrey sii
(Giinther), and it is with this form that Lebetus appears at present to be most closely
connected. Lacking scales on the head and predorsal regions, and with meristic
characters of D 2 1/8-9, A I/y-8, and Sc. 1.1.25-30 (Duncker, 1928 ; personal
observation), B. Jeffrey sii displays a resemblance to Lebetus which is also evident in
the arrangement of the lateral line system (Text-fig. 21). On the cheek absence of
transverse rows of papillae, and the short row of large papillae with an intermediate
papilla between this and row a, recall the conditions in Lebetus. The lateral preorbital
rows c 1 and c 2 are identical in both genera. On the dorsum of the head there is a
correspondence in the reduced number and the arrangement of the papillae. Lebetus
differs from B. Jeffrey sii in the disappearance of cephalic canals and their replacement
by papillae of rows a 1 , u l , i 1 , and the anterior part of n, together with a certain
" condensation " of other rows (d 1 , preoperculo-mandibular, opercular, and other
series) and loss of rows d and b. The pattern of sensory papillae in Lebetus is obviously
more specialized than that in Buenia, but the features present in both denote a
greater affinity between these two gobies than is exhibited between Lebetus and other
genera. In the persistence of cephalic canals and various rows of papillae, Buenia
displays more ostensibly primitive characters than Lebetus. The former in addition
retains the anterior pelvic membrane, is not so small as Lebetus, and does not show
pronounced sexual dimorphism, although this does occur in growth of the first dorsal
fin rays and probably in coloration of this fin. In the male genitalia, the testes are
long and there is no free sperm duct. B. Jeffrey sii may accordingly show greater
resemblance to the common stock from which the two are derived. The Mediterranean

246

P. J. MILLER

species B. reticulatus (C.&V.) (=B. affinis (Kolombatovic)) is more removed from
Lebetus in having a greater number of papillae especially in rows b and d, and a
somewhat higher scale count (more than 33), than B. jeffreysii (Sanzo, 1911, pi. 9,
figs. 9, 10 ; De Buen, 19300). In Manx waters B. jeffreysii has an offshore distribution

FIG. 21. Sensory papillae and orifices of cephalic canals in Buenia jeffreysii, male, 31
mm. standard length. (A) Lateral, (B) dorsal views of head. Abbreviations as in Fig. 7.
Orifices of canals cross-hatched, with greek lettering as in Sanzo (1911).

like Lebetus and has even been taken from the same localities (see also Le Danois,
1910), but differences in precise habitat are perhaps reflected in the different body
form and coloration (Holt & Byrne, 1903, fig. 4).

Some of the characters of Lebetus may be interpreted as adaptations to a mode of life
involving intimate association with dead shells, nodules of calcareous algae, and
stones. The small size of the fish is obviously related to this type of habitat, and the
normal coloration may have cryptic value against a background of Lithothamnion.
Te Winkel (1935) has discussed those anatomical features of the gobiid Mistichthys

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 247

luzonensis which may be correlated with the extremely small size of this species at
maturity (standard length 10-0-11-0 mm.). Although Lebetus is diminutive, it is
appreciably larger than Mistichthys and the only obvious comparable modification
in this genus is the reduced size of the testis, to which may be related the form of the
urogenital papilla as considered above. A more detailed examination of the viscera
than was attempted in the present study would be needed to investigate this question.
Reduction in number of sensory papillae has already been mentioned, and may be
linked with small body size (Barlow, 19616). The significance of loss of the cephalic
canals is unknown ; these are also lacking in Lesueurigobius, which attains 100 mm.
in L. friesii, but are present in smaller forms such as Odondebuenia, Buenia, etc.
Another feature of probable adaptive importance is the absence of the anterior
pelvic membrane, since this is also missing in Odondebuenia which appears to have a
similar habitat and which in addition exhibits almost complete separation of the pelvic
fins. While the exact affinities of Odondebuenia await investigation, it seems likely that
in view of the resemblances in arrangement of sensory papillae between this genus and
Gobius the condition of the pelvics is derived from an original gobiid disc structure in
response to environmental demands, and is not primitive as in the Eleotrids. The
advantages conferred by the alteration of the pelvic disc in Lebetus and Odondebuenia
are not obvious. The loss of the anterior membrane seems to have occurred inde-
pendantly in the evolution of the two genera as the exploitation of similar ecological
niches proceeded in two areas from different stocks.

It must be stressed that the above phylogenetic conclusions may at present be
regarded as tentative. Not only does the position of Odondebuenia remain doubtful,
but the alignment of Lebetus with Buenia may well be affected by future work on the
lateral line system of exotic gobies and faunistic exploration in little known areas,
which together could disclose closer resemblances and intermediate forms between
these and Lebetus.

SUMMARY

The teleostean genus Lebetus Winther 1877 (Percomorphi-Gobioidea) has been
studied from new Irish Sea material dredged off the south of the Isle of Man, and
from other examples, including types, in the collections of various British and Scandi-
navian museums. A redefinition of the genus includes the disposition of sensory
papillae ; there is a reduction in number of the latter, and cephalic canals and an
anterior membrane to the united pelvic fins are lacking. Two species were previously
recognized : orca Collett 1874 and scorpioides Collett 1874. Among the various
criteria coloration, meristic characters, body proportions used in the past to
distinguish these two species, it was found that only in coloration and development of
the dorsal fins could the present Lebetus material be divided into two groups, termed
orca and scorpioides. The former consisted entirely of males, nearly all maturing or
mature, the latter of females and immature males. The conclusion is reached that
there is only one valid species, by page priority Lebetus orca. A detailed account of
the external anatomy and osteology of this species is provided.

In distribution L. orca is restricted to the European Atlantic boreal region, and a
full list of records is provided in an appendix. The species is known chiefly from

248 P. J. MILLER

coarse grounds, and has a wide bathymetric range from about 2 to 375 m. Investiga-
tion of gut contents, using a points system of assessment, shows that L. orca is
exclusively predatory, feeding on small demersal animals which largely comprise
Crustacea and polychaetes. A number of endoparasites are listed. The male repro-
ductive organs are peculiar in the relatively small size of the testis and the long free
sperm duct ; " seminal vesicles " are present. Sexual differences exist in the form
of the urogenital papilla, which in mature males has an unusual terminal circlet of
vascularized papillae. In Manx waters, the breeding season probably commences in
March and may extend to August. 140 to 270 ripening oocytes have been counted in
mature females. Sexual dimorphism is found in coloration, and size of dorsal and
anal fins. Examination of scales indicates that duration of life may be at least two
years, with sexual maturation at one or two years. Maximum total length recorded
is 39-0 mm. Investigation of life history and growth in the Manx specimens was
complicated by the high selectivity of the sampling gear employed.

In a discussion of the systematic position of Lebetus, it is shown that the skeleton is
typically gobiid, and the arrangement of sensory papillae together with other features
suggest that the nearest relative of this genus among the European gobies is Buenia
Jeffrey sii (Giinther). Previous grouping of Lebetus with two Mediterranean genera
lacking an anterior pelvic membrane (Odondebuenia De Buen and Cabotichthys
Whitley) is criticized. The validity of Cabotichthys is doubted, and the suggestion is
made that loss of the anterior membrane in Lebetus and Odondebuenia occurred
independently during their evolution and occupation of a similar habitat in different
areas.

ACKNOWLEDGEMENTS

I wish to thank Messrs J. S. Colman and A. B. Bowers, Marine Biological Station,
Port Erin, and Mr. A. C. Wheeler, British Museum (Natural History), for their
helpful criticism of this paper. My thanks are also due to Mr. D. Eggleston, Dr. R. G.
Hartnoll and Dr. M. C. Miller for collecting specimens of Lebetus while dredging off
the Isle of Man. For arranging the loan of other material, I am most grateful to Mr.
E. K. Barth, Zoologisk Museum, Oslo ; Dr. E. Bertelsen, Carlsbergfondets Dana-
Ekspeditioner, Charlottenlund Slot ; Mr. J. Nielsen, Universitetets Zoologiske
Museum, Copenhagen ; Mr. A. C. Wheeler, British Museum (Natural History) ;
and Mr. J. F. Willgohs, Zoologisk Museum, Bergen. Professor H. Brattstrom,
Biological Station, Espegrend ; Dr. H. O. Bull, Dove Marine Laboratory, Cullercoats;
Mr. B. Christiansen, Zoologisk Avdeling, Tromso Museum ; Dr. R. S. Clarke, Royal
Scottish Museum, Edinburgh ; Mr. S. Mathiasson, Naturhistoriska Museet, Goteborg;
Dr. B. B. Rae, Marine Laboratory, Aberdeen ; Dr. B. Swedmark, Kristinebergs
Zoologiska Station, Fiskebackskil ; and Dr. A. E. J. Went, Fisheries Division of Dept.
of Lands, Dublin, have kindly provided information on distribution or material
present in collections under their care. Messrs J. W. Coles and S. Prudhoe, British
Museum (Natural History), identified nematodes and platyhelminthes encountered in
this study, while Dr. Eve C. Southward, Plymouth Laboratory, and Drs. N. S. Jones
and D. I. Williamson, Port Erin Marine Biological Station, identified polychaetes,
molluscs and arthropods among the gut contents. I am also most grateful to Mr. C. A.

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 249

Grainge, Zoology Dept., Liverpool University, for preparation of sections, and to
Messrs P. J. Evennett, Zoology Dept., Leeds University, and D. J. Slinn, Port Erin
Marine Biological Station, for photographic assistance. Most of the above investiga-
tion was carried out during the tenure of a D.S.I.R. Research Studentship at the
Marine Biological Station, Port Erin, and has been completed in the Zoology Dept.,
Glasgow University. I am therefore indebted to Mr. J. S. Colman and Professor
C. M. Yonge, F.R.S., for provision of research facilities at Port Erin and Glasgow
respectively.

REFERENCES

AFLALO, F. G. 1904. British Salt-water Fishes. 328 pp. London.

ANDERSSON, K. A. 1942. Fiskar och Fiske i Norden. 1, 540 pp. Stockholm.

BAL, D. V. 19400. A Study of Fish Eggs and Larvae from Manx Waters. 126 pp. Ph.D. Thesis,

Liverpool University.

19406. Some recent additions of fish eggs and larvae to the fauna of Port Erin. Rep. mar.

biol. Sta. Port Erin, No. 53 : 14-17.
BARLOW, G. W. 19610. Causes and significance of morphological variation in fishes. Syst.

Zool., 10: 105-117.
19616. Gobies of the genus Gillichthys, with comments on the sensory canals as a taxonomic

tool. Copeia 1961, No. 4 : 423-437

BERTIN, L. 1958. Sexualit6 et f6condation. Traite de Zoologie, 13 : 1584-1652.
BOHLKE, J. E. & ROBINS, C. R. 19600. Western Atlantic gobioid fishes of the genus Lythrypnus,

with notes on Quisquilius hipoliti and Garmannia pallens. Proc. Acad. nat. Sci. Philadelphia,

112 : 73-101.

19606. A revision of the gobioid fish genus Coryphopterus. Ibid., 112 : 103-128.

BRUUN, A. F. & PFAFF, J. R. 1950. Fishes. List of Danish Vertebrates, 19-60.
CANTACUZENE, A. 1956. Inventaire de la faune marine de Roscoff. Poissons. Trav. Sta. biol.

Roscoff, Supplement 8, 67 pp.
CLARK, R. S. 1914. General report on the larval and post-larval teleosteans in Plymouth

waters. /. mar. biol. Ass. U.K., 10 : 327-394.

1920. The pelagic young and early bottom stages of teleosteans. Ibid., 12 : 159-240.

COLLETT, R. 1874. On two apparently new species of Gobius from Norway. Ann. Mag. nat.

Hist. (4), 13 : 446447.

- 18750. Bidrag til Kundskaben om Norges Gobier. Forh. Vid. Selsk. Christiania 1874 :
151-179.

18756. Norges Fiske, med Bemaerkninger om deres Udbredelse. 240 pp. Christiania.

- 1885. Meddelelser om Norges Fiske i aarene 1879-83. Nyt. Mag. Naturv. 29 : 47-123.

- 1896. Poissons provenant des campagnes du yacht 1'Hirondelle (1885-1888). Res. Camp,
sci. Monaco, Fasc. 10, 198 pp.

1902. Meddelelser om Norges Fiske i aarene 1884-1901. I. Forh. Vid. Selsk. Christiania

1902, No. i, 121 pp.
COUJARD, R. 1941. Sur 1'existence d'une glande testiculaire et d'une glande gnitale annexe

chez les gobies. C.R. Soc. biol. Paris, 135 : 570-574.
CRAWSHAY, L. R. 1912. On the fauna of the outer western area of the English Channel.

/. mar. biol. Ass. U.K., 9 : 292-393.
DE BUEN, F. 1923. Gobius de la Peninsula Iberica y Baleares. Grupos Lesueurii, Colonianus,

Affinis, y Minutus. Mem. Inst. Espanol Oceanogr., 3 : 121-266.

- 19300. Sur une collection de Gobiinae provenant du Maroc. Essai de synopsis des especes
de 1'Europe. Bull. Soc. Sci. nat. Maroc. 10 : 120-147.

19306. Lebetus Winther 1877, Odondebuenia nov. gen. y Cabotia nov. gen. Trab. Inst.

Espanol Oceanogr., No. 5, 30 pp.

250 P. J. MILLER

DE BUEN, F. 1931. NotasalafamiliaGobiidae. Observaciones sobre algunos generos y sinopsis de
las especies ib6ricas. Not. Res. Inst. Espanol Oceanogr. (2), No. 54, 76 pp.

1932. Lebetus scorpioides. Faun, ichthyol. I'Atlantique Nord, Plate 300.

- 1940. Les Gobiidae p61agiques ou vivant sur les fonds d'algues calcaires de 1'Europe
Occidentale. Bull. Inst. oceanogr. Monaco, No. 790, 16 pp.

1951. Contribuciones a la ictiologia. V-VI. Sobre algunas especies de Gobiidae de la

coleccion del Laboratorio Arago (Banyuls-sur-Mer, Francia) y description de un nuevo
genero (Austrogobius) sudamericano. Bol. Inst. Paul. Oceanogr., 2 : 55-69.

DOTU, Y. 1955. On the life history of a gobioid fish, Eutaeniichthys gilli Jordan et Snyder.
Bull, biogeogr. Soc. Japan, 16-19 : 338-344.

- I 957# The bionomics and life history of a gobioid fish, Paleatogobius uchidai Takagi.
Jap. J. Ichthyol., 6 : 97-104.

- I957&- The life history of the small, transparent goby, Gobius lidwilli McCulloch. Sci.
Bull. Fac. Agric. Kyushu Univ., 16 : 85-92.

- i958a. The bionomics and life history of two gobioid fishes, Tridentiger undicervicus
Tomiyama and Tridentiger trigonocephalus (Gill) in the innermost part of Ariake Sound.
Ibid., 16 : 343-358.

19586. The life history of the gobioid fish, Acentrogobius masago (Tomiyama). Ibid., 16 :
359-37-

D6ru, Y. & MITO, S. 1958. The bionomics and life history of the gobioid fish, Luciogobius

saikaiensis D6tu. Ibid., 16 : 419-426.

DUNCKER, G. 1928. Teleostei Physoclisti 9. Gobiiformes. Tierw. N.-u. Ostsee, 12 g : 121-148.
EGAMI, N. 1960. Comparative morphology of the sex characters in several species of Japanese

gobies, with reference to the effects of sex steroids on the characters. /. Fac. Sci. Univ.

Tokyo, Section 4,9: 67-100.
EGGERT, B. 1931. Die Geschlechtsorgane der Gobiiformes und Blenniiformes. Z. wiss. Zool.,

139 : 249-517.
EHRENBAUM, E. 1936. Naturgeschichte und wirtschaftliche Bedeutung der Seefische Nord-

europas. Handbuch der Seefischerei Nordeuropas, 2, 337 pp.
EHRENBAUM, E., KYLE, H. M., SCHNAKENBECK, W., DUNCKER, G., & MOHR, E. W. 1929.

Pisces (Nachtrage und Berichtigungen) . Tierw. N.-u. Ostsee, 12 h : 141-164.
EKMAN, S. 1953. Zoogeography of the Sea. 417 pp. London.

ELMHIRST, R. 1926. Notes on fishes from the Firth of Clyde. Scot. Nat. 1926 : 151-158.
FAGE, L. 1907. Essai sur la faune des poissons des iles Bal6ares et description de quelques

especes nouvelles. Arch. Zool. exp. gen. (4), 7 : 69-93.

1918. Shore-fishes. Rep. Dan. oceanogr. Exped. Medit. 1908-1910, 2, A3, 154 pp.
FARRAN, G. P. 1912. A biological survey of Clare Island in the county of Mayo, Ireland and

of the adjoining district. Part 19. Pisces. Proc. roy. Irish Acad., 31, 6 pp.
FILIPPI, F. 1862. Note zoologiche. Arch. Zool. Anat. Fisiol., 1 : 52-76.
GOSLINE, W. A. 1955. The osteology and relationships of certain gobioid fishes, with particular

reference to the genera Kraemeria and Microdesmus. Pacific Sci., 9 : 158-170.

1959- Four new species, a new genus, and a new suborder of Hawaiian fishes. Ibid., 13 :
67-77.

GREGORY, W. K. 1933. Fish skulls : a study of the evolution of natural mechanisms. Trans.

Amer. phil. Soc. Philadelphia, 23 : 75-481.
GRIEG, J. A. 1899. Ichthyologiske Notiser, II. Bergens Mus. Aarb. 1898, No. 3, 23 pp.

1913. Bidrag til Kundskapen om Hardangerfjordens Fauna. Ibid., 1913, No. i, 147 pp.

GUNTHER, A. 1867. Additions to the British fauna. Ann. Mag. nat. Hist. (3), 20 : 288-291.

1888. Report on the fishes obtained by Mr. J. Murray in deep water on the north west
coast of Scotland, between April 1887 and March 1888. Proc. roy. Soc. Edinburgh, 15 :
205-220.

HARTNOLL, R. G. 1961. Studies on the Biology of Manx Spider Crabs. 177 pp. Ph.D. Thesis,
Liverpool University.

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 251

HOLT, E. W. & BYRNE, L. W. 1898. Notes on the reproduction of teleostean fishes in the

south-western district. /. mar. biol. Ass. U.K., 5 : 333-340.
1903- The British and Irish gobies. Rep. Sea Inl. Fish. Ireland 1901, Part 2 : 37-66.

1910. Fourth report on the fishes of the Irish Atlantic slope. List of recorded species,

with references. Sci. Inv. Fish. Ireland 1908, No. 5, 7 pp.

HYNES, H. B. N. 1950. The food of fresh-water sticklebacks (Gasterosteus aculeatus and Pygo-

steus pungitius), with a review of methods used in studies of the food of fishes. J. anim Ecol

19 : 36-58.

ILJIN, B. S. 1930. Le systeme des Gobiides. Trab. Inst. Espanol Oceanogr., No. 2, 63 pp.
JENKINS, J. T. 1925. The Fishes of the British Isles. 376 pp. London. (2nd Edition, 1936,

408 pp.)
JOHANSEN, A. C. 1925. On the diurnal vertical movements of young of some fishes in Danish

waters. Med. Komm. Hav. Fisk., 8, No. 2, 26 pp.

JONES, N. S. 1949. Faunistic notes. Ann. Rep. mar. biol. Sta. Port Erin, No. 61 : 27-28.
1950. Records of the Offshore Fauna of the South of the Isle of Man. MS., Port Erin Marine

Biological Station Library.

1951. The bottom fauna off the south of the Isle of Man. /. anim. Ecol., 20 : 132-144.

JORDAN, D. S. 1919. The Genera of Fishes. Part III. 285-410 pp. Stanford.
1920. Ibid., Part IV. 412-576 pp. Stanford.

1923. A classification of fishes including families and genera as far as known. Stanford

Univ. Pub. Biol. Sci., 3 : 79-243.

KINZER, J. 1960. Zur Ethologie und Biologic der Gobiiden, unter besonderer Beriicksich-

tigung der Schwarzgrundel Gobius jozo L. Zoo/. Beitr., 6 : 207-290.
KOUMANS, F. P. 1931. A Preliminary Revision of the Genera of the Gobioid Fishes with United

Ventral Fins. 174 pp. Lisse.

I 953- Gobioidea. Fishes of the 1 ndo- Australian Archipelago, 10, 423 pp.

LE DANOIS, E. 1910. Sur la faune ichthyologique du mae'rl. Bull. Soc. zool. France, 35 : 167-

171.

1913. Contribution a 1'etude systematique et biologique des poissons de la Manche

Occidentale. Ann. Inst. Oceanogr. Monaco, 5, Fasc. 5, 214 pp.

1948. Les Profondeurs de la Mer. 303 pp. Paris.

I953- La Vie Etrange des Rivages Marins. 191 pp. Paris.

LELE, S. H. & KULKARNI, R. D. 1938. The skeleton of Periophthalmus barbarus (Linn.).

(i) The skull. /. Univ. Bombay, 6 : 76-91.
1939. The skeleton of Periophthalmus barbarus (Linn.). II. Branchial arches, vertebral

column and appendicular skeleton. Ibid., 7 : 123-134.

LILLJEBORG, W. 1884. Sveriges och Norges Fiskar. Vol. i, Pt. 3 : 497-782. Uppsala.
LONNBERG, E. & GUSTAFSON, G. 1935- Fran Kristinebergs zoologiska station sommaren 1935.

Fauna och Flora, Uppsala 1935 : 241-247.

MALM, A. W. 1877. Goteborgs och Bohusldns Fauna. Ryggradsdjuren. 674 pp. Goteborg.
MARINE BIOLOGICAL ASSOCIATION. 1931. Plymouth Marine Fauna. 2nd Ed. 371 pp. Ply-
mouth. (3rd Ed., 457 pp., I957-)
MATSUBARA, K. & IWAI, T. 1959. Description of a new sandfish, Kraemeria sexradiata, from

Japan, with special reference to its osteology. /. Wash. Acad. Sci., 49 : 27-32.
MILLER, P. J. 19610. Age, growth, and reproduction of the Rock Goby, Gobius paganellus L.,

in the Isle of Man. /. mar. biol. Ass. U.K., 41 : 737-769.

19616. The species of the teleostean genus Lebetus. Nature, Land., 192 : 675-676.

NORMAN, J. R. 1935. Fishes. List of British Vertebrates, 40-66.

MS. A Draft Synopsis of the Orders, Families and Genera of Recent Fishes and Fish-Like

Vertebrates. 649 pp. British Museum (Natural History).

OTTERSTR0M, C. V. 1912. Fisk. I Pigfmnefisk. Danmarks Fauna, 6, 198 pp.
PATIENCE, A. 1906. On the occurrence of Gobius orca, Collett, within the Clyde Sea Area.

Trans, nat. Hist. Soc. Glasgow, 8 : 74-76.

252 P. J- MILLER

PELLEGRIN, J. & FAGE, L. 1907. Description d'un Eleotris Mediterranean nouveau. Bull.

Soc. zool. France, 32 : 11-12.
PETERSEN, C. G. J. 1917. On the development of our common gobies (Gobius) from the egg

to the adult stages etc. Rep. Dan. biol. Sta., 24 : 3-16.

1919. Our gobies (Gobiidae). From the egg to the adult stages. Ibid., 26 : 47-66.

PETIT, G. 1941. Sur la morphologic cranienne de deux poissons cavernicoles aveugles :

Typhleotris madagascariensis Petit et Coecobarbus geertsi Blgr. C.R. Fac. Sci. Marseille

(i), 1 : 36-40.
PRASAD, R. R. 1959. Reproduction in Clevelandia ios (Jordan and Gilbert), with an account

of the embryonic and larval development. Proc. Nat. Inst. Sci. India, 25, B : 12-30.
REGAN, C. T. 1911. The osteology and classification of the gobioid fishes. Ann. Mag. nat.

Hist. (8), 8 : 729-733.
ROBINS, C. R. & BOHLKE, J. E. 1961. A new gobioid fish from the Antilles and comments on

Ctenogobius fasciatus and C. curtisi. Copeia 1961, No. i : 46-50.
RUSSELL, F. S. i926. The vertical distribution of marine macroplankton. II. The pelagic

young of teleostean fishes in the daytime in the Plymouth area, with a note on the eggs of

certain species. /. mar. biol. Ass. U.K., 14 : 101-159.
19266. The vertical distribution of marine macroplankton. III. Diurnal observations on

the pelagic young of teleostean fishes in the Plymouth area. Ibid., 14 : 387-414.
I93oa. The vertical distribution of marine macroplankton. IX. The distribution of the

pelagic young of teleostean fishes in the daytime in the Plymouth area. Ibid., 16 : 639-676.

19306. The seasonal abundance and distribution of the pelagic young of teleostean fishes
caught in the ring-trawl in offshore waters in the Plymouth area. Ibid., 16 : 707-722.

1935. The seasonal abundance and distribution of the pelagic young of teleostean fishes

caught in the ring-trawl in offshore waters in the Plymouth area. Part II. Ibid., 20 : 147-179.

1936. The seasonal abundance of the pelagic young of teleostean fishes in the Plymouth
area. Part III. The year 1935, with a note on the conditions as shown by the occurrence of
plankton indicators. Ibid., 20 : 595-604.

1939. On the seasonal abundance of young fish. VI. The year 1938. Ibid., 23 : 381-386.

1940. On the seasonal abundance of young fish. VII. The year 1939, January to August.

Ibid., 24 : 265-270.
SAEMUNDSSON, B. 1908. Oversigt over Islands fiske med oplysninger om deres forekomst,

vigtigste biologiske forhold og okonomiske betydning. Skr. Komm. Havunders., No. 5,

140 pp.
1926. Fiskarnir (Pisces Islandiae.) Islenzk dyr, 1, 291 pp.

1927. Synopsis of the fishes of Iceland. Rit. Vis. Isl. Reykjavik, 2 : 1-66.
1949. Marine Pisces. Zool. Iceland, 4, Part 72, 150 pp.

SANZO, L. 1911. Distribuzione della papille cutanee (organi ciatiformi) e suo valore sistematico

nei Gobi. Mitt. zool. Sta. Neapel, 20 : 249-328.
SCHMIDT, J. 1904. Fiskeriundersogelser ved Island og Faeroerne i sommeren 1903. Skr.

Komm. Havunders., No. i, 148 pp.

1912. Introduction. Rep. Dan. oceanogr. Exped. Medit. 1908-1910, 1 : 1-49.

SCHNAKENBECK, W. 1928. Rassenuntcrsuchungen am Hering. Ber. Deutsch. Wiss. Komm.

Meeresf. (N.F.), 3 : 91-204.
SMITH, J. E. 1932. The shell gravel deposits, and the infauna of the Eddystone grounds.

/. mar. biol. Ass. U.K., 18 : 243-278.
SMITH, J. L. B. 1951. The genus Paragobioides Kendall & Goldsborough 1911. Ann. Mag. nat.

Hist. (12), 4 : 518-527.
1959- Gobioid fishes of the families Gobiidae, Periophthalmidae, Trypauchenidae, Taen-

ioididae, and Kraemeriidae of the Western Indian Ocean. Ichthyol. Bull. Rhodes Univ., No.

13 : 185-225.

SMITT, F. A. 1892. A History of Scandinavian Fishes. Parti. 566 pp. Stockholm.
1900. Preliminary notes on the arrangement of the genus Gobius, with an enumeration of

its European species. Ofr. Svensk. Vet. Akad. Fork. 1899 : 543-555.

TAXONOMY AND BIOLOGY OF THE GENUS LEBETUS 253

SOUTHWARD, E. C. 1957- The distribution of polychaeta in offshore deposits in the Irish Sea.

J. mar. biol. Ass. U.K., 36 : 49-75.
SPRINGER, V. G. & MCERLEAN, A. J. 1961. Spawning seasons and growth of the code goby,

Gobiosoma robustum (Pisces : Gobiidae), in the Tampa Bay area. Tulane Stud. Zool., 9 : 87-98.
TAKAGI, K. 1950. On the glossohyal bone of the gobioid fishes of Japan, with some phylo-

genetic considerations. Jap. J. Ichthyol., 1 : 37-52.
TANING, A. V. 1940- Island-Faera-Omraadets Kutlinger (Gobiidae). Vid. Medd. Dansk.

naturh. Foren., 104 : 267-282.
TE WINKEL, L. E. 1935- A study of Mistichthys luzonensis with special reference to conditions

correlated with reduced size. /. Morph., 58 : 463-535.
VIVIEN, J. H. 1939- fitude preliminaire du tractus genital chez Gobius paganellus L. Trav.

Sta. zool. Wimereux, 13 : 713-721.

WENT, A. E. J. 1957. List of Irish Fishes. 31 pp. Dublin.
WHITLEY, G. P. 1930. Additions to the check-list of the fishes of New South Wales (No. 3).

Austr. Zool. Sydney, 6 : 117-123.

- 1931. Studies in ichthyology, No. 5. Rec. Austr. Mus. Sydney, 18 : 138-160.
1940. The Nomenclator Zoologicus and some new fish names. Austr. Nat. Sydney 1940 :

241-243.
WINTHER, G. 1877. Om de danske fiske af slaegten Gobius. Naturh. Tidsskr. Kjobenhavn (3),

11 : 41-56.
1879. Prodromus ichthyologiae danicae marinae : fortegnelse over de i de danske farvande

hidtil fundne fiske. Ibid., 12 : 1-68.

APPENDIX

Distribution of Lebetus orca

A list of published and unpublished records known to the author is given below. A complete
sequence of information is as follows : locality, number of specimens, sex, museum number,
standard length + caudal fin length or total length in millimetres (mm.), depth of water in
metres (m.), nature of ground, date of capture, and authority in the case of published records.
When a number of records from the same locality or area are given in the one publication, these
are separated by semicolons and the authority placed at the end. Abbreviations employed are :
c., coarse ground ; cl., clay ; d., damaged; gr., gravel; h., hard ground ; Lith., Lithothamnion ;
m., mud ; midw., midwater ; Mod., Modiolus ; n.d., no date ; s., sand ; sh., shells ; St., stones ;
surf., surface ; t.L, total length ; w., weed.

ICELAND : Off S. and W. coasts, Eystra Horn to BreiSifjordur, " pelagic fry . . . rather rarely "
(Taning, 1940; Saemundsson, 1949).

FAEROES : Off N. coast, and Faero Bank, postlarvae (Taning, 1940).
ROCKALL : On and near Rockall Bank, 1908 (Tailing, 1940).

NORWAY

Hemnefjord (Trondelag) : one <J (ZMO 14001), 23-0 + d.mm., July 1882 (Collett, 1885).

S0ndfjord (Sogn og Fjordane) : Batalden, one ?<J (ZMO J4OOO), t.l. 22-0 mm., 375 m., summer
1875 (Lilljeborg, 1884 ; Collett, 1902).

Fedjefjord (Hordaland) : Herdla, two 6*6* (2MB 4174, 4175), 24-25 + d. and 26-0 + 7-25 mm.,
n.d. and 20 Aug. 1918.

S0rfjord (Hordaland) : Between Bruvik and Stamnes, one 6* (ZMB 5294), 24-0 + 6-25 mm.,
20-30 m., gr.s., ii July 1953.

Hordaland : Bergen area, one, t.l. 29-0 mm. (Lilljeborg, 1884 ; Collett, 1902). Solsvig, one
6* (ZMB 771), t.l. 31-0 mm., 56 m., sh.s., July 1899 (Collett, 1902). Herl0vaer, one $ (ZMB
536), t.l. 28-0 mm., 37 m., sh.s., July 1898 (Grieg, 1899 ; Collett, 1902).

254 P- J- MILLER

Hardangerfjord : Espevaer, one 6* (ZMO J3999), 26-0 + 6-0 mm., 145-180 m., c.s., July
1873 ; Lyngholmen, one ? (ZMO 54021), 30-0 + 7-0 mm., no m., August 1873 (Collett, 1874,
18750). Nordre Bratholmen (Hjelvefjord), one Q*, 23-5 + 7-0 mm., 50-100 m., sh.s., 30 Mar.
1903 ; Bognestremmen, one <J (2MB 1966), 26-5 + 6-0 mm., 20-40 m., gr.s., 26 Apr. 1903 ;
Guldholmen (Solsvik), one - (ZMB 2009), 28-0 + 6-0 mm., 30-60 m., 25 Sept., 1903 ; Djupevik,
one 6* (ZMB 1893), 23-0 + 6-0 mm., 40-80 m., sh.gr.s., and one ?$, 27-5 + 5-0 mm., 20-60 m.,
sh.gr., 2 July 1909 ; Straumastein (Jondal), one <J (ZMB 1887), 18-5 + 4-0 mm., 100-200 m.,
sh.s., 15 June 1909 (Greig, 1913).

Stavangerfjord : Hvitings0, one $ (ZMO J402o), t.l. 28-0 mm., 37 m., July 1872 (Collett,
1874, 18750).

Rogaland : Egersund, one 6* (ZMO J4023), 24-0 + 6-0 mm., 55 m., s., 12 Aug. 1880 (Collett,
1885).

SWEDEN

Kosterfjord (Goteborg) : May 1895 (Duncker, 1928). Sneholm, one, t.l. 35 mm., 100 m., st.;
Sacken, one, t.l. 39 mm., 80 m., coldwater coral reef (Lonnberg & Gustafson, 1935).

Gullmai fjord (Goteborg och Bohus) : August 1926 (Duncker, 1928). Flatholmen, two, 15-20
m., red algae (Lonnberg & Gustafson, 1935).

DENMARK

Kattegat : Anholt Light NW. by W^N., two, 45 m., 18 Apr. 1902 (Otterstr0m, 1912). Anholt,
one, t. 1. i6mm., May 1902 (Petersen, 1919). SW. Anholt Harbour (56 37' N., n 22' E.),
fifty-four, upper and midw. in 18 m., 14 Apr. 1925 ; S. Anholt (56 30-5' N., n 35' E.), sixteen,
upper and midw. in 31 m., 14 Apr. 1925; E. Store Middelgrund (56 33-5' N., 12 13' E.), one,
midw. in 48 m., 8 Apr. 1925 ; 56 15' N., n 53' E., nine, 13.0-17-0 mm., upper and
midw. in 25-26 m., m.s., 20/21 Apr. 1925 (Johansen, 1925). Hessel0, about twenty young,
August 1918 (Petersen, 1919). Yderflakket (between Sjaellands Odde and Hjelmen), one <$
(ZMC 91), 14-4 + 4' mm., n m., st.gr.w., 18 Aug. 1876 (Winther, 1877). Sjaellands Odde, one <
(ZMO 89), 14-5 + 4-5 mm., 22 May 1913.

Great Belt : " fairly often single or a few young specimens ", c. 18-20 m. (Petersen, 1919).
T0rres0 (N. of Fiinen), June 1912 ; Oddens Havn, May 1913 (Duncker, 1928). SE. Roms0
(55 30' N., 10 52' E.), five, midw. in 30 m., 16 Apr. 1925 (Johansen, 1925). Kerteminde, one,
t.l. 15-0 mm., 4 m., 19 May 1904 ; Nyborg, " small young ", Aug. 1917 ; E. Palegrund, two,
t. 1. : 3-5 and 4-5 mm., 6 Aug. 1917 ; E. Vresen's buoy, one, t. 1 : 5-5 mm., 24 m., 19 Sept.
1917 (Petersen, 1919).

Fehmern Belt : SE. Staberhuk (54 22' N., n 23' E.), one, midw. in 23 m., 17 Apr. 1925
(Johansen, 1925).

BRITISH ISLES

Northumbeiland : Off Alnmouth (55 24' N., i 31-5' W.), larvae, 8 Sept. 1925 (Ehrenbaum
et al. (1929) attribute this record to Schnakenbeck (1928) but a search through this paper has
failed to reveal it) .

Clyde Sea Area : Kilbrennan Sound, one Q* (BMNH 90.10.20.9-10), t.l. 37-5 mm., 22 Mar.
1888 (Giinther (1888) as <$ Gobius Jeffrey sii ; Collett, 1896). Between Scalpsie Bay and Cock of
Arran (55 44' N., 5 n' N.), one <, 26-0 + 8-0 mm., 135-160 m., m.; Largs Channel (55 45' N.,
4 54' W.), one $, 29-0 + 7' mm., 35 m., h.; Firth of Clyde (55 41' N., 5 W.), one 6*. 27-0 -{- 7-0
mm., 93 m., m. (Patience, 1906). E. Cumbrae, off Cock of Arran, in over 146 m. (Elmhirst,
1926).

Ireland : Inislyre Harbour, Clare Island (Mayo), one, 9 m., sh.st., May 1909 (Farran, 1912).
Ballynakill Harbour (Galway), several, including two $$, 12-0 + d. and 16-0 + 4-0 mm., and
two $$, 10-5 + d. and 17-0 + 5-0 mm. (BMNH 1903.4.14.4-7), 2-5-5 m - sh.st.Lith.; 30 mi
WNW. Cleggan Head (Galway), one, ?9, 135 m. (Holt & Byrne, 1903).

Irish Sea : Up to 2 mi off Bradda Head (Isle of Man), eight, t.l. 3-5-8-0 mm., at 4-5-7-2 m.,
18 May-ig July 1939 (Bal, 19400). 2^ mi WSW. Chicken Rock (Isle of Man), one, 37 m., sh.,
29 May 1947 ; if mi SW. by W. Chicken, one, 37 m., sh.gr.m., 5 Aug. 1947 ; 2 mi. WSW. Chicken,

255

one, 37 m., sh., n Sept. 1947 ; 3^ mi WNW. Chicken, four, 49 m., sh.s., 6 Aug. 1948 ;

3 mi SW. by S. Chicken, two, 42 m., sh., 21 Sept. 1948 ; 3 mi W. 5 N. Chicken, one, 51 m., sh.st.,
21 Sept. 1948 ; 4^ mi ESE. by 5 E. Langness (Isle of Man), one, 35 m., st.Mod., 24 Sept. 1948 ;
6| mi ESE. by 5 E. Langness, one, 42 m., sh.st.gr., 25 Sept. 1948 ; 3 mi ESE. Douglas Head
(Isle of Man), three, including $ 18-0 + 5-75 mm., 29 m., sh.Lith., 22 Nov. 1948 ; i\ mi WSW.
Chicken, one <J (PEM), 26-5 + 6-5 mm., 44 m., m.s., 10 May 1949 ; 4 mi SW. Chicken, one $
(PEM), 28-0 + 6-2 mm., 40 m., sh.st. (Jones, 1949, 1950, 1951). Five mi E. Douglas Head, one <J,
23-5 + 6-5 mm. and one $, 23-5 + 5-5 mm., 33 m., sh.Lith., 22 Sept. 1948 ; J mi NE. Calf
Stack (Isle of Man), one <J, 26-0 + 6-8 mm., sh.st., 7 May 1957 ' I" 1 m i N. Calf Island, eleven
(Jo*, 20-5 + 5-5 to 26-75 + 7-25 mm., and nine $$, 23-75 + 6-25 to 28-0 -f 6-5 mm., 18-35 m -.
sh.st.Lith., 25 Mar. 1958 to 31 Oct. 1960 ; i mi N. by E. Calf, one $, 25-0 -f- 6-0 mm., sh.m.,

4 Apr. 1960 ; 4 mi S. Spanish Head (Isle of Man), one $, 28-0 + 7-0 mm., 53 m., Mod., n Mar.
1958 ; Off Aldrick Bay (Isle of Man), two $$, 23-5 + 5-5 and 25-75 + 6-5 mm., 20 Oct. 1958 and

5 Nov 1959 ; i mi W. Bradda Head, two o*6\ 23-25 + 6-5 and 30-0 + 7-5 mm., one $ 23-0 +
6-0 mm., 25 Nov. 1958, and two $$, 17-0 + 4-75 and 24-5 + 6-5 mm., 14 Apr. 1959, 27 m.,
sh.st.m.s.w.; 2^ mi S. Perwick Bay (Isle of Man), four 0*6* (including BMNH 1961 . 10. 17. i, 3),
23-0 + 6-2 to 28-0 + 6-3 mm., one $ (BMNH 1961.10.17.2), 29-0 + 7-0 mm., 10 Feb. 1959, and
two (Jo*, 17-0 + 4-5 and 20-0 + 5-0 mm., 16 Nov. 1959, 35 m., sh.st. gr.Lith.; J mi W. Port Erin
Breakwater (Isle of Man), one (J, 15-75 + 4-4 mm., 9-27 m., st.gr. w., 28 Sept. 1959 ; W. Calf
Sound, one (J, 23-0 + 5*75 mm., sh.st. m.s., 29 Sept. 1959 ; Aldrick Bay, one $, 14-5 + 4-0
mm., 18 m., sh.gr., 2 Oct. 1959 ; 2| mi W. Calf Sound, one $, 26-75 + 6-75 mm., sh.st.gr.,
16 Oct. 1959 ; \ mi N. Aldrick Rock, one <J. 24-25 + 5'75 mm., sh.m.s.w., 9 Dec. 1959 ;
4 mi SW. Chicken, one 9- (BMNH 1961.10.17.4), 28-25 + d. mm., 51-58 m., sh.gr., 10 Feb.
1960 ; 6 mi W. Sound, one $, 22-0 + 6-0 mm., and two $$, 25-25 + d. and d. mm., 49-64 m.,
sh.gr.m.s., 3 May 1960 ; 5 mi W. Sound, one (J, 24-5 + 6-75 mm., 55 m., 8/14 June 1960 ; 9^
mi W. Port Erin, one $, 23-5 + 6-0 mm., 60-65 m., sh.m.s., 8 July 1960 ; i J mi N. Chicken, one
<J, 26-0 + 6-0 mm., 42 m., sh. and abundant Flustra, 3 Feb. 1961 ; 3 mi WSW. Chicken, one (J
(BMNH 1961.10.17.5), 25-0 + 6-5 mm., 42 m., sh., 10 May 1961 ; also from Manx area, exact
localities unrecorded, three (Jo*. J 6'O + 4'5 to 26-5 + 7-0 mm., and three $-, 22-0 + 5-0 to 28-5
+ 7-25 mm.

Western English Channel : 6 mi W. Rame Head (Cornwall), one, t.l. n-o mm., midw. in 48 m.,

24 Oct. 1913 (Clark, 1914). Rame Penlee, one, 6-5 mm., 13 m., 3 July 1919 ; 5! mi WSW. Rame,
one, 5-0 mm., 38 m., 21 July 1919 ; 4! mi S. by W. Rame, two, 4-0 and 7-0 mm., 31 July 1919 ;
3J mi NE. by E. Eddystone, one, 6-0 mm., midw. in 55 mm., 5 Aug. 1919 ; 3! mi S. 14 W. Rame,
two, 5-0 and 6-0 mm., 41 m., 5 Aug. 1919 ; 4 mi S. 17 W. Rame, one, 7-0 mm., midw. in 41 m.,
8 Aug. 1919 ; mi E. Eddystone, two, 6-0 and 7-0 mm., 37 m., 8 Aug. 1919 ; 5 mi SW. by W.
Rame, one, 5-0 mm., 50 m., u Aug. 1919 ; 20 mi S.W. 5 W. Eddystone, eleven, 5-0 to 12-0 mm.,
surf, and midw. in 75 m., 9 Sept. 1919 (Clark, 1920). Off Plymouth Sound (50 17' to 18' 30* N.,
4 10' to n' W.), three at 3-8-18-1 m., 13/27 June 1924 ; 10 mi SW. Eddystone, eight, 17 June
1924, and two, 15 July 1926, at 12-4-57-8 m.; 5 mi N. Eddystone, eighteen, at surf, to 52 m.,

25 June 1924/30 June 1926 ; c. 2% mi N. Eddystone, six, at 14-30 m., 17/19 June 1925 ; 2 to 3
mi E. Eddystone, one hundred and thirty, at 3-3-43-1 m., i July 1925/22 Sept. 1926 ; other
postlarvae recorded May/Oct. and Dec. 1927 ; July/Oct. 1929 ; June/Oct. 1930/34 ; May, Sept.
and Oct. 1935 ; June, July and Oct. 1938 ; Aug. 1939 (Russell, 1926^, b, ig^oa, b, 1935. I 936,
1939, 1940)- 200 to 300 yds. (4 to 6 cables) NNW. Eddystone, occasional, sh.gr. (Marine Biological
Association, 1957 ' ground described by Smith, 1932). Falmouth Harbour (Cornwall), one $,
t.l. 21-0. mm., 33 m., sh., 13 July 1897 (Holt & Byrne, 1898). 34 to 47 mi S. 17 to 27 W.
Eddystone, one, t.l. 33-0 mm., 90-97 m.; 46-4 mi S. 25 W. Eddystone, one, t.l. 31-0 mm., 91
m., sh.st. gr.s. (Crawshay, 1912).

FRANCE

Western Channel : estuaries of Morlaix and Penze, near Roscoff (Finisterre), fifteen, 10-0 -f
3-0 to 17-5 + 4-5 mm., 5-8 m., " maerl " (Le Danois, 1910, 1913)- Baie de Morlaix, Roscoff,
" rare ", 4-15 m., sh. maerl (Cantacuzene, 1956). Off NW. Finisterre (48 40' N., 5 4' W., and

256

P. J. MILLER

48 43' N., 4 45' W.), one, t.l. 7-0 mm., in 113 m., s., 7 May 1906, and two, t.l. n-o and 12-0 mm.,
in over 100 m., st., 19 Sept. 1910, respectively (Fage, 1918 ; Schmidt, 1912).

Bay of Biscay : Sables d'Olonne (46 27' N., 6 30' W.), two, t.l. 24-0 mm., 166 m., m.s., 20
July 1886 (Collett, 1896).

DOUBTFUL RECORDS

Fage (1918) regarded larval fishes from the Straits of Gibralter (Atlantic) as possibly referable
to Gobius scorpioides, at the following stations : 35 45' N., 5 59' W., sh.st., in 58 m., 21 Feb.
1909 ; 35 50' N., 6 3' W., st., in 490 m., 21 Feb. 1909 ; 35 57' N., 6 o' W., m., in 275 m. (
23 June 1910 ; 35 51' N., 5 58' W., clay, in 343-720 m., 8 Sept. 1910. Station details are from
Schmidt (1912).

Bull. B.M. (N.H.) Zool. 10, 3

PLATE I

if
w * ' /

PLATE i

Male L. ovca (BMNH 1961 . 10. 17.5) with extended first dorsal fin. From colour-transparency
by Mr. D. J. Slinn. Scale 5 mm.

THE FRESHWATER GASTROPOD

MOLLUSCS OF WESTERN ADEN

PROTECTORATE

C. A. WRIGHT

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 4

LONDON: 1963

THE FRESHWATER GASTROPOD MOLLUSCS
OF WESTERN ADEN PROTECTORATE

C. A. WRIGHT

Pp. 257-274; Plates I-II ; 12 Text-figures, i Map

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 4

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. 4 of the Zoological
series.

Trustees of the British Museum 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued March 1963 Price Ten Shillings

THE FRESHWATER GASTROPOD MOLLUSCS
OF WESTERN ADEN PROTECTORATE

C. A. WRIGHT

IN the spring of 1962 a collection of freshwater gastropod molluscs was made during
the course of an investigation into the transmission of human schistosomiasis in the
Western Aden Protectorate. Material was collected in four areas, Abyan, an irri-
gated part of the coastal plain about 30 miles east of Aden, the Wadi Yeshbum in
Upper Aulaqi near the border with the Eastern Protectorate, the Wadi Hatib south
of Nisab, also in Upper Aulaqi and at Museimir in Haushabi, on the Wadi Tiban near
to the Yemen border.

I am indebted to Dr. C. R. Jones, the Health Adviser, Aden Protectorate, and to
Dr. J. Markham, Assistant Health Adviser, Western Aden Protectorate, for the
facilities to carry out this work and for practical help in the field. I also wish to
thank Mr. Andrew Fuller, Assistant Adviser, Upper Aulaqi, for his help and Mr.
D. Claugher for his assistance both in the field and in the laboratory. The work was
made possible by grant number 3650 from the U.S. Public Health Service.

HISTORICAL

Paladilhe (1872) published descriptions of several land and freshwater molluscs
from the area around Aden and his work provided the basis for subsequent reviews
by Jousseaume (1889) and Ancey (1906), both of whom added further species to the
known molluscan fauna of the area. Connolly (1941) reported on the non-marine
mollusca collected by the British Museum (Natural History) expedition to Yemen
and the Western Aden Protectorate and included an account of specimens collected
by the Lord Wakefield expedition to the Hadhramaut. At the end of that paper
Connolly gave a faunal list for Southern Arabia, based on Ancey's work but with
additions. In recent years the only references to freshwater molluscs in South-
west Arabia have been in reports of schistosomiasis surveys by Petrie &Seal (1943),
Kuntz (1952), Ayad (1956), Azim & Gismann (1956) and Farooq (1961) but only the
first of these was directly concerned with the Western Aden Protectorate.

Topography of the areas visited

The general topography of South-west Arabia was described by Scott (1939). The
Western Aden Protectorate lies at the western end of the southern shore of the
Arabian peninsula and extends inland for several hundred miles at its border with
the Eastern Protectorate. There is a wide coastal plain of desert, broken in some
places by barren outcrops of hills, and inland the folded ridges of mountains reach

ZOOL. 10,4 i5

2 6o C. A. WRIGHT

heights of over 7,000 feet. To the north lie the mountainous areas of Yemen and to
the north-east is the edge of the vast desert of Rub' al Khali.

Throughout most of the region bodies of permanent, fresh, surface water are rare.
Rainfall is erratic in the lower hills, sometimes torrential, sometimes absent and
much of the land surface is porous. In the highland areas there are more regular,
heavy monsoon rains during which rivers flow through the inland valleys but few
of them reach the sea except when occasional flash-floods descend, scouring the beds
of the wadis which dissect the coastal plain. Cultivation in the uplands is mostly
confined to valleys where the water table is high and wells provide a barely adequate
water supply. In some of these areas springs occur, and short, permanent streams
run above ground for distances varying from a few hundred yards to several miles

YEMEN

EASTERN

ADEN

PROTECTORATE

RED } f WESTERN ADEN PROTECTORATE

SEA

ADEN

GULP OP ADEN
Map showing approximate positions of the areas in which collections were made.

before disappearing. This is the situation in the Wadi Hatib near Nisab in Upper
Aulaqi State. Near the lower (north) end of the valley at Ma'rbah there are a number
of small springs whose waters spread out to form a marshy area in which dense
thickets of large reed-mace (Typha] thrive. In the rivulets and pools of this marsh
there are many fish, frogs, aquatic insect larvae and snails. Further up the valley
at Rassaiss there are one or two large springs which give rise to a clear gently flowing
stream with deep, rocky pools in which fish are abundant and freshwater turtles
also live. Snails are common in the shallow parts of the stream, also in a rock pool
not connected with it in a cleft of the cliff above the main spring. Higher up the
wadi at Tarbak a small river rising from a spring has cut a gorge in the otherwise
barren floor of the valley. During its course of about a mile this stream passes
through wide, almost still pools then runs over shallow, shingled beds before flowing

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 261

rapidly through a narrow, rocky channel and down a waterfall into a deep pool and
on down the gorge. In this short but varied course there is a wide variety of habitats
and this is reflected in the occurrence of a richer snail fauna than was found anywhere
else in the Protectorate.

Further to the east in Upper Aulaqi State is Wadi Yeshbum, a relatively fertile
valley with a high water table. A spring near As Sufal gives rise to a small, shallow
stream which flows several miles toward the escarpment marking the boundary
between the Eastern and Western Protectorates. Near Said and Yeshbum there are
narrow gorges entering the main valley and in some of these there are deep rock
pools which persist for long periods and may even be permanent.

In complete contrast to this terrain is Abyan, a low-lying irrigated district in the
coastal plain about thirty miles east of Aden. This irrigation scheme owes its exist-
ence to the Wadi Bana, one of the few permanent rivers which reaches the coastal
plain, and the Wadi Hasan which, although not wholly permanent has sections
where springs give rise to small perennial streams in the river bed. Near the north
end of the district the two wadis almost converge and the irrigation works have
provided a link so that water from either source can be fed into the canal system.
Near to this point is a permanent spring, independent of either wadi, whose water
provides irrigation for vegetable gardens in the near-by village of Bateis. The
general irrigation system is only in use seasonally and the principal snail habitats
appear to be confined to a few parts of the permanent, natural waters in the wadis
and the spring.

The Wadi Tiban at Museimir in Haushabi was the only other locality visited and
it is in some respects intermediate in character between the habitats in Upper Aulaqi
and those in Abyan. There is a permanent shallow stream flowing over the shingle
bed of the wadi in a low-lying valley surrounded by hills. Despite its lower altitude
(about 500 feet) the snail fauna at Museimir is the same as that at As Sufal in the
Wadi Yeshbum (about 3,000 feet).

Family PLANORBIDAE

Biomphalaria rueppelli (Dunker)

(PL I, 1-8}

Planorbis ruppellii Dunker, 1848, p. 42.

Planorbis arabicus Melvill & Ponsonby, 1896, p. 3, PI. I, figs. 15-17.

Biomphalaria arabica Connolly, 1941, p. 33.

Biomphalaria boissyi arabica Kuntz, 1952, p. 25.

Biomphalaria sp. (in part) Ayad, 1956, p. 90.

UPPER AULAQI : Wadi Yeshbum at Marbum, As Sufal and Sumayfah, either in
rock pools or gently flowing water, 3,000-3,200 ft.

Wadi Hatib, Ma'rbah and at Tarbak and Rassaiss, in streams, about 4,000-4,500 ft.

HAUSHABI : Museimir, in the Wadi Tiban in shallow, flowing water, about 500 ft.

The whorls of the shell are flattened above and bluntly angled beneath, the inner-
most are deeply sunk on the upper surface and the umbillicus is narrow. Apart from
a well-marked spiral sculpture on most of the specimens this material does not differ

262 C. A. WRIGHT

significantly in shell-form from material collected in the highlands of Ethiopia
(Wright & Brown, 1962). The largest specimen found in Aden (at Sumayfah)
measured 10-5 mm. in maximum diameter, 3-2 mm. umbilical diameter and 4-1 mm.
in height. These dimensions are almost the same as those of the holotype of
B. arabica (10.8, 3.65, 4.0).

The mantles of all the specimens dissected were uniformly darkly pigmented and
without distinct markings. The penis sheath and preputium are approximately
equal in length in most individuals but in a few the sheath is a little shorter than the
preputium. The number of primary prostate diverticula is at least twelve in adult
snails.

The mesocones of the lateral radula teeth are all triangular ; none of the arrow-
head type were seen. Subdivision of the marginal ectocones is general but in some
radulae the marginal ectocones may be undivided in one or two longitudinal rows

(Fig. S).

This species is widely distributed in South-west Arabia and is the dominant
basommatophoran in areas away from the coast. The type locality for B. arabica is
Dhofar in the Hadhramaut and Connolly (1941) records the species from several
localities in the Western Aden Protectorate and Yemen. Through the kindness of
Dr. Naguib Ayad of the Ministry of Public Health, Cairo, I have been able to examine
shells of the specimens which he collected in Yemen and there is no doubt that his
material also belongs to this species. Kuntz (1952) treated arabica as a sub-species
of B. alexandrina (= boissyi) but the present study shows that it is correctly placed
in the pfeifferi group. Jousseaume (1889) referred to the young planorbid recorded
by Paladilhe (1872) from alluvial debris at Kursi, near Aden, as belonging to the
adowensis group and it has recently been shown that B. adowensis is almost certainly
not distinct from B. rueppelli (Wright & Brown, 1962).

B. ruepelli is found in practically every body of permanent fresh water in the part
of Upper Aulaqi which was visited. The snails reach maturity at a relatively small
size and lay large numbers of eggs. The population density in some rock pool habitats
is very great and, with the large numbers present and their wide distribution, it is
surprising that infection with Schistosoma mansoni is not a more serious public health
problem than it appears to be.

Gyraulus convexiusculus (Hutton)

(PI. I, 9-17)
Planorbis convexiusculus Hutton, 1849, p. 657 ; Connolly, 1941, p. 36.

UPPER AULAQI : Sumayfah, Wadi Yeshbum, from deep rock pools and shallow,
flowing water, about 3,200 ft.

Tarbak, Wadi Amhadu (part of Wadi Hatib), from a slow-moving part of the
stream, abundant on aquatic vegetation.

LOWER YAFAI : Nabwa spring, Bateis, Abyan, three very young specimens, prob-
ably belonging to this species.

The shells are small, flattened and the whorls do not increase in diameter so rapidly
as they do in the common African species, G. costulatus. The mean dimensions of a

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN

263

FIG. i. Gyraulus convexiusculus. Complete reproductive system.

FIGS. 2 & 3. Bulinus reticulatus. Male copulatory organs of two individuals from Rassaiss.
FIG. 4. Ancylus fluviatilis . Complete reproductive system of specimen from Tarbak.

264 C - A - WRIGHT

small sample of adult specimens from Tarbak are, diameter 4-3 mm., umbilicus
1-6 mm. and height i-i mm. There is no obvious angulation of the whorls and the
aperture is roughly oval. There is a superficial, somewhat coarse spiral sculpture on
most of the specimens, recalling the appearance of G. kigeziensis nyanzae Mandahl-
Barth 1954.

Anatomically G. convexiusculus differs from G. costulatus in the less compact
arrangement of the accessory genital glands, a reflection of the differences in shell-form
between the two species (Text-fig, i). The penis sheath and preputium are roughly
equal in length but the proximal dilatation of the penis sheath is wider than the
maximum diameter of the preputium. The structure of the penial stylet is the same
as that described for G. costulatus (Wright, in press). The prostate in G. convexiusculus
is long and has as many as eighteen primary diverticula without marked secondary
branching. The radula teeth are very small (Text-fig. 6) and the cusps of the laterals
are longer and more lanceolate than those of G. costulatus. There are six tricuspid
laterals in each half-row, about five transitional teeth which resemble the laterals
except in the division of the ectocone and about six marginals. The marginals have
the ecto- and endocones divided into two cusps each, resulting in five-cusped teeth
instead of the seven-cusped marginals in Angolan G. costulatus.

Connolly (1941) recorded this species from the Hadhramaut at Hureidha, Wadi
Jedd. The present material is so close to a series collected at Kandahar (Afghanistan)
by Hutton (possibly the type material, B.M. (N.H.) collection number 56.9.15.79)
that there is no reason to consider it to be a separate species unless subsequent
anatomical studies of topotype material reveal any marked differences.

Bulinus sericinus (Jickeli)
(PI. II, i-4)

Isidora sericina Jickeli, 1874, p. 194, PI. Ill, fig. 2.
Bulinus truncatus Connolly, 1941 (in part), p. 33.

UPPER AULAQI : Tarbak, Wadi Amhadu (part of Wadi Hatib), from rushes and
emergent grasses in an almost still pool, also on stones and rocks in a narrower, more
rapidly flowing part of the stream, about 4,500 feet.

A large sample was collected and it includes a wide range of shell forms. Wright
& Brown (1962) have discussed variation in this polymorphic species from the
Ethiopian highlands and this material from Aden includes most of the forms
described. The majority of the smaller specimens, some smooth-shelled, with or
without inflated whorls, some strongly ribbed (var. harpula Pollonera), were collected
on vegetation in an almost still part of the stream while larger individuals, most of
them with exserted spires (shackoi Jickelli) were more abundant on rocks in rapidly
flowing water.

Anatomically the Aden material differs from the Ethiopian populations studied
earlier in that the majority are aphallic. This condition was found in only one of a
large number of Ethiopian specimens. Euphallic individuals in the Tarbak sample
have a copulatory organ similar to that described by Mandahl-Barth (1960) for
B. mutandensis. The penis sheath is longer than the preputium and, at its maximum

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 265

diameter, it is also a little wider. The sheath is regularly banded with dark rings,
the epiphallus is short and, in one specimen a slight proximal dilatation of the epi-
phallus was seen. The radula teeth of B. mutandensis are described by Mandahl-
Barth as about twice the size of those of B. tmncatus but the teeth of the specimens
from Tarbak (Fig. 7) do not differ significantly from those of Ethiopian B. sericinus.

Ayad (1956) mentioned that Mandahl-Barth identified all of his bulinids from the
highlands of Yemen as B. sericinus (Jickeli) and, after examining the Yemeni speci-
mens named as B. truncatus by Connolly (1941), I think that this is a valid action.
Connolly's material was all found between altitudes of 7,900 and 9,300 ft. A sample
of six specimens in the collection of the British Museum (Natural History) collected
by Dr. Haythornthwaite from Lower Yafai do not have an exact locality and the
altitude at which they were found is not known. At the present time it seems that
my sample from Tarbak is from the lowest altitude yet recorded (about 4,500 ft.)
for this highland species.

The role of B. sericinus as an intermediate host for schistosomes is not yet certain.
One of the specimens from Tarbak shed schistosome cercariae about two weeks after
it was collected but the species of parasite has not been identified. A few snails
from this sample were exposed to miracidia of Schistosoma haematobium hatched
from a urine sample collected at Said but no cercariae were produced.

Bulinus reticulatus Mandahl-Barth 1954

(PL II, 5-8)
Bulinus truncatus (var.) Connolly, 1941, p. 37.

UPPER AULAQI : Wadi Yeshbum at Marbum, near Said, in rock pools in a narrow
gorge, about 3,200 ft.

Rassaiss, Wadi Hatib, south-east of Nisab, in a moderate-sized pool in a cleft of
a small cliff at the head of a gulley, about 4,000 ft.

The shells of these specimens compare well with those of a small sample from the
type locality, Kisumu in Kenya. Those from Said tend to have the spire more ex-
serted and those from Rassaiss are uniformly globular with a pronounced spire. The
largest specimen collected (from Said) had a shell-length of 6-1 mm., aperture length
of 3-7 mm. and maximum width of 4-0 mm. The mean length of twenty shells from
Said was 4-9 mm. and from Rassaiss 5-1 mm. The mean ratio of shell-length to
aperture length is i-6in both samples but the more globular specimens from Rassaiss,
have a lower length /width ratio (1-30) than the narrower specimens from Said (1-42).

The most marked macroscopic feature of the shell in this species is the broadly
reflexed, straight collumella and wide umbilicus. On dead shells collected at Said
there is a very strongly marked reticulate sculpture which can be seen clearly with
a hand lens but fresh, translucent shells require more careful examination.

Preliminary observations on the anatomy of this material showed that it differed
in several respects from the descriptions published by Mandahl-Barth (1954 & 1957).
Three specimens from Kisumu, Kenya, were therefore dissected for comparison.
The mantle in the snails from Aden appears uniformly black, due to the dense and

266 C. A. WRIGHT

even distribution of chromatophores, and it has a pale grey margin which is often
reflexed over the edge of the aperture ; the mantle markings of the Kisumu speci-
mens are regular black spots and blotches on a light grey ground. The male copu-
latory organ is very large relative to the size of the animals, about 3'5~4'0 mm. total
length in a 5 mm. shell (Text-figs. 2 and 3). According to Mandahl-Barth the
preputium is exceptionally big and it is wider and longer than the sheath. In material
from both localities in Aden the sheath is a little longer than the preputium and its
maximum diameter is at least equal to that of the distal part. The specimens from
Kisumu have the sheath and preputium roughly equal in length and in only one of
the three dissected was the preputium conspicuously wider than the sheath. The
epiphallus is short. Mandahl-Barth refers to the radula teeth of B. reticulatus as
very small and with the endo- and mesocones not completely separated. The teeth
of the Kenyan specimens are certainly small and the cusps are fine and lanceolate but
there is scarcely any fusion of the inner and middle cusps of the laterals (Text-fig. 9) .
The Aden specimens, on the other hand, have larger teeth, the mesocones of the
laterals are very big and spatulate and the endocones are reduced to small points on
the inner margins of the mesocones (Text-fig. 8) . There are eight or nine laterals in
the Aden radulae and twelve to fourteen marginals while those from Kisumu have six
or seven laterals and sixteen to eighteen marginals. The marginals of the Kenyan
specimens are more delicately pectinate than those from Aden.

B. reticulatus in East Africa is known only from temporary habitats and it seems
likely that the pools at Said and Rassaiss may not be permanent. Water originating
from some springs in the Wadi Hatib has a high mineral content which results in a
white crust being deposited around the edge of drying pools. Well-water from Nisab
has a total dissolved solids content of about 1750 p.p.m. of which 480 p.p.m. are total
hardness and about 550 p.p.m. are sodium chloride. Analysis of a small sample of
water from the pool at Rassaiss (carried out by the laboratory of the Government
Chemist, London) showed total dissolved solids of only 200 p.p.m. of which 40 p.p.m.
were total hardness and 60 p.p.m. sodium salts. This suggests that the pool is filled
largely by rainwater rather than from a spring and it is almost certain, therefore, to
dry out periodically. The only other animals found in the Rassaiss pool were a
single pair of waterbeetles, Hydaticus jucundus (Reiche) and a number of anostracan
Crustacea, kindly identified for me by Dr. J. Harding as Streptocephalus neumanni
Thiele ; both of these species are characteristic elements of ephemeral waters.

Connolly (194 T 1 described in some detail a large sample of what he considered to
be a variety of B. truncatus from Al Bahr in the Hadhramaut and he mentioned that
they were so different from the typical form that they might be considered to be a
separate species. These specimens have been re-examined and proved to be B. reti-
culatus. The isolated and obscure habitats in which the snails occur at Rassaiss and
Said, also their probably temporary nature, suggest that this species is likely to be
overlooked in any routine survey for medically important snails. It was because the
Marbum pools at Said were reputed to be the local source of infection with urinary
schistosomiasis that they were visited and the pool at Rassaiss was found only by
chance. From three records for the species in Southern Arabia it is impossible to
guess at its distribution but it is likely to be found in the middle heights throughout

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 267

the Protectorate and possibly in Yemen also. Epidemiological evidence points to
this snail as the intermediate host of Schistosoma haematobium at Said and specimens
from Rassaiss have proved to be excellent hosts in the laboratory for the Said strain
of parasite. The snails breed easily in the laboratory, the egg-masses are relatively
large for a small bulinid with up to fourteen eggs in each and the size at hatching is
smaller than in B. forskali but the " infant mortality " rate is lower.

Bulinus beccari (Paladilhe)
(PI. II, 9-14)

Physa beccarii Paladilhe, 1872, p. 23, PI. I, figs. 7 & 8.
Bulinus for sk alii Azim & Gismann, 1956, p. 436, figs. 18, 19.

FAHDLI : Wadi Hasan at Dirgag, Abyan, in shallow, running water with dense
aquatic vegetation and a sandy bottom.

LOWER YAFAI : Bateis, Abyan, in the Nabwa spring in cool, relatively deep gently
flowing water on Ceratophyllum and other aquatic plants, also a single living specimen
and one dead shell in the Wadi Bana under a stone in shallow, fast-running water.

The shell is high-spired and elongate, translucent white, without marked ribbing
and with a slight carination forming an ill-defined shoulder on the earlier whorls.
The columella is slightly twisted at its upper end and its margin is reflexed, almost
closing the umbilicus. The largest specimens seen (from Dirgag) have a shell-length
of 7 -3 mm. but the majority of adult specimens in both of the large samples do not
exceed 6-0 mm. shell-length, with maximum width 3-4 mm. and aperture length
also 3-4 mm.

The anatomy of the animal is in no way exceptional. The mantle is light grey,
densely spotted with black and the rest of the body is darkly pigmented so that the
living snail appears black. The male copulatory organ is like that of most members
of the B. forskali group, the sheath is a little longer and narrower than the preputium
and the epiphallus is short. The radula teeth are smaller than in B. forskali, the cusps
of the laterals are long and lanceolate and, in the majority of specimens there is an
accessory cusp high up on the outer edge of the ectocones (Text-fig. 10).

The affinities of B. beccari have never been clear. Paladilhe recognized that it was
closely related to many species characteristic of the African fauna but he was able
to distinguish between it and lamettosa Roth and vitrea " Parreys ", both now regarded
as synonyms for B. forskali (Ehrenberg). Germain (1921) included beccari with the
Mauritian species cernicus in the synonomy of forskali and Mandahl-Barth (1957),
following Germain, provisionally placed beccari in the synonomy of cernicus. The
general appearance of the shell of beccari is close to some forms of senegalensis from
the Gambia (Wright, 19596) but it is distinguished from that species by the presence
(on most specimens) of a weak carination on the upper part of the early whorls.
B. beccari differs from most forms of B. forskali in the poorly developed ribbing of
even the early whorls and in its generally more obese body whorl and less exserted
spire. The radula teeth are smaller than senegalensis, forskali or scalaris and there
is no fusion of the endo- and mesocones of the laterals as in cernicus.

The importance of B. beccari as an intermediate host for Schistosoma haematobium

268

C. A. WRIGHT

Radula teeth.

FIG. 5. Biomphalaria rueppelli. Central, i, n and 16.
FIG. 6. Gyraulus convexiusculus. Central, i and 15.

FIG. 7. Bulinus sericinus. Central, i, 6 and 12.
FIG. 8. B. reticulatus from Rassaiss. Central, I, 2, 9 and 16.

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 269

in the coastal plain regions of Aden is described elsewhere (Wright, in press). The
ability of the species to act as a host for this parasite suggests possible affinities with
either B. senegalensis or B. cernicus. Ecologically beccari differs from senegalensis, a
snail so far only known from temporary rain pools with a low calcium content, in.
that it is found in permanent, flowing water with a high calcium content.

Family LYMNAEIDAE

Lymnaea truncaluta (Miiller)

(PI. I, 18-20)

Buccinum truncatulum Miiller 1774, vol. II, p. 130.
Lymnaea (Galba) truncatula Connolly, 1941, p. 33.

UPPER AULAQI : Tarbak, Wadi Amhadu (part of Wadi Hatib) , under small stones
and amongst gravel both in the water and on the edge of the stream at a point where
it was flowing rapidly.

Several dozen specimens of this species were collected from a very localized and
unusual habitat. The shells do not differ in any significant respect from specimens
from the Ethiopian highlands (Wright & Brown, 1962) and the anatomy of the repro-
ductive system shows no unusual characters. The radula teeth are interesting in
that the first five laterals are all bicuspid, the endocone is reduced to a slight swelling
on the inner edge of the mesocone and this cusp and the ectocone are unusually long
and dagger-like (Text-fig, n). The endocone appears as a discrete cusp near the tip
of the mesocone in the sixth row and in the next few rows the division becomes
more marked. Transition to the marginals occurs by subdivision of the endocone,
the ectocone remains undivided and eventually disappears as do most of the divis-
ions of the endocone so that the extreme marginals are bicuspid.

Connolly (1941) recorded two specimens of this species collected from a muddy
brook near San'a, Yemen. The present record merely extends its known distribution
a little to the south and east.

Family ANCYLIDAE

(PL II, 15-22)
Ancylus fluviatilis Miiller. 1774, Vol. II, p. 201

Pseudancylus abyssinicus Connolly, 1941, p. 34.

Pseudancylus argenteus Connolly, 1941, p. 37, PI. Ill, figs. 11-12.

UPPER AULAQI : Sumayfah, Wadi Yeshbum, from a deep, shaded rock-pool in a
side gulley near Yeshbum, about 3,200 ft.

Tarbak, Wadi Amhadu (part of Wadi Hatib), from the rocky sides of a swift-
flowing, shallow channel, in the stream bed, about 4,500 ft.

There are no obvious differences between the ancylids collected in Upper Aulaqi
and European specimens of A. fluviatilis. The description of the genital system
of this species by Lacaze-Duthiers (1899) differs only from the Aden material in that

270 C. A. WRIGHT

he shows the flagellum to open into the penis sheath while in my specimens the duct
appears to open separately into the preputium (Text-fig. 4) ; this is a character
which could so easily be affected by fixation methods that it cannot be regarded as
having any taxonomic importance.

Connolly (1941) described Pseudancylus argenteus from pools in Shab Samua,
Hureidha in the Hadhramaut and recorded P. abyssinicus (Jickeli) from a pond near
Haz in Yemen. In the collection of the British Museum (Natural History) there is
a sample of P. argenteus (identified by Connolly) from Mesayid, Yemen. The shells
of my material from Sumayfah compare well with Connolly's abyssinicus and the
larger specimens from Tarbak are only distinguishable from Connolly's Yemen
material of argenteus by their darker colour. The type series of argenteus have the
apex of the shell less sharply recurved than in the other samples. Connolly noted
that in argenteus the ectocone of the marginal radula teeth is unicuspid while in
abyssinicus and fluviatilis the ectocone is subdivided. This observation was based
on radula preparations made from the type series of argenteus and from the Yemen
sample of abyssinicus. Re-examination of the original slides shows that this character
is slightly variable. In the majority of my specimens the ectocone is unicuspid but
in one radula it is undivided in the marginals on one side and bicuspid on the other
(Text-fig. 12).

Attention was drawn to the similarity between the radulae of Ancylus fluviatilis
and abyssinicus by Jickeli (1874) and the close relationship between the two species
was emphasized by Walker (1914). When preserved material from the highlands of
Ethiopia becomes available it is almost certain that the anatomical identity of
abyssinicus with fluviatilis will be confirmed. The range of this Palearctic species
must now be extended to include the highland areas of South-west Arabia.

Family THIARIDAE
Melanoides tuberculata (Miiller)

(PI. II, 23)

Nerita tuberculata Miiller, 1774, Vol. II, p. 191.
Melanoides tuberculata Connolly, 1941, p. 35.

FAHDLI : Wadi Hasan at Dirgag, Abyan.

LOWER YAFAI : Nabwa spring and Wadi Bana, Bateis, Abyan.

UPPER AULAQI : As Sufal, Wadi Yeshbum and several places in the Wadi Hatib.

HAUSHABI : Wadi Tiban at Museimir.

This species is almost universal in its distribution in South-west Arabia, usually
in gently flowing water. Specimens from Abyan in the coastal plain reach a greater
maximum size than those from the hill regions.

DISCUSSION

All of the species in this collection have been obtained before in various parts of
South-west Arabia but this is the first material which has been subjected to detailed
anatomical examination to elucidate the relationships of some of the species. Thus,

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 271

Biomphalaria arabica is shown to be indistinguishable from B. rueppelli of Ethiopian
origin, the Bulinus truncatus of most previous authors is more closely related to the
highland species, B. sericinus, Connolly's B. truncatus var. from the Hadhramaut is
the rare East African B. reticulatus and Pseudancylus argenteus and P. abyssinicus
are only local forms of Ancylus fluviatilis . The specimens collected in Upper Aulaqi
state are the first from that area and all thus provide new locality records which
either extend or fill in gaps in known distribution patterns.

Connolly (1941) noted a marked difference between the non-marine molluscan
fauna of South-west Arabia and that of Hejaz and Nejd. Mattingly and Knight
(1956) found that the boundary line between the Palearctic and Ethiopian zoogeo-
graphies! regions defined by Chapin (1923) on the basis of the avifauna also holds
good for the mosquitoes of Arabia. This line runs roughly south-east from a point on
the Red Sea coast near Jedda to the coast of Oman opposite the Kuria Muria Islands,
and, as a result, places the whole of Yemen and the Aden Protectorates in the Ethio-
pian region. Scott (1939) in a general account of the natural history of South-west
Arabia pointed out that the major regional component of the flora and fauna of the
hot valleys is African modified by an Oriental influence while in the highlands there
is a well-marked Palearctic element. An analysis of the regional components of the
mosquito fauna of Arabia as a whole by Mattingly and Knight showed that the bulk
of the species are Palearctic but that there is a strong Ethiopian element in the
South-west and the Oriental influence is small and mostly confined to the coast.
These authors also concluded that the mosquito fauna of Socotra is predominantly
Palearctic.

The freshwater molluscs appear to conform well to the general zoogeographical
picture in South-west Arabia. Many of the older records are not sufficiently reliable
for consideration but, through the kindness of Dr. Ruth Turner of the Museum of
Comparative Zoology at Harvard I have been able to examine material collected
by Dr. R. E. Kuntz in Yemen and identified by Dr. J. Bequaert. These specimens,
in conjunction with my own and material in the collection of the British Museum
(Natural History) provide a reasonable basis for discussion.

The principal influence is clearly Ethiopian and, in the highland areas, this
influence can be seen in the narrow political-geographical sense rather than the
broader zoogeographical meaning of the adjective. Both Biomphalaria rueppelli
and Bulinus sericinus are characteristic of the highlands of Abyssinia and the rela-
tionship is further reinforced by Kuntz's specimens of Segmentorbis angusta (Jickeli)
found near Ma'bar, Yemen. Mattingly and Knight (loc. cit.) pointed out that the
number of Ethiopian species in the Arabian mosquito fauna is inclined to be exag-
gerated by the inclusion of forms whose distribution in the Ethiopian region is
limited and which are more accurately assigned to the Palearctic. This is also the
case with the freshwater mollusca in that Lymnaea truncatula, Ancylus fluviatilis and
Pisidium casertanum are common to both the Abyssinian and South-west Arabian
highlands but are, in fact widely distributed Palearctic species. Regardless of their
wider distribution, the presence of these three species isolated in the highland areas
on both sides of the southern Red Sea provides very strong evidence in favour of a
common origin of the two faunas.

272 C. A. WRIGHT

Both Bulinus beccari and B. reticulatus are obvious African elements. B. beccari
appears to be an endemic Arabian form but more detailed investigations of the
coastal regions of North-east Africa may reveal its presence there. It is very close
to some forms of B. forskali and it has been confused with that species in the past.
It is probably merely a local race of B. forskali, no more entitled to specific distinction
than many other races, but for its obvious epidemiological importance. B. reticulatus
presents exactly the opposite problems to those posed by B. beccari. It is not a well-
known species and the full range of its variation has not been studied ; the material
from Aden differs markedly in characters of the radula and male copulatory organ
from topotype specimens from Kenya and the Aden form is also a proven host for
Schistosoma haematobium but the intermediate host potential of African B. reticulatus
has not been investigated. In another paper (Wright, in press) I have suggested
that B. crystallinus and B. camerunensis may be relicts of an ancestral bulinid stock
and it is possible that B. reticulatus also falls into this category. If this is the case
then considerable differences are to be expected between isolated populations of an
ancient form and morphological differences such as those found between Arabian and
Kenyan B. reticulatus suggest a long period of separation.

A further possible Ethiopian influence in the Arabian fauna is Lymnaea arabica
Smith, 1894. Hubendick (1951) considers that the species is probably a local form
of the extremely variable African L. natalensis. Specimens collected by Kuntz in
Yemen were identified by Bequaert as L. exserta, a form of L. natalensis and several
samples in the collection of the British Museum (Natural History) from Yemen,
Western Aden Protectorate and Dhurfar have been assigned to L. caillaudi, also a
synonym of L. natalensis. However, Connolly (1941) recorded L. auricularia from
near San'a, Yemen and Dr. Bequaert also identified samples collected by Kuntz
from near M'bar and Dhamar as this species. Hubendick (1951) treats L. natalensis
as the African part of the L. auricularia super-species which extends throughout
the Palearctic and Oriental regions. However, L. auricularia, L. natalensis and some
of the Oriental forms of L. auricularia can be distinguished from one another by the
chromatographic pattern of fluorescent substances in their body-surface mucus
(Wright, 19590). Unfortunately no material belonging to this complex was found
during the present survey but, if living specimens become available for study in the
future, it should be possible to decide whether they are derived from Ethiopian,
Palearctic or Oriental origins.

The only clearly Oriental element in the present collection is Gyraulus convexiu-
sculus, a species described originally from Afghanistan but with a wide range in
South-east Asia. Kuntz collected a large sample of this species near San'a, Yemen
and Connolly records it from Wadi Jedd in the Hadrhamaut. Pallary (1928) reported
Indoplanorbis exustus from Muscat and Connolly included this species in his faunal
list for Southern Arabia but expressed doubts as to the correct identification of the
material. However, Pallary's specimens are in the collection of the British Museum
(Natural History) and I am able to confirm that they are shells of I. exustus. Two
samples in Kuntz's Yemen collection identified as this species are in fact Biomphalaria
rueppelli and Pallary's specimens are the only confirmed record for I. exustus in
Southern Arabia. Muscat lies in the extreme south-eastern corner of the Arabian

FRESHWATER CASTOPOD MOLLUSCS OF WESTERN ADEN 273

peninsula on the Persian Gulf, the commonly accepted boundary between the Pale-
arctic and the Oriental regions, and the presence of this typically Oriental species
is not, therefore, surprising. A collection of freshwater gastropods from Socotra
has been examined and in five large samples the dominant species present was
I. exustus, an Oriental influence which is in contrast to Mattingly and Knight's
comments on the predominantly Palearctic nature of the island's mosquito fauna.
The only other species present in the Socotran collection was Melanoides tuberculata
an almost ubiquitous prosobranch with a range extending from West Africa through
to South-east Asia. In the Western Aden Protectorate this is the most abundant
species of snail and, with the possible exception of Gyraulus convexiusculus, is the
only one that occurs both in the coastal plain and in the highlands.

The conclusions reached by this examination of the relationships of the freshwater
gastropods of South-west Arabia are in accordance with the general zoogeographical
trends in the area. The principle influence is Ethiopian with a Palearctic component
in the highlands (a component which is shared with the Ethiopian highlands) and
a small Oriental element which appears to have a wide altitudinal range. Recent
surveys by medical parasitologists have given the impression that the molluscan
fauna of the area is of the Mediterranean type. This impression was created by the
treatment of Biomphalaria arabica as a sub-species of the alexandrina group rather
than as a member of the pfeifferi complex, by the failure to recognize the differences
between the highland Bulinus sericinus and the lowland B. truncatus and by the
automatic inclusion of B. beccari in the synonomy of B. forskali. That these appar-
ently academic points have an epidemiological significance is shown by the failure
of attempts to infect Middle Eastern strains of B. truncatus with two strains of
Schistosoma haematobium from the Western Aden Protectorate (Wright, in press).
They explain also Witenberg and Saliternik's (1957) failure to infect Israelis, truncatus
with S. haematobium obtained from Yemeni immigrants.

SUMMARY

Anatomical study of a small collection of freshwater gastropods from the Western
Aden Protectorate has permitted a reassessment of the affinities of some of the species
recorded from South-west Arabia. The species represented in the collection are
Biomphalaria rueppelli, Gyraulus convexiusculus, Bulinus sericinus, B. reticulatus,
B. beccari, Lymnaea truncatula, Ancylus fluviatilis and Melanoides tuberculata. The
zoogeographical aspects of the basommatophoran fauna are discussed and the epi-
demiological significance of the bulinids as intermediate hosts for human blood-
flukes is mentioned.

REFERENCES

ANCEY, C. F. 1906. Releve des mollusques terrestres et fiuviatiles de la peninsule Arabique.

/. de Conchyl., 53 (3) : 257-271 (1905).
AYAD, N. 1956. Bilharziasis survey in British Somaliland, Eritrea, Ethiopia, Somalia, the

Sudan and Yemen. Bull. Wld Hlth Org., 14 : 1-117.
AZIM, A. M. & Gismann, A. 1956. Bilharziasis survey in South-western Asia. Bull. Wld Hlth

Org., 14 : 403-456.
CHAPIN, J. P. 1932. The birds of the Belgian Congo. Bull. Amer. Mus. nat. Hist., 65 : 1-756.

274 c - A - WRIGHT

CONNOLLY, M. 1941. South Arabian non-marine mollusca. British Museum (N.H.) Expedition

to South-west Arabia, 1937-8, 1 (4) : 17-42.
FAROOQ, M. 1961. Report on a visit to Saudi Arabia. Wld Hlth Organisation, EM/BIL/ig,

Saudi Arabia 24.

GERMAIN, L. 1921. Faune malacologique des Isles Mascareignes. Angers.
HUBENDICK, B. 1951. Recent Lymnaeidae. Kungl. Svenska. Vetenskap. Hand., 3 (i) : 1-221.
HUTTON, T. 1849. Notices of some land and freshwater shells occurring in Afghanistan. /.

Asiatic Soc. Bengal, 18 (2) : 649-661.
JICKELI, C. F. 1874. Fauna der Land- und Susswasser Mollusken Nord-Ost-Afrikas. Nova

Ada Leop. Carol., 37 : 1-352.
JOUSSEAUME, F. 1889. Especes nouvelles des environs d'Aden suivies d'un apercu sur la faune

malacologique de la peninsule Arabique. Bull. Soc. Malac. France, 6 : 345-362.
KUNTZ, R. E. 1952. Schistosoma mansoni and S. haematobium in the Yemen, Southwest Arabia :

with a report of an unusual factor in the epidemiology of schistosomiasis mansoni. /.

Parasit., 38 (i) : 24-28.
LACAZE-DUTHIERS, H. 1899. Des organes de la reproduction de I'Ancylus fluviatilis. Arch.

Zool. exp. Gen., 7 : 33-120.
MANDAHL-BARTH, G. 1954. The freshwater mollusks of Uganda and adjacent territories. Ann.

Mus. Roy. Congo Beige Tervuren Zool., 32 : 1-206.
1957- Intermediate hosts of Schistosoma. African Biomphalaria and Bulinus II. Bull.

Wld Hlth Org., 17 : 1-65.
1960. Intermediate hosts of Schistosoma in Africa, some recent information. Butt. Wld

Hlth Org., 22 : 565-573.
MATTINGLY, P. F. & KNIGHT, K. L. 1956. The mosquitoes of Arabia. I. Bull. Brit. Mus. (Nat.

Hist.) Entom., 4 (3) : 91-141.
MELVILLE, J. C. & PONSONBY, J. H. 1896. Descriptions of seven new species of terrestrial and

fluviatUe mollusca from the Hadhramaut, South Arabia. Proc. malac. Soc. Lond., 2 (i) :

i-3-
PALADILHE, A. 1872. Descriptions de quelques especes nouvelles des environs d'Aden. Ann.

Mus. civ. Stor. not. Genova, 3 : 12-26.
PALLARY, P. 1928. Mollusques continentaux de sud de 1'Arabie collectes en 1926 par M. Lees.

Proc. malac. Soc. Lond., 18 (i) : 39-42.
PETRIE, P. W. R. & SEAL, K. S. 1943. A medical survey of the Western Aden Protectorate (1939-

1940). Colonial Office, London. (Middle East No. 66.)
SCOTT, H. 1939. Some considerations regarding the natural history of the Western Aden

Protectorate and the Yemen. Proc. Linn. Soc. Lond., 151 (3) : 164-171.
SMITH, E. A. 1894. On a small collection of land and freshwater shells from Oman, Arabia.

Proc. malac. Soc. Lond., 1 (4) : 141-142.

WALKER, B. 1914. Notes on the Ancylidae of North Africa. Nautilus 27 (10) : 113-117.
WITENBERG, G. & SALiTERNiK, Z. 1957- Studies on vectors of Schistosoma in Israel. Butt. res.

Council Israel, 6 B (1-2) : 107-141.
WRIGHT, C. A. i959. The application of paper chromatography to a taxonomic study in the

molluscan genus Lymnaea. J. Linn. Soc. Lond. Zool., 44 (296) : 222-237.

I959&- A note on the distribution of Bulinus senegalensis. West Afr. med. J., 8 (4) : 142-148.

1962. The significance of infra-specific taxonomy in bilharziasis. Ciba Found. Symp.

Bilharziasis.
1963. The freshwater gastropod molluscs of Angola. Bull. Brit. Mus. (Nat. Hist.) Zool.,

10 (8).
1963. Schistosomiasis in the Western Aden Protectorate ; a preliminary survey, Trans.

Roy. Soc. trop. Med. Hyg. 57.

& Brown, D. S. 1962. On a collection of freshwater gastropod molluscs from the Ethio-
pian highlands. Bull. Brit. Mus. (Nat. Hist.) Zool, 8 (6) : 285-312.

PLATE I

FIGS. 1-3. Biomphalaria rueppelli. Holotype of Planorbis arabica.
FIGS. 4-6. B. rueppelli. Sumayfah, Upper Aulaqi.
FIGS. 7-8. B. rueppelli. Tarbak, Upper Aulaqi.
FIGS. 9-14. Gyraulus convexiusculus. Tarbak, Upper Aulaqi.
FIGS. 15-17. G. convexiusculus. Kandahar, Afghanistan.
FIGS. 18-20. Lymnaea truncatula. Tarback, Upper Aulaqi.

(All figures X 4.)

Hull. B.M. (N.H.) Zoo/. 10, 4

PLATE i

PLATE

FIGS. 1-4. Bitlinits sericinns. Tarbak, Upper Aulaqi.
FIGS. 5 and 6. B. reticulatus. Rassaiss, Upper Aulaqi.
FIGS. 7 and 8. B. reticulatus. Marbum, Upper Aullqi.
FIGS. 9 and 10. B. beccari. Bateis, Abyan, Lower Yafai.
FIGS. 11-14. B- beccari. Dirgag, Abyan, Fahdli.

FIGS. 15-17. Ancylns fluviatilis. Holotype of Psendancylus argentens.
FIGS. 18 and 19. A. fluviatilis. Tarbak, Upper Aulaqi.
FIGS. 20-22. A. fluviatilis. Sumayfah, Upper Aulaqi.
FIG. 23. Melanoides tuberculata. Musemir, Haushabi.

(All figures X 4.)

Bull. B.M. (N.H.) Zool. 10,4

P L A T E 2

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

OBSERVATIONS ON THE

CHAETOTAXY OF THE LEGS IN

THE FREE-LIVING GAMASINA

(ACARI : MESOSTIGMATA)

G. OWEN EVANS

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 5

LONDON: 1963

OBSERVATIONS ON THE CHAETOTAXY OF

THE LEGS IN THE FREE-LIVING GAMASINA

(ACARI : MESOSTIGMATA)

G. OWEN EVANS

Department of Zoology, British Museum (Natural History)

Pp. 275-303 ; 15 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 5

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. 5 of the Zoological
series. The abbreviated titles of the periodicals cited
follow those of the World List of Scientific Periodicals.

Trustees of the British Museum, 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued May, 1963 Price Thirteen Shillings

OBSERVATIONS ON THE CHAETOTAXY OF

THE LEGS IN THE FREE-LIVING GAMASINA

(ACARI : MESOSTIGMATA)

By G. OWEN EVANS

CONTENTS

Page

SYNOPSIS ............ 277

INTRODUCTION ........... 277

TERMINOLOGY ........... 277

ONTOGENETIC DEVELOPMENT OF THE CHAETOTAXY OF THE LEGS IN Perga-

masus (PARASITIDAE) ......... 278

CHAETOTAXY OF THE LEGS IN ADULTS OF THE FREE-LIVING GAMASINA . 285

DISCUSSION ........... 301

REFERENCES ........... 303

SYNOPSIS

Studies on the chaetotaxy of the legs of the free-living Gamasina have shown that segmental
chaetotactic patterns provide valuable taxonomic criteria for the classification of the group. The
application of leg chaetotaxy to the classification of the Gamasina at species, genus and family
levels is discussed. A system of setal nomenclature is introduced for individual leg segments.

INTRODUCTION

During the last ten years, studies on the chaetotaxy of the idiosoma and the gnatho-
soma of the Mesostigmata have resulted in the introduction of many new taxonomic
criteria for the classification of the group. These studies are now extended to the
ambulatory appendages which have been largely ignored in both descriptive and
classificatory works on the Mesostigmata.

The present contribution is restricted to an account of the leg chaetotaxy in the
free-living Gamasina. My studies are as yet incomplete, particularly as regards the
ontogenetic development of the chaetotaxy, but they do indicate that leg chaetotaxy
can provide taxonomic characters of comparable value to those based on the chaeto-
taxy of the gnathosoma and idiosoma. It is hoped that this paper will stimulate
other acarologists to investigate further this interesting aspect of the morphology of
the Mesostigmata.

TERMINOLOGY

For the purpose of setal nomenclature, each segment of the legs is considered to
have four setae-bearing surfaces, namely, a dorsal (d), a ventral (v), an antero-lateral
(al) and a postero-lateral (pi). The anterior and posterior faces of the segment (or
leg) refer to the positions adopted when the leg is extended laterally, that is, at right
angles to the longitudinal axis of the idiosoma. The setae on the dorsal and ventral

ZOOL. 10, 5 15$

278 G. OWEN EVANS

surfaces of a segment are divided into an anterior and a posterior series according to
their positions relative to the anterior and posterior faces of the segment. The setae
are numbered from the distal to the proximal end of the segment. When the distri-
bution of the setae is such that their division into an anterior and posterior series is
not possible, they are simply referred to as dorsal (d) and ventral (v) setae. The
unpaired setae which appear during ontogeny on the dorsal and ventral surfaces of
tarsi II-III are termed medio-dorsal (md) and medio-ventral (mv), respectively.
The chaetotaxy of tarsus I is not considered in the present study.

It is often necessary to refer to the number of setae on a particular segment of
the leg in the course of descriptive work. A reference to the total number of setae is
of limited value since it gives no indication of the distribution of the setae on a seg-
ment. Considerably more information is given when the numerical data are divided
into four sections to coincide with the four setae-bearing faces of the segment. Thus,
the data may be expressed as follows :

(antero-laterals / I postero-laterals) .

/ ventrals /

Each segmental formula is accompanied by a specific reference to the segment and
to the postembryonic developmental stage to which the data apply. For example,
Femur I, L (2 / f / 2) implies that the femoral segment of leg I in the larva bears
two antero-laterals, four dorsals, two ventrals and two postero-laterals. The abbrevi-
ations L, P, D, A (<^ or $) are used, respectively, for larva, protonymph, deutonymph
and adult (male and female). This type of segmental formula has been found to be
adequate for representing the chaetotactic data of the trochanter and femur but
not of the genu and tibia. In many instances, it has been found that the total number
of setae on the dorsal and ventral surfaces of the genu and tibia may be the same in
widely separated groups of species although the actual chaetotactic patterns of these
segments may be quite different owing to differences in the numbers of setae in
the anterior and posterior series of the dorsal and/or ventral surfaces of the segments.
These differences in chaetotactic patterns may be shown numerically by extending
the above chaetotactic formula to include the number of setae in the anterior and
posterior series of the dorsal and ventral surfaces, thus,

, , / antero-dorsals postero-dorsals /

(antero-laterals / -, - - / postero-laterals).

/ antero-ventrals postero- ventrals /

This formula is used for the chaetotaxy of the genua and tibiae throughout this paper.
In the diagrammatic representation of segmental chaetotaxy, the open cicrcles
(o) refer to dorsal setae, the black dots ( ) to ventral setae and the half-open circles
(c) to lateral setae.

THE ONTOGENETIC DEVELOPMENT OF THE CHAETOTAXY OF THE
LEGS IN PERGAMASUS (PARASITIDAE)

The following account of the development of the leg chaetotaxy in the larval and
protonymphal stages of Pergamasus applies to the majority of the free-living Gama-

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 279

sina I have examined. The adult chaetotaxy which is determined at the deuto-
nymphal stage is considerably more variable.

al o

FIGS, i a-l. Diagrammatic representations of the chaetotaxy of the coxae and trochanters
of legs I-IV in Pergamasus (Parasitidae) . a-d, chaetotaxy of the coxae of legs I-IV.
e-h, chaetotaxy of the trochanters of legs I-IV in the larva and protonymph. i-l, chae-
totaxy of trochanters I-IV in the deutonymph and adult.

ad, antero-dorsal seta ; al, antero-lateral seta ; av, antero-ventral seta ; pd, postero-
dorsal seta ; pi, postero-lateral seta ; pv, postero-ventral seta; I-IV, legs I-IV.

Coxae I-IV (Text-figs. ia-d)

The chaetotaxy of coxae I-III is determined at the larval stage ; each segment
bearing two ventral setae (av and pv). There is a remarkable constancy in the relative
positions of the setae. Coxa IV bears a single ventral seta in the protonymph and
in the succeeding developmental stages. This seta is possibly homologous to seta av
of coxae II and III but has moved to a median position on the segment. The migra-
tion of a seta to a median position on the dorsal, ventral and lateral surfaces of
a segment when the corresponding member of the pair is missing, is a common feature
of segmental chaetotaxy.

280 G. OWEN EVANS

Trochanters I -IV (Text-figs. ie-l)

Trochanter I has four setae in the larva and protonymph, comprising two laterals
(al and pi] and two ventrals (f>v l and pv z ) . Two setae are added to the segment in
the deutonymph, one dorsal (d) and one ventral (av). This chaetotaxy is retained
by the adult. The chaetotaxy of trochanter II in the larva and protonymph is similar
to that of trochanter I at those stages but in the deutonymph only one seta (av) is
added to the segment. Four setae occur on trochanter III in the larva and on tro-
chanters III and IV in the protonymph, and five setae on each of these segments in
the deutonymph and adult. The distribution of these setae is more variable than the
setae of trochanters I and II. This is probably due to the asymmetrical development
of the trochanters (and femora) of legs III and IV as the result of the posterior
direction of legs III and IV.

Femora I -IV (Text-figs. 2a-i)

Femora I and II have a completely different chaetotaxy from femora III and IV
as regards the number and the distribution of the setae.

Femur I in the larval and protonymphal stages bears ten setae four dorsals
(ad lt ad z , ad 3 andpd^, four laterals (al lt al 2 , pl^ and_/ 2 ) and two ventrals (v^ and v z ).
The deutonymphal complement of thirteen setae is formed by the addition of one
dorsal (pd 2 ) and two ventrals (v 3 and i> 4 ). This number is retained by the adult. There
are no setae situated proximal of the lyriform fissure. Setae pd 2 and ad 3 assume a
median position on the dorsal surface of the segment.

Femur II in the larva (Text-fig. 2b) has seven setae (ad^ ad 2 , pd v pd 2 , v lt al^ and
pli). In the protonymph (Text-fig. 20) a single ventral seta (v z ) is added to the seg-
ment whilst in the deutonymph (Text-fig. 2g) three setae appear, one lateral (al z ),
one dorsal (ad 3 ) and one ventral (v 3 ] . Setae al 2 and v 3 are situated proximal to the lyri-
form fissure. The chaetotaxy of the female resembles that of the deutonymph but
in the male the segment becomes swollen and seta v^ and often v 2 hypertrophy and
form spur-like structures. The main femoral spur develops from seta v v and the
axillary spur from seta v 2 . Seta v 3 in the male moves to a more distal position and
never lies proximal to the lyriform fissure.

Femur III in the larva and protonymph carries five setae, comprising three dorsals
(ad lt ad 2 3.ndpd^), one lateral (al) and one ventral (v). Only one seta (pd z ) is added to
the segment in the deutonymph. The chaetotaxy of femur IV in the protonymph
resembles that of femur III at that stage except for the absence of the ventral seta
(v) which appears with pd z in the deutonymph. The deutonymphal chaetotaxy of
femora III and IV is retained by the adult.

Genua I-IV (^a-h)

Genu I carries eight setae in the larva and protonymph four dorsals (ad lt ad 2 ,
pd t , pd z ), two laterals (a/ x and^/j) and two ventrals (av-^ and^Wj). This number is
increased to thirteen in the deutonymph by the appearance of one ventral (av z ), two
laterals (al z and pl z ) and two dorsals (ad 3 and pd 3 ). Seta av z migrates medially so
that the three ventral setae form an inverted triangle. Six setae occur on genu II

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 281

8.0. " S -

~S -00
>

^ lo

-60

JSi

-^~

"SP

-e-

II 3
|| a

d o O

gs. rt

OCL

a CL

>-
c

Qo<

ZOOL. 10, 5

282

G. OWEN EVANS

-DO
o

8 -

-50
a

-00

-e-

"go

-e-

"go

> o"^*

2>"

o 0"

o-S.

DO.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 283

fel

-e-

-oO'

T3Q >

O W O

"go

-e-

& *?

O ^

^"H"o

O c: i

__J DCL

1*8

O 4J

PL,

CO O

CL.

-e-

.n-tg.
<u

be O

o .

284

G. OWEN EVANS

in the larva and protonymph comprising ad lt ad z , pd lf pd z , al v and pl v There are
no ventral setae. Five setae appear in the deutonymph, one dorsal (ad 3 ), two lateral
(al z and pl z ) and two ventral (av-^ and pVj). The chaetotaxy of genu III in the larva
and protonymph is similar to that of genu II at those stages but only two setae,
the ventrals av-^ and pv v are added in the deutonymph. Genu IV bears five setae
(ad^ ad z , pd lt pd z and al : ) in the larva and protonymph. The deutonymph adds
five setae (av v pv lt al z , pl^ and pd 3 ) .

The deutonymphal chaetotaxy of the genua is retained by the adult. In the males
seta av-L hypertrophies and takes the form of a spur-like structure.

Tibiae I -IV (Text-figs, ^a-h)

Tibia I in the larva and protonymph bears eight setae (ad^ ad z , pd v pd z , al v pl lt
av-i and PVJ). Six setae are added proximally in the deutonymph. These comprise
two dorsals (ad 3 and pd 3 ), two laterals (al z and pl z ) and two ventrals (av z and pv z ).
The deutonymphal complement of setae is retained by the adult. The tibiae of
legs II and III in the larva and protonymph bear seven setae. Their chaetotaxy
resembles that of tibia I in these stages except for the absence of ad z and the migration
of ad l to a median position on the segment. Tibia II in the deutonymph shows an

Q b C

FIGS. $a-c. Diagrammatic representations of the chaetotaxy of tarsi II-IV in the larva
(a), the protonymph (b) and the deutonymph (c) of Pergamasus.

md, medio-dorsal seta ; mv, medio-ventral seta. Other abbreviations as in Fig. i.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 285

increase of three setae a dorsal (ad 2 ) and two laterals (al 2 and pl 2 ). Only one seta
(al 2 ) is added to tibia III in the deutonymph. The chaetotaxy of tibia IV in the
protonymph is similar to that of tibiae II and III at that developmental stage. In
comparison with tibia III, however, three setae appear on the segment in the deuto-
nymph (setae al 2 , pd 3 and pl 2 ) . In the male seta av may hypertrophy and form a
spur-like structure.

Tarsi II-IV (Text-figs. $a-e)

The chaetotaxy of tarsi II and III in the larva consists of sixteen setae three
pairs of dorsals, three pairs of laterals and two pairs of ventrals. In the protonymph
an unpaired seta (md) is added in the distal half of each of tarsi II-IV and a further
unpaired seta (mv) appears ventrally in the proximal half in the deutonymph making
a total of eighteen setae. This number is retained by the adult.

The chaetotactic formulae for the trochanters, femora, genua and tibiae of legs
I-IV in the adults and immature stages of Pergamasus are as follows :

Trochanter

L, P, I (i-S-i), II (i--i), HI (i-J-o), P, IV (i-J-o).
D, A, I (i-J-i), II (i-J-i), III (i-J-o), IV (i-J-o).

Femur

L, I (2-1-2), II (!-$-!), HI (l-f-0).

P, I (2-|-2), II (I-|-I), HI (I-J-O). IV (1-8-0).

D, A, I (2-|-2) II (2-1-1), III (i-*-o), IV (i-f-o).
Genu

L, P, i (i-f, f-i), ii (1-8, s-i), ni (i-s, s-i), P, iv (i-8. -o).

D, A, I (2-|, f-2), II (2-f, f-2), III (I-f , f-1), IV (2-f , f-1).

Tibia

T p T l-r 22 r \ TT / T 12 T \ TTT / T 12 T \ p TW /T * 2 T^

-L, r, i ^i T , T i;, 11 \L T , T ij, 111 ^i T , T L), r, iv ^i T , T i;.
D, A, I (2 , 2), II (2 f , f 2), III (2 J, f i), IV (2 \, f 2).

CHAETOTAXY OF THE LEGS IN THE ADULTS OF THE
FREE-LIVING GAMASINA

Coxae I-IV

The chaetotaxy of the coxae is constant in the free-living Gamasina ; the details
given for the developmental stages of Pergamasus (p. 279) applying throughout the
group. The form of the coxal setae varies considerably. In Scarabaspis (Eviphididae),
for example, both setae of coxa I and seta pv of coxa II take the form of sclerotized
oval protuberances.

286 G. OWEN EVANS

Trochanters I -IV

Trochanter I in the majority of the species I have examined bears six setae two
laterals (al and pi], three ventrals (av, pv lt pv 2 ) and one dorsal (d). The Microsejidae,
however, retain the larval complement of four setae in the adult, and in the families
Macrochelidae and Phytoseiidae (with few exceptions) and the genus Thinoseius
(Eviphididae) there are only five setae on this segment ; the dorsal seta (d) being
absent. The relative positions of the setae are remarkably constant.

The chaetotaxy of trochanter II is essentially the same in all the species investi-
gated and comprises five setae (al, av v pv lt pv 2 and pi). With the exception of
Microgynium (Microsejidae), the occurrence of five setae on trochanters III and IV
is also constant although the distribution of the setae on the segments is variable.
In Microgynium, trochanter III has only four setae, thus retaining the larval pattern.

The setae on the trochanters appear to be concentrated towards the external face
of the segment when the legs are in the normal position for ambulation, that is, on
the posterior surfaces of trochanters I and II and the anterior surfaces of trochanters
III and IV.

Femur I
Four types of chaetotaxy of femur I have been observed :

X-type (2-f-i).

The chaetotaxy comprises setae al lt al z , ad lf ad 2 , pd lt pd 2 , pl^ v^ v 2 and v 3 (Text-fig.
6a). The three ventral setae form an inverted triangle. I have observed this chaeto-
tactic pattern only in the Microsejidae (Microsejus and Microgynium}.

Xl-type (2-1-2).

The following setae are present : al lt al 2> ad lt ad 2 , ad 3 , pd lf pd 2 , pl v pl z> v l and v 2
(Text-fig. 66). This type occurs in the genera Sejus, Platyseius and Plesiosejus
(Aceosejidae) and Thinoseius (Eviphididae).

Xll-type (2 f 2).

Seta v s is added to those occurring in the Xl-type making three ventral setae
(Text-fig. 6c). This chaetotaxy occurs in the families Phytoseiidae and Ameroseiidae,
and in the genera Asca and Halolaelaps (Rhodacaridae) and the genera Lasioseius,
Proctolaelaps, Melichares, Leioseius, Arctoseius, iphidozercon and Zerconopsis
(Aceosejidae).

XHI-type (2-1-2).

This type is characterized by having four ventral setae (Text-fig. 6d). It is the
most widely distributed chaetotactic pattern in the suborder and has been observed
in the families Parasitidae, Veigaiaidae, Arctacaridae, Laelaptidae, Zerconidae,
Epicriidae, Macrochelidae and Pachylaelaptidae, and in the genera Eviphis, Alliphis,
Pelethiphis and Scardbaspis (Eviphididae), and Rhodacarus, Rhodacarettus, Eury-
parasitus, Gamasellus, Ologamasus and Gamasiphis (Rhodacaridae).

CHAETOTbXY OF THE LEGS IN THE FREE-LIVING GAMASINA 287

al,

ad,

av,
av 2

ad 2

pi.

a I,

Ol:

) ad,

pdi
o

V 2

ad 2

> ad 3
o 3

pd 2

pl:

al,

ad,

pd,

o .

ad 2
o

(

V 2

pd 2

ad 3

V 3 (

pi.

al,

o P d
ad,

9
ad,

ad 3
o

op! 2

FIGS. >Or-d. Diagrammatic representation of the chaetotaxy of femur I in the free-living
Gamasina, a, X-type (Microsejidae). b, Xl-type (Aceosejidae). c, Xll-type (Amero-
seiidae). d, Xlll-type (Eviphididae) .
Abbreviations as in Fig. i.

Genu I

I have observed five types of chaetotactic patterns of germ I ; the number of
setae ranging from ten to thirteen.

288

G. OWEN EVANS

X-type ( 2 -f , f-2).

The following setae are present : ad v ad 2 , pd lt pd z , av lt pv lt al v al 2 , pl lt pl z (Text-
fig. 7). This type is characteristic of members of the families Phytoseiidae and
Microsejidae.

ad,

all <

) Pd,<

>pl|

o pd.

av, (}

PV.

al| <

) a ;' , <

pi,

ad 2

a! 2 <

> ad 2

>p! 2 al2<

pd,

f^i_

UI 2

pd 2

ad,
av, pd,

ad 3

a al ' (

pv,
av 2 r

pi, b

ad 2

> P I 2

ad,

ad 3

a o' P dl

O ,

ah <

> ^o' <

>pl, Q all <

av.

ipll

av,

O pv,

pVi
Of

ad 2 o

2 pd 2

a! 2 i

' o l d '

^P^ a bi

o pd 3

>pk

ad, pdl

ads O

FIGS. ya-. Diagrammatic representations of the chaetotaxy of genu I in the free-living
Gamasina. a, X-type (Phytoseiidae) . b, Xlb-type (Macrochelidae) . c, XIa-type (Evi-
phididae). d, XH-type (Aceosejidae). e, XHI-type (Parasitidae).
Abbreviations as in Fig. i.

XIa-type (i-|, f- 2 ).

This chaetotactic pattern differs from all others on genu I in having a single

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 289

antero-lateral (al-^ seta. The setal complement is al v ad v ad z , ad z , pd v pd 2 , pl v pl z ,
av lt av z and pv^ (Text-fig. 7^) . I have only observed this pattern in the Eviphididae.

Xlb-type (2-f, f-2).

This differs from the X-type in having an additional dorsal seta, ad z (Text-fig. 7&).
It is characteristic of the Macrochelidae.

XH-type (2-|, |-2).

This chaetotactic pattern is similar to Xlb except for the presence of an additional
ventral seta (av z ). The three ventral setae form an inverted triangle owing to the
migration of seta av z to a median position on the segment (Text-fig. 7^). It is found
in the Ameroseiidae and in the genera Arctoseius, Iphidozercon and Zerconopsis
(Aceosejidae), Asca, Digamasellus and Halolaelaps (Rhodacaridae) and Pachyseius
(Pachylaelaptidae) .

Xin-type (2-f, f-2).

The XUI-type is characterized by having six dorsal (ad-^-ad^ and pdf-pd a ), in
addition to two antero-laterals, two postero-laterals and three ventrals (Text-fig. 70).
I have observed this genual pattern in the families Parasitidae, Veigaiaidae, Arcta-
caridae, Laelaptidae, Zerconidae and Epicriidae, and in the genera Lasioseius
Proctolaelaps, Melichares, Leioseius, Sejus, Platyseius and Plesiosejus (Aceosejidae),
Pachylaelaps and Olopachys (Pachylaelaptidae) and Rhodacarus, Rhodacarellus,
Euryparasitus, Gamasellus, Ologamasus and Gamasiphis (Rhodacaridae).

Tibia I

The X-XIII-types of chaetotaxy observed on genu I also occur on tibia I as well
as a XlV-type characterized by having four ventral setae.

X-type (2-f, f-2).

This type (Text-fig. 8a) is found in the families Phytoseiidae and Microsejidae.

Xl-type (i-f , f-2).

Only the XIa-type of chaetotaxy of genu I occurs on the tibia and as in the case of
the genu it appears to be restricted to members of the family Eviphididae (Text-fig.
86).

XH-type (2-|, f-2).

This type (Text-fig. 8c) occurs in the families Macrochelidae, Pachylaelaptidae
and Ameroseiidae, and in the genera Arctoseius, Iphidozercon, Zerconopsis and
Plesiosejus (Aceosejidae), and Digamasellus and Halolaelaps (Rhodacaridae).

XIH-type (2-f, f-2).

I have observed this chaetotactic pattern (Text-fig. 8d) in the free-living Laelaptidae
and in the genera Lasioseius, Proctolaelaps, Melichares, Leioseius, Sejus, Platyseius
and Plesiosejus (Aceosejidae), and Asca (Rhodacaridae).

2QO

G. OWEN EVANS

a! 2 o

ad,
o

pa.

Opli

ad 2

al.

a! 2 o

av, pv.

opli

t %

ad,

al,

pd,
o

ad 2 o

pi,

ad n

ad i
pd,

av,

av 2
o
ad 2

ads

pd 2

PV.

ad 3 o

p'd

FIGS. 8a-e. Diagrammatic representations of the chaetotaxy of tibia I in the free-living
Gamasina. a, X-type (Phytoseiidae) . b, Xl-type (Eviphididae). c, XH-type (Aceoseji-
dae). d, XHI-type (Aceosejidae). e, XlV-type (Parasitidae) .
Abbreviations as in Fig. i.

XlV-type (2-|, |-2).

The four ventral setae arranged in the form of a rectangle readily distinguishes this
pattern from the others found on tibia I (Text-fig. Se). It occurs in the families Para-
sitidae, Veigaiaidae, Arctacaridae, Zerconidae, Epicriidae and in the genera Rhoda-
carus, Rhodacarellus , Cyrtolaelaps, Euryparasitus , Gamasellus, Ologamasus and
Gamasiphis (Rhodacaridae) .

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 291

-e-

-oO
a

(3P

O
>-> o
^> c3

^Tab

If 8,

s*>

s~~

4) f-H JH

^
<D <U

rt fa

-g'o "go

-e-

"8.

-S s -s -a

5 o .&-^TJ

s|ll

fi tS - -^ "3

3 a S 1 -

Plli

lw| 2

O 3

Sfr.
h*^

S.
Q

rt
bb ts

IH V

a i- 1

3
H fa

s
.a

* a

II sal

^ll^5

a 5 ^

| fa ^ ^

I O <^T "-i '5J*

bo "cL o

co '> rt J >

O ^H 'd *t3 W "s

2 Q2 G. OWEN EVANS

Femur II

The following four types of chaetotactic pattern have been observed on this
segment :

IX-type (2-f-i).

This type has the following chaetotaxy ; al v , al z , ad v ad z , pd lt pd 2 , pl v av { and pv^
(Text-fig. go). It occurs in the Microsejidae.

Xa-type (2-|-i).

The chaetotaxy comprises ad v ad z , pd lt pd z , al v al z , pl lt v lt v 2 and v z . This type
has been found in the family Ameroseiidae.

Xb-type (2-f-i).

The chaetotaxy differs from the Xa-type in the presence of five dorsal setae (ad s
additional) but only two ventral setae. It is characteristic of the Phyto-
seiidae, the genera Sejus, Platyseius and Plesiosejus (Aceosejidae), and also occurs
in the genus Thinoseius (Eviphididae). In the latter, the ventral setae are widely
separated so that one occurs in the proximal and the other in the distal half of the
segment (Text-fig, gc). In the Phytoseiidae, however, both ventral setae are usually
situated in the proximal half of the segment (Text-fig, gb).

Xl-type (2-|-i).

This type has five dorsal setae and three ventrals in addition to the two antero-
laterals and one postero-lateral (Text-fig, gd). It is the most widely distributed
chaetotactic pattern of femur II in the Gamasina and has been observed in the families
Parasitidae, Veigaiaidae, Arctacaridae, Rhodacaridae, Zerconidae, Epicriidae,
Laelaptidae, Macrochelidae, Pachylaelaptidae, Aceosejidae (with the exception of
Sejus, Platyseius and Plesiosejus) and Eviphididae (excluding Thinoseius).

Seta al 2 is always situated proximal to the lyriform fissure in all four types of
chaetotaxy.

Genu II
Genu II carries seven, nine, ten or eleven setae.

VH-type ( 2 -, g-i).

In this type there are no ventral setae (Text-fig. 100) ; the chaetotaxy comprising
setae al lt al z , ad lt ad 2 , pd lt pd 2 and pl v It occurs in the majority of the Phytoseiidae.

IX-type (2-f , f-i).

I have, to date, only observed this pattern in Amblyseius graminis Chant (Phyto-
seiidae). The chaetotaxy shows an addition of two ventral setae (av-^ and pvj to
the VH-type (Text-fig. io&).

Xa-type (2-f, -2).

This differs from the IX-type in having two postero-lateral setae (Text-fig. 100).
This pattern is characteristic of the Microsejidae.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 293

a! 2 (
a! 2 (

ad,
> ad,

pd,
O

o <

pd 2

ipl

ali I

ad,
O

pd,
o

ad,
O

pd,

N av,

, i

av,

pv * U '

o
ad 2

> o

at<

) _

pd 2 (

ad 2

pd 2

O
ad 3

ad,

pd,

pv,

2
Q

fi pv,
pd 2

ad 3

FIGS. loa-f. Diagrammatic representations of the chaetotaxy of genu II in the free-
h'ving Gamasina. a, VH-type (Phytoseiidae). b, IX-type (Amblyseius graminis Chant :
Phytoseiidae) . c, Xc-type (Sejus : Aceosejidae). d, Xb-type (Aceosejidae). e, Xa-type
(Microsejidae). /, XI -type (Parasitidae) .
Abbreviations as in Fig. i.

Xb-type (2-f, 5-i).

This type (Text-fig. iod) which occurs in the genera Arctoseius, iphidozercon and
Zerconopsis (Aceosejidae) is characterized by having two antero-laterals (al^ and al 2 ),
five dorsals (ad^ 3 and pd^), two ventrals (av^ and PV-^) and one postero-lateral

Xc-type ( 2 -f, -2).

The Xc-type has two postero-laterals but only one ventral seta (Text-fig. loc).
It occurs in the genera Sejus, Platyseius and Plesiosejus (Aceosejidae).

Xd-type (i-f , |-2).

This pattern differs from all others on genu II in having only one antero-lateral
seta. I have observed this chaetotaxy in the genus Thinoseius (Eviphididae) only.

294 G - OWEN EVANS

XI-type (2-1 f-2).

The XI-type (Text-fig. io/) is by far the most widely distributed chaetotactic
pattern of this segment and occurs in the families Parasitidae, Veigaiaidae, Arcta-
caridae, Rhodacaridae, Zerconidae, Epicriidae, Macrochelidae, Pachylaelaptidae,
Laelaptidae, Ameroseiidae and Eviphididae (excluding Thinoseius) and in the genera
Lasioseius, Leioseius, Melichares and Proctolaelaps (Aceosejidae).

Tibia II
In the species I have examined, tibia II bears seven, nine or ten setae.

VH-type (i-f , i-i).

This type is characteristic of the Phytoseiidae and comprises setae ah, ad v ad z , pd lt
pl v av! andpvi (Text-fig,

IXa-type (i-f , f-2).

This type, observed only in Thinoseius (Eviphididae), has one antero-lateral seta

K).

IXb-type (2-1 f-i).

The chaetotaxy comprises al lt al z , ad v ad z , pd lt pd 2 , pl v av^ andpv 2 (Text-fig. n&).
This type occurs in the genera Arctoseius, iphidozercon and Zerconopsis (Aceosejidae).

IXc-type (2-f , 1-2).

The chaetotaxy comprises setae al v al z> ad v ad z , pd v pl lt pl 2 , av lf pv (Text-fig, nc).
I have observed this type in the genera Platyseius and Plesiosejus (Aceosejidae).

IXd-type (2-1, |-2).

This differs from the IXc-type in the setation of the dorsal surface of the segment ;
the three setae being ad lt pd 1 and pd 2 (Text-fig, n^). It occurs only in the Micro-
sejidae.

X-type (2-f, f-2).

The following setae are present : al lt al z , ad^ ad z , pd v pd 2 , pl v pl 2 , av^ and pv t
(Text-fig. lie). This chaetotactic pattern is present in the families Parasitidae,
Veigaiaidae, Arctacaridae, Rhodacaridae, Zerconidae, Epicriidae, Laelaptidae,
Ameroseiidae, Macrochelidae, Pachylaelaptidae, Eviphididae (excluding Thinoseius)
and in the genera Lasioseius, Leioseius, Melichares, Proctolaelaps and Sejus (Aceoseji-
dae).

Tarsi II-IV

The chaetotaxy of tarsi II-IV is extremely constant throughout the free-living
Gamasina ; the details given for Pergamasus (p. 278) being applicable throughout.
There is some variation in the relative position of the setae due, in some instances,
to the elongation of the segment and in the form of setae ad and pd^.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 295

al,

al 2 *

'ad,
O

pd,
O

av,

ad.

av, pv,

adi

pd,
O

> av,

(
PV,

pd 2

ad 2

FIGS, iia-e. Diagrammatic representations of the chaetotaxy of tibia II in the free-
living Gamasina. a, VH-type (Phytoseiidae) . b, IXb-type (Zerconopsis : Aceosejidae).
c, IXc-type (Platyseius : Aceosejidae). d, IXd-type (Microsejidae). e, X-type (Laelap-
tidae).

Abbreviations as in Fig. i.

Femur III

The number of setae on femur III is markedly constant. With the exception of
Thinoseius, all the species I have examined have six setae (Text-fig, ge). Owing to the
difference in shape of this segment in certain species, the distribution of the setae is
subject to some variation. In Thinoseius, seven setae are present ; the additional
seta apparently being developed ventrally (Text-fig, o,/).

Genu III

This segment bears six to ten setae.

29 6 G. OWEN EVANS

Vl-type (i-g, -i).

The following setae are present : al v ad lt ad z , pd v pd z and pl t (Text-fig. I2).
There are no ventral setae. I have seen this type only in the genus Phytoseius
(Phytoseiidae).

VH-type (i-, g-i).

The chaetotaxy comprises setae al lt ad lt ad 2> pd lt pd z , pl and av^ (Text-fig. 126).
It occurs in members of the family Macrochelidae and in the genera Arctoseius,
Iphidozercon and Zerconopsis (Aceosejidae), Pachylaelaps and Olopachys (Pachylae-
laptidae), Thinoseius (Eviphididae) and Iphiseius and Typhlodromus (Phy toseiidae) .

Villa-type (i-J, f-i).

The chaetotaxy differs from the VH-type in the addition of pv- L (Text-fig. I2c).
This chaetotactic pattern occurs in the Eviphididae (excluding Thinoseius).

VHIb-type (2-1 8-i).

This chaetotactic pattern has been found in the genera A sea and Saprolaelaps
(Rhodacaridae), Sejus, Platyseius and Plesiosejus (Aceosejidae), Pachyseius (Pachy-
laelaptidae) and Epicriopsis (Ameroseiidae). The eight setae comprise ah, al 2 , ad v
ad z> pd, pd z , pl and av t (Text-fig, xzd}.

IXa-type (2-}, f-i).

Setae o/ lf al z , ad v ad 2 , pd v pd z , pl lf av and pv l are present in this type (Text-fig.
120). This pattern occurs in the families Parasitidae, Veigaiaidae, and in the genera
Rhodacarus, Rhodacarellus, Cyrtolaelaps, Euryparasitus, Gamasellus, Hydrogamasus,
Ologamasus and Gamasiphis (Rhodacaridae), Lasioseius, Leioseius, Melichares and
Proctolaelaps (Aceosejidae), Hypoaspis, Coleolaelaps and Ololaelaps (Laelaptidae)
and Neocypholaelaps (Ameroseiidae).

IXb-type (2-f, -2).

I have observed this type only in the Microsejidae (Text-fig. I2/).

X-type (2-|, ?-2).

The ten setae comprise al lt al 2 , ad^, ad 2 , pd lf pd z , pl lt pl z , av^ and^Vj (Text-fig. ~L2g}.
It occurs in the families Arctacaridae, Zerconidae and Epicriidae, and in the genera
Ameroseius and Kleemannia (Ameroseiidae).

Tibia III

The following three types of chaetotactic pattern have been observed on this
segment :

VH-type (i-l, }-i).

This type occurs in the families Phytoseiidae, Eviphididae, Macrochelidae and
Pachylaelaptidae, and in the genera Arctoseius, Iphidozercon and Zerconopsis
(Aceosejidae). The chaetotaxy comprises a/ 1} ad v pd v pd 2 , pl lt av and pv (Text-fig.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 297

ad, pd,

O O

a!, <

)

>pl,

al, <

av,

o o

ad 2 pd 2

ad, pd,
o o

ad, pd,
>

ad (

av, pv,
)

ipl.

av, (

>pl|

o<L <^

a) 2 i

>Q d ' P d 2

ad,

av, py,

pd,

_ o
H

al,(

ad, pd,
) O O (

p|, al, <

o I

> ad, p d, <

all i

) av, p v ,

avi

a! 2 <

ad 2 o

> Pd 2

a! 2 <

> o o

ad 2 pd 2

1C 1 ^ 2 '

>0 , 9 C

aa 2 pa 2

FIGS. i2a-g. Diagrammatic representations of the chaetotaxy of genu III in the free-
living Gamasina. a, Vl-type (Phytoseius : Phytoseiidae) . b, VH-type (Macrochelidae) .
c, Villa-type (Eviphididae) . d, VHIb-type (Asca : Rhodacaridae) . e, IXa-type
(Parasitidae). /, IXb-type (Microsejidae). g, X-type (Zerconidae) .
Abbreviations as in Fig. i.

2Q8

G. OWEN EVANS

ad,

pd,

al,<

) av,

pv, (

pd 2

ad,

pd,
O

al, <

.
> av,

PV, <

a) 2 <

pd,

ad.
)

pd,

* <

av,

pv,

)

o (

pd 2

FIGS. i3-c. Diagrammatic representations of the chaetotaxy of tibia III in the free-
living Gamasina. a, VH-type (Phytoseiidae) . b, VHI-type (Parasitidae) . c, IX-type
(Microsejidae).
Abbreviations as in Fig. i.

VHI-type (2-\, f-i).

There is an addition of one antero-lateral seta to the VH-type ; the seta being
tf/ 2 (Text-fig. 136). This pattern is present in the families Parasitidae, Veigaiaidae,
Rhodacaridae, Laelaptidae, Ameroseiidae (excluding Kleemannia) and in the genera
Lasioseius, Leioseius, Melichares and Proctolaelaps (Aceosejidae).

IX-type (2-1, f-2).

Seta_/>/ 2 is additional to the VHI-type (see Text-fig. I3c). It occurs in the families
Arctacaridae, Zerconidae, Epicriidae and Microsejidae and in the genus Kleemannia
(Ameroseiidae).

Femur IV

The presence of six setae on this segment is extremely constant in the free-living
Gamasina ; the only exception being members of the Microsejidae. In the majority
of the species I have examined the chaetotaxy comprises al v ad v ad 2 , pd lt pi and vi
(Text-fig, gh) but there are variants of this pattern as, for example, in the aceosejid
Plesiosejus (Text-fig. gg) which has four dorsals but no postero-lateral setae. The
microsejids have seven setae on this segment (al v ad^, ad 2 , pd v pd 2 , av 1 and^vj.

Genu IV
Five types of chaetotaxy of genu IV have been observed.

Vl-type (i-f, -o).

The following setae are present : al it ad lt ad 2 , pd v pd z and av (Text-fig. 140;).
This pattern is found in certain species of the genus Macrocheles s.lat. (Macrochelidae)
and has been used as a taxonomic character by Evans & Hyatt (1962).

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 299

VII-type (i-f, g-i).

This shows an addition of seta pi to the Vl-type (Text-fig. 146). It is found in the
families Pachylaelaptidae, Eviphididae and Phytoseiidae, in the genera Arctoseim,
iphidozercon and Zerconopsis (Aceosejidae), Digamasellus (Rhodacaridae) and certain
species of the Macrochelidae.

ad, pd,
o o

al, o av,

al,

ad,
o

pd,

)
av,

) O

p3 2 <

J
od 2

al, <

odi pd,
> o o

a! 2 <

> o Q
ad 2 pd2

ali

a! 2

ad,

av,

,
ad

<rP''

ad,

pd,
o

av,

pv,

pd 2 c

pd 3

pi,

FIGS, i^a-e. Diagrammatic representations of the chaetotaxy of genu IV in the free-
living Gamasina. a, Vl-type (Macrocheles : Macrochelidae). b, VII-type (Eviphididae).
c, VHI-type (Microsejidae). d, IX-type (Lasioseius : Aceosejidae). e, X-type (Parasiti-
dae).

Abbreviations as in Fig. i.

300 G. OWEN EVANS

VHI-type (2-|, S-i).

Setae al v al 2 , ad v ad z , pd v pd 2 , pl and av l are present (Text-fig. 14^). This type
occurs in the family Microsejidae.

IX-type (2-|, g-i).

Seta pd s is added to the VHI-type (Text-fig. 14^). It is found in the families
Laelaptidae, Ameroseiidae and in the genera Lasioseius, Leioseius, Melichares,
Proctolaelaps, Sejus, Platyseius and Plesiosejus (Aceosejidae), and in A sea and
Halolaelaps (Rhodacaridae).

X-type (2-f , f-i).

This type (Text-fig. 141?) has two ventral setae (pv^ being additional to the ventral
chaetotaxy found in the other types). It occurs in the families Parasitidae,
Veigaiaidae, Arctacaridae, Zerconidae, Epicriidae and in the genera Rhodacarus,
Rhodacarellus , Cyrtolaelaps , Euryparasitus, Gamasellus, Ologamasus, Hydrogamasus
and Gamasiphis (Rhodacaridae).

Tibia IV

Six types of chaetotaxy of tibia IV have been observed ; the number of setae on
the segment ranging from six to ten.

Vl-type (I-*, J-i).

This type comprising setae al lt ad lt pd v pd z , pl and av 1 occurs in members of the
family Phytoseiidae (Text-fig. 150).

VH-type (i-l -i).

The chaetotaxy of this type differs from VI in the addition of one postero-ventral
seta (Text-fig. 156). It is found in the families Macrochelidae, Pachylaelaptidae and
Eviphididae, and in the genera Arctoseius, iphidozercon and Zerconopsis (Aceosejidae)
and Digamasellus (Rhodacaridae).

VHI-type (2-1, f-i).

I have found this type in the genus Halolaelaps (Rhodacaridae) and the chaeto-
tactic pattern differs from the Vll-type in the addition of seta al z (Text-fig. 150).

IXa-type ( 2 -f , }-i).

The chaetotaxy comprises al lt al 2 , ad lt ad z , pd lt pd%, pl v av t a.ndpv l (Text-fig. 150).
The pattern occurs in the family Ameroseiidae.

IXb-type (2-1, f-2).

In contrast to type IXa, the segment bears only one antero-dorsal seta (adj) but
two postero-lateral setae (p^ and pl 2 ). This type of chaetotaxy occurs in the families
Zerconidae and Microsejidae (Text-fig.

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 301

ad o

ad,
O

pd,

r o

pd 3

al,

al 2

av,

.pi,

pd,

av,

ad,

pd,
O

PV.

pd 2

pl 2

pd,

) ^

pv,

adi

pa 2

ad 2

av,

pl ' 1-

all I

)

ad,

pv, c

pd 2

a! 2 4

)

(

pa;

pi.

FIGS. I5-/. Diagrammatic representations of the chaetotaxy of tibia IV in the free-
living Gamasina. a, Vl-type (Phytoseiidae) . b, VH-type (Eviphididae). c, VHI-type
(Halolaelaps : Rhodacaridae) . d, IXb-type (Microsejidae). e, IXa-type (Ameroseiidae).
/, X-type (Parasitidae) .

Abbreviations as in Fig. i.

X-type (2-1 |-2).

This type shows an addition of postero-dorsal seta (pd s ) to the setal complement of
IXb (Text-fig. I5/). It is found in the families Parasitidae, Veigaiaidae, Arcta-
caridae, Epicriidae and Laelaptidae and Aceosejidae (excluding genera under VII-
type), and in the genera Rhodacarus, Rhodacarellus, Cyrtolaelaps, Euryparasitus,
Gamasellus, Ologamasus, Hydrogamasus and Gamasiphis.

DISCUSSION

The wide variety of leg segmental chaetotactic patterns displayed by the free-living
Gamasina will undoubtedly provide useful morphological criteria for their classifica-
tion. Taxonomic characters based on chaetotactic patterns are readily definable quan-

302

G. OWEN EVANS

titatively and are thus suitable for mathematical analysis. The value of chaetotactic
patterns as indicators of phylogenetic relationships is difficult to assess although in
free-living forms their stability at species, genus and family levels is surprisingly
high. However, the nature of the ontogenetic development of the chaetotaxy,
particularly the importance of localized neoteny in determining specific segmental
patterns in specialized parasitic forms, allows for the widespread occurrence of con-
vergence in chaetotactic patterns. The results of the present study suggest that the
overall chaetotaxy of the legs, especially of legs I and II, may be of significance in
the classification of the suborder at familial and suprafamilial levels whereas differences
in the chaetotaxy of individual segments may be of more value at species and genus
level.

Segmental chaetotaxy of the legs appears to be subject to little or no intraspecific
variation. In the free-living Gamasina, differences in chaetotactic patterns between
species of the same genus are usually restricted to a single segment as, for example,
the presence of six or seven setae on genu IV in Macrocheles (Evans & Hyatt, 1962).
Specialized parasitic forms exhibit considerably more interspecific variation. Accord-
ing to Till (1962), species of the genus Androlaelaps show differences in the chaetotaxy
of tibia I, tibia III, genu III and genu IV. This variability is even more marked in
the endoparasitic forms where differences in overall leg chaetotaxy may be of impor-
tance only at species level. Differences in the chaetotaxy of one or two leg segments
are also evident at generic level and are particularly useful in providing " key
characters" for the differentiation of genera belonging to the same family. For
example, the chaetotactic patterns of genu III in the Ameroseiidae, of genua III and
IV in the Phytoseiidae and genu III in the Pachylaelaptidae.

An analysis of the leg chaetotaxy in confamilial genera indicates a greater stability
in the chaetotaxy of legs I and II than of legs III and IV. With few exceptions, the
familial concepts based on leg chaetotaxy support the present classification of the
free-living Gamasina based on other morphological criteria. The notable exception
is the family Rhodacaridae. The genera Rhodacarus, Rhodacarellus, Cyrtolaelaps ,
Euryparasitus, Gamasellus, Ologamasus, Gamasiphis, Hydrogamasus and Sessiluncus
form a natural assemblage which can be accommodated in the family Rhodacaridae,
whereas the genera Digamasellus, Asca and Halolaelaps do not appear to be confamilial
with the Rhodacarus-group. Their affinities appear to be with certain genera of the
Aceosejidae but whether they should be placed in that family or in a distinct family
(Digamasellidae) can only be decided by a comprehensive study of the complex.
Certainly the emphasis placed on the nature of the dorsal sclerotization of the idiosoma
in the classification is not justified in the light of the conclusions reached on the basis
of chaetotactic studies.

The criterion of leg chaetotaxy also gives interesting results when applied to the
classification of the Aceosejidae. At present this family is divided into two sub-
families (Evans, 1957), namely, the Aceosejinae and Platyseiinae. On the bases of
segmental patterns, however, three groups of genera may be recognized of which two
groups are more closely related to each other than to the third. The three groups are :

(a) Lasioseius, Leioseius, Proctolaelaps , Melichares and Zercoseius (Aceosejinae) ;

(b) Sejus, Platyseius and Plesiosejus (Platyseiinae) and (c) Arctoseim, Iphidozercon

CHAETOTOXY OF THE LEGS IN THE FREE-LIVING GAMASINA 303

and Zerconopsis. In group (c) the genera Arctoseius and Iphidozercon are at present
placed in the Aceosejinae and Zerconopsis in the Platyseiinae. This revised grouping
of the genera appears to be more satisfactory since members of group (c) which may
be referred to as the Arctoseiinae, differ from the other two groups in the nature of
the chaetotaxy of the opisthonotal region of the dorsal shield in the female. The
Arctoseiinae also show affinities with the Digamasellus-gTOup mentioned above.

The chaetotaxy of legs I and II also provide good " key characters " for the
separation of certain f amilies which have been difficult to distinguish on such characters
as idiosomal sclerotization and chaetotaxy. Two such cases immediately come to
mind, namely, the separation of the Aceosejidae and Phytoseiidae, and of the Evi-
phididae and Laelaptidae. Reference to the chaetotactic patterns of the genua and
the tibiae of legs I will now remove much of the uncertainty associated with the recog-
nition of members of these families in existing key works.

The results of isolated observations on the leg chaetotaxy of the Uropodina and
Antennophorina (or Fedrizzina) suggest that segmental chaetotaxy is also of impor-
tance at all levels of their classification. In the Uropodina, for example, there is a
marked and constant difference between the segmental chaetotaxy of the Lower
Uropodina (Trachytid-type) and the more specialized Higher Uropodina (Uropodid-
type) ; the latter exhibiting considerably more localized neoteny in leg chaetotaxy
to the extent of certain segments retaining the larval chaetotaxy in the adult stages.
A comprehensive study of the leg chaetotaxy in the Uropodina is now being under-
taken by Mr. Brian Ainscough.

The Mesostigmata, in common with the Anactinochaeta as a whole, display little
variety in setal types. Unlike the Actinochaeta which have a variety of setal types
that can be distinguished on the basis of their optical and chemical properties (Grand-
jean, 1935), the setae of the leg segments of the Mesostigmata referred to in
the present work appear to be of one type. They are solid and lack " actinochitin ".
The sensory field distally on tarsus I, however, has a variable number of hollow setae
which bear a superficial resemblance to the solenidia of certain Actinochaeta.

REFERENCES

EVANS, G. O. 1957. An introduction to the British Mesostigmata with keys to families and
genera. J. Linn. Soc. Lond. Zool. 43 (No. 291) : 203-259.

EVANS, G. O. & HYATT, K. H. 1963. Mites of the genus Macrocheles Latr. (Mesostigmata)
associated with Coprid beetles in the collections of the British Museum (Nat. Hist.). Bull.
Brit. Mus. (not. Hist.) Zool. 9, 9 :

GRAND j EAN, F. 1935. Observations sur les Acariens (i ere serie) . Bull. Mus. Hist. nat. Paris
(2), 7 : 119-126.

TILL, W. M. 1963. Ethiopian mites of the genus Androlaelaps Berlese s. lot. (Acari : Meso-
stigmata). Bull. Brit. Mus. (nat. Hist.) Zool. 10, i : 1-104.

ZOOL. 10, 5

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED
BARTHOLOMEW PRESS, DORKING

A REVISION OF THE RECENT
ROUND HERRINGS

(PISCES: DUSSUMIERIIDAE)

P. J. P. WHITEHEAD

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 6

LONDON : 1963

A REVISION OF THE RECENT

ROUND HERRINGS
(PISCES: DUSSUMIERIIDAE)

P. J. P. WHITEHEAD

Department of Zoology, British Museum (Natural History)

- 305-380 ; 33 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 6

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. 6 of the Zoological
series.

Trustees of the British Museum 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued May 1963 Price Twenty Seven Shillings

A REVISION OF THE RECENT

ROUND HERRINGS
(PISCES: DUSSUMIERIIDAE)

By P. J. P. WHITEHEAD

CONTENTS

Page
SYNOPSIS ............ 308

INTRODUCTION ........... 308

THE FAMILY DUSSUMIERIIDAE ........ 309

THE SUBFAMILY DUSSUMIERIINAE ....... 310

THE GENUS Dussumieria . . . , . . . , .311
Dussumieria acuta . . . . . . . .312

THE GENUS Etrumeus .......... 320

Etrumeus teres . . . . . . . . .321

THE SUBFAMILY SPRATELLOIDINAE . . . . . . 328

THE TRIBE EHIRAVINI (new tribe) . . . . . . 330

THE GENUS Ehirava . . . . . . . . . .331

Ehirava malabaricus . . . . . . . .331

E. madagascariensis ........ 333

THE GENUS Gilchristella ......... 333

Gilchristella aestuarius ....... 334

THE GENUS Sauvagella ......... 335

Sauvagella bianalis . . . . . . . .336

THE TRIBE SPRATELLOIDINI . . . . . . . -337

THE GENUS Spratelloides ......... 338

Spratelloides gracilis ........ 338

S. gracilis gracilis, S. g. atrofasciatus ..... 345

Spratelloides delicatulus ....... 345

S. delicatulus delicatulus, S. d. robustus ..... 348

THE GENUS Jenkinsia .......... 348

Jenkinsia lamprotaenia . . . . . . 351

/. lamprotaenia lamprotaenia, J. I. viridis .... 354

Jenkinsia majua (sp. nov.) ....... 355

/. majua majua, J. m. woodsi (new subsp.) .... 356

SYSTEMATIC CHARACTERS ......... 356

1. Scutes 356

2. Jaw elements .......... 359

3. Branchiostegal rays . . . . . . . . .360

4. Shape of gill opening and bones of opercular series . . .361

5. Other skull features ......... 363

SPECIATION IN THE DUSSUMIERIIDAE ....... 366

ORIGINS AND RELATIONSHIPS OF THE DUSSUMIERIIDAE .... 367

ACKNOWLEDGEMENTS .......... 37 1

REFERENCES . . . . . . . . . . . 37 1

ZOOL."io, 6

3 o8 P. J. P. WHITEHEAD

SYNOPSIS

The recent round herrings, the Dussumieriidae, are here maintained as a family distinct from
the Clupeidae or true herrings. Two subfamilies are recognized, the Dussumieriinae containing
the genera Dussumieria and Etrumeus ; and the Spratelloidinae, which is further split into the
tribes Spratelloidini and Ehiravini. The former contains the genera Spratelloides and Jemkinsia,
and the latter Ehirava, Gilchristella and Sauvagella.

In those genera with a wide distribution (e.g. Dussumieria, Etrumeus, Spratelloides), morpho-
metric and meristic data, from samples covering the entire geographical range, indicate that
many former species are no more than geographical variants, or in some cases subspecies. Ten
species are recognized, and these are described and figured. Keys are given for all taxa.

The systematics and phylogeny of the Dussumieriidae are examined in the light of certain
hitherto neglected characters, principally the development of abdominal and pelvic scutes, the
formation of the posterior cranial fontanelles, the shape of the cleithral outline of the gill opening,
the shape of the bones in the opercular series, and the number and shape of the supra-maxillary
bones. It is concluded that the Dussumieriidae are modern representatives of a primitive non-
scuted clupeid.

INTRODUCTION

THE Dussumieriidae, or round herrings, are small clupeid fishes fairly widely distri-
buted in tropical and temperate seas, mainly in the Indo-Pacific region. They are
usually separated from the Clupeidae by their absence of abdominal scutes, thus
having rounded rather than keeled bellies. Jordan & Gilbert (1883) , Giinther (1868) ,
and Weber & de Beaufort (1913), among earlier writers, placed the round herrings as
a subfamily of the Clupeidae, but nowadays they are usually given family status (e.g.
since Jordan, 1925). Svetovidov (1952) however retained them in a subfamily of the
Clupeidae, but the absence of scutes in all but one species of round herring is here con-
sidered evidence that the evolution of this group predated the evolution of the scuted
clupeid groups.

The most recent revision of the family was that of Bertin (1943), who recognized
eight genera containing recent species within two subfamilies, the Dussumieriinae
and the Spratelloidinae. I have here rejected three of these genera (Montalbania,
Perkinsia and Halecula] but have accepted two others (Ehirava and Jenkinsia}.
Comparatively little systematic work has been published recently on the group,
being for the most part descriptions of species and some notes on bionomics. Chap-
man (1948) published a useful description of the osteology of the round herrings,
comparing them with the clupeids, but considered only a single genus, Etrumeus.
Ridewood (1904) had earlier compared the skull of Dussumieria with certain clupe-
oids, but scarcely any work has been done on the osteology of the spratelloidine
round herrings (except caudal anatomy by Hollister, 1936).

The recent discovery of abdominal scutes in a round herring (Whitehead, I962a)
raises the problem of the relationship of the Dussumieriidae to the Clupeidae ; it
has also thrown more light on the division between the Spratelloidinae and the
Dussumieriinae. I have found here that, although the Spratelloidinae approach
the Clupeidae in many respects, and especially in the form of the pelvic scute, those
species which most closely approach the Clupeidae are at the same time those which
also most closely resemble the Dussumieriinae in other characters. On the other
hand, in certain characters the division between the Spratelloidinae and the Dus-
sumieriinae is as marked as is the division between either and the Clupeidae. But

A REVISION OF THE RECENT ROUND HERRINGS 309

the presence or absence of scutes seems to me to be of such fundamental importance
that the round herrings should be separated from the clupeids at family level.

Although Gosline (1951) drew attention to the Ruling of the International Com-
mission for Zoological Nomenclature concerning the use of the generic name Stole-
phorus, it is worth repeating that, under Opinion 93, the genotype of Stolephorus
Lacepede is S. commersonianus 1 Lacepede, by description and figure an undeniable
anchovy, so that Spratelloides Bleeker must replace Stolephorus as a round herring
genus, and the round herrings are the Dussumieriidae, not the Stolephoridae, as
Fowler (1941, 1958), Smith (1955), and others have termed them. The error arose
through an inadequate description by Houttuyn of a fish later included by Lacepede
in his engraulid genus Stolephorus ; this is discussed further on p. 340.

This revision is based on collections in the British Museum (Natural History) and
on material lent or donated by other institutions, for which I would like to thank
particularly, Dr. R. Rosenblatt of the Scripps Institute of Oceanography; Professor
C. R. Robbins of the Institute of Marine Studies, University of Miami; Professor
J. L. B. Smith of Rhodes University; Mr. A. Ben-Tuvia, of the Sea Fisheries
Research Station, Haifa ; Dr. L. Woods of the Chicago Natural History Museum ;
Dr. M. Blanc of the Museum Nationale d'Histoire Naturelle, Paris ; Dr. J. Nielsen,
Universitetets Museum, Copenhagen, and Dr. H. Steinitz of the Hebrew University,
Jerusalem.

Family DUSSUMIERIIDAE

Diagnosis

Clupeoid fishes usually with elongate, fusiform bodies and rounded bellies (body
more compressed in the tribe Ehiravini). One or two abdominal scutes associated
with the pelvic fin ; pre- or post-pelvic scutes entirely absent except for the former
in one instance (Gilchristella aestuarius] ; neither the pelvic scutes nor, where present,
the pre-pelvic scutes are keeled.

Anal fin normally equal to or shorter than dorsal, exceptionally longer. Pelvics
slightly in front, below or a little behind dorsal. Pectorals set low on body. Anal
always behind dorsal.

Mouth terminal, lower jaw more or less projecting. Premaxilla small, edentulous
or with a single series of small conical teeth which are often deciduous. Small,
conical and sometimes deciduous teeth on dentary, along lower edge of maxilla, on
glossohyal, suprabasal (where present), mesopterygoids and palatine. A well-
developed posterior supra-maxilla overlapping distal tip of maxilla and produced
anteriorly into a pointed shaft ; a second, plate-like supra-maxilla sometimes
present, lying between the shaft and the upper edge of the maxilla.

Hyomandibular with two separate cranial heads articulating with both sphenotic
and pterotic ; ceratohyal with or without indented ventral edge. Branchiostegal
rays from six to twenty.

Pseudobranchiae well-developed ; gill membranes separate, free from isthmus ;
gillrakers fine and slender but rarely more than about forty. Pyloric caecae
numerous. Adipose tissue often entirely covering eye.

1 A cheironym which should be rejected in favour of the name actually used by Lacepede, commersonii.

310 P. J. P. WHITEHEAD

Sensory canals of head well-developed, with superficial ramifications extending
on to pre-operculum, sub-orbitals, operculum, and sometimes on to maxilla, part of
articular, and sub-operculum.

Scales cycloid, covering entire body except head, often highly deciduous ; elongate
axillary scales in angle of pectoral and pelvic fins and elongate scales on upper and
lower lobes of caudal. Vertebrae 30-56.

The Dussumieriidae are small, often brilliantly silvery fishes chiefly found in the
Indo-Pacific region between latitudes 40 N. and 40 S., but with a few species along
the Pacific and Atlantic coasts of North America. They are found in coastal regions
and in estuarine and tidal lagoons and, although rarely exploited by any large
fishery, they are acceptable in some places as food fishes when caught in sufficient
number. One of the three fishes constituting the " Iwashi" fishery of Japan, is
the round herring, Etrumeus.

The Dussumieriidae fall into two very distinct groups, the Dussumieriinae, larger
fishes with more branchiostegal rays (14-19) ; and the Spratelloidinae, comprising
species which rarely exceed no mm. and have 6-7 branchiostegal rays. The former
subfamily appears to be the more primitive, while the latter approaches the Clupeidae
so nearly in certain cases (e.g. Gilchristella aestuarim], that it would be tempting to
place the Spratelloidinae with the clupeids were it not for the closely related but
more typical spratelloidines, such as Spratelloides gracilis.

I. Branchiostegal rays 14-19 ; adult size 150-350 mm. ; pelvic scute w-shaped ; no

posterior cranial fontanelles in adults; ceratohyal not excavated ventrally; dorsal

rays 16-21 ........... Dussumieriinae

II. Branchiostegal rays 6-7 ; adult size 50-110 mm. ; pelvic scute often with ascending

process ; a pair of posterior cranial fontanelles in adults of most species ; cerato-
hyal excavated ventrally ; dorsal rays i i-i 6 ..... Spratelloidinae

Subfamily DUSSUMIERIINAE

Diagnosis

Dussumieriid fishes, with 14-19 branchiostegal rays, the first seven to nine
attached to the ceratohyal whose ventral edge is not excavated. Premaxilla
toothed, teeth not deciduous ; maxilla with a narrow posterior supra-maxilla whose
depth is about half that of the maxilla at its widest point ; a second, anterior supra-
maxilla present in some cases. Ventral scutes absent except for a modified w-shaped
scute surrounding the base of the pelvic fins and sometimes a second triangular,
plate-like scute immediately behind the pelvics.

No posterior cranial fontanelles in adults, this portion of the head forming a
shallow, triangular depression. Posterior margin of pre-operculum not vertical
but inclined forwards ; ventral margin of operculum not horizontal but rising
posteriorly.

A small, usually little developed, fleshy eminence on the postero-ventral angle of
the cleithrum. Inter-operculum barely exposed in lateral view.

Dorsal rays 16-21 ; anal 9-19. Transverse scales on body 11-15. Vertebrae
52-56.

A REVISION OF THE RECENT ROUND HERRINGS 311

Adult size 150-350 mm.

Two genera, Dussumieria and Etrumeus.

A. Pelvic fins under dorsal base ; two supra-maxillae present ; anal rays 14-19 ;

exposed portion of sub-operculum sub-rectangular .... Dussumieria

B. Pelvic fins behind dorsal base ; a single supra-maxilla ; anal 9-13 ; exposed portion

of sub-operculum triangular ......... Etrumeus

Etrumeus and Dussumieria are obviously closely related, but Fowler (1958) used
the differences in pelvic position, number of anal rays, and presence or absence of
a second supra-maxilla to erect a tribal division in the Dussumieriinae. This does
not seem justified and the differences between these two genera are small compared
with those used here to split the Spratelloidinae into two tribes (p. 329).

Genus DUSSUMIERIA Valenciennes

Dussumieria Valenciennes, 1847, Hist. Nat. Poiss., 20 : 467 (Type : Dussumieria acuta Valenc.).
Montalbiana Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 7. (Type : Etrumeus (Mont-
albania) albulina Fowler, 1934.)

Body elongate, more or less compressed. Snout pointed, jaws equal. Two supra-
maxillae, each about half the length of the maxilla (Text-fig. 28). Proximal half of
maxilla thickened along its dorsal edge, distal half flat, tip of maxilla rounded,
almost entire ventral edge bearing small conical teeth. Premaxillae toothed. A
single w-shaped scute surrounding base of pelvic fins. Dorsal rays 17-21, anal rays
14-19, 20-34 gillrakers on lower part of first gill arch. Branchiostegals 14-20.
Vertebrae 54-56. Anal well behind the dorsal, whose origin is a little nearer the
caudal base than the tip of the snout. Pelvic origin below middle of dorsal fin. A
slight fleshy eminence on the angle of the posterior margin of the cleithrum
(cleithral flap), not so developed as in Spratelloides , but more developed than that
of Etrumeus (see Text-fig. 3ob). Gill filaments of first arch shortened to accom-
modate this eminence. Ventral margin of operculum nearer to horizontal than in
Etrumeus (Text-fig. 3ob).

A single species recognized here, D. acuta.

This genus is entirely Indo-Pacific in its natural distribution, but since the opening
of the Suez Canal there have been several records of Dussumieria, and especially the
Erythrean form, being caught in the eastern Mediterranean (Lissner, 1949, and Ben-
Tuvia, 1953). I have examined twenty such specimens and they have the high
gillraker count typical of the Red Sea population (Text-fig. 3). The Red Sea form
evidently is not so closely adapted to hot, hypersaline conditions that it cannot
survive elsewhere. On the other hand it has been able to survive passage through
the even more saline Bitter Lakes. It will be interesting to see whether the Mediter-
ranean population will now lose the rather distinctive characters of the Red Sea
form.

312 P. J. P. WHITEHEAD

Dussumieria acuta Valenciennes
(Text-fig. 5)

Dussumieria acuta Valenciennes, 1847, Hist. Poiss. Nat., 20 : 467, pi. 606 (Type locality : Bombay,

Coromandel) ; Cantor, 1849, /. Asiat. Soc. Beng., 18: 1268; Day, 1865, Fishes of Malabar:

226; Kner, 1865, Reise Novarra, Fische: 330; Giinther, 1868, Cat. Fish. Brit. Mus., 7: 466;

Bleeker, 1872, Atlas Ichth. Ind. Neerland., 6: 94, pi. 271, fig. i; Day, 1878, Fishes of India,

pt. 4 : 647, pi. 166, fig. 4 ; Weber and Beaufort, 1913, Fishes Indo-Aust. Arch., 2 : 21, fig. 13 ;

Fowler, 1928, Mem. Bernice P. Bishop Mus., 10: 30; Roxas, 1934, Philipp. J.Sci.,55: 251,

pi. i, fig. 5 ; Fowler, 1941, Bull. U.S. not. Mus., No. 100 : 570 ; Berlin, 1943, Bull. Inst.

oceanogr. Monaco, No. 853 : 3, figs, i (scale) and 2 ; Liang, 1948, Quart. J. Taiwan Mus., 1 :

2; Monroe, 1955, Marine and freshwater fish. Ceylon : 28; Fowler and Steinitz, 1956, Bull. Res.

Counc. Isreal, 5 B (3-4) : 261.
Dussumieria elopsoides Bleeker, 1849, Verh. Bat. Gen., 22 : 12 ; Giinther, 1868, Cat. Fish. Brit.

Mus., 7 : 466.
Dussumieria hasselti Bleeker, 1850, Natuurk. Tijdschr. Ned. Ind., 1 : 422 ; Idem, 1872, Atlas

Ichth. Ind. Neerland., 6 : 95, pi. 271, fig. 2 ; Day, 1878, Fishes of India, pt. 4 : 647, pi. 166,

fig- 5 ; Weber and Beaufort, 1913, Fish. Indo-Aust. Arch., 2 : 23 ; Roxas, 1934, Philipp. J.

Sci., 55: 250; Fowler, 1941, Bull. U.S. nat. Mus., No. 100 : 572; Bertin, 1943, Butt. Inst.

oceanogr. Monaco, No. 853 : 6, fig. 2 ; Schultz and Wellander, 1953, Bull. U.S. nat. Mus., No.

202 ; 25 ; Monroe, 1955, Marine and freshwater fish, Ceylon : 28 ; Fourmanoir, 1961, Mem. Inst.

sci. Madagascar, (F) 4 : 84, fig. i.
Dussumieria productissima Chabanaud, 1933, Bull. Inst. oceanogr. Monaco, No. 627 : 4, figs. 3-6 ;

Idem, 1933, Bull. Soc. zool. France, 58 : 289 ; Gruvel and Chabanaud, 1937, Mem. Inst. egypt.

(Egypte), 35 : 3, fig. 3 ; Fowler, 1941, Bull. U.S. nat. Mus., No. 100 : 570 ; Bertin, 1943, Bull.

Inst. oceanogr. Monaco, No. 853 : 6 ; Ben-Tuvia, 1953, Bull. Sea Fish. Res. Stat., (Israel), No. 8 :

6, fig. i.
Etrumeus (Montalbania) albulina Fowler, 1934, Proc. A cad. nat. Sci. Philad., 85: 244, fig. 7;

Idem, 1941, Bull. U.S. nat. Mus., No. 100 : 574, fig. 14.
Montalbiana albulina Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 8.

Notes on Synonymy

The most important references are cited here ; full synonymies are given by
Fowler (1941) under D. acuta, D. productissima, and D. hasselti.

Fowler (1941), Bertin (1943) and Herre (1953) amongst others, have all included
Dussumieria elopsoides Bleeker in the synonymy of D. acuta Valenciennes, recognizing
two further species, D. hasselti Bleeker, which extends to the eastern limits of the
range of this genus (China), and D. productissima Chabanaud from the extreme
western fringe (Gulf and Isthmus of Suez). Giinther (1868) had, however, placed
D. hasselti in synonymy with D. elopsoides. Examination of the holotypes of the
last two species has revealed no possible basis for specific distinction on preserved
characters (see Table I). At the same time, the specimens in this museum labelled
D. acuta and D. elopsoides (sensu Giinther) both show a parallel variation in several
characters which can clearly be correlated with geographical distribution. Specimens
from the intermediate part of the range of these two species could be assigned to
either species, and in fact Bertin (1943) stated that all but two of Valenciennes types
of D. acuta should be referred to D. hasselti.*

1 I have examined these specimens through the courtesy of Dr. M. Blanc of the Museum Nationale
d'Histoire Naturelle, Paris, and wish to thank him for allowing me to borrow them.

A REVISION OF THE RECENT ROUND HERRINGS 313

Unfortunately Valenciennes did not state the number of scales in lateral series in
his D. acuta, and neither did Chabanaud for D. productissima. The scales in
Dussumieria are highly deciduous ; amongst 29 specimens of D. acuta in this museum,
there is only one in which a scale count can even be estimated. In this case 36
scales are actually present, but at least 12 scales, more likely 15, are missing but
fairly well represented by scale pockets. This would place the fish in the range of
D. elopsoides (i.e. 52-56 scales ; cf. 42-44 reported for D. acuta). In the majority
of descriptions it is rarely stated whether the scale count is an actual one based on
the specimens examined, or whether it merely follows previous descriptions. An
exception is Blegvad (1944), who counted 42-44 scales in specimens from the Gulf
of Iran and assigned them to D. acuta ; I have examined these fishes, but all traces
of even the scale pockets are now obliterated.

It seems therefore that scale counts are an unreliable character for separating
D. acuta from D. elopsoides. Delsman (1925) came to the same conclusion and
found little difference in vertebral counts in a specimen each of D. acuta and D.

f *

- ^

/" A*

* AA j*

x **'

<a*

ft /*\ Sr ^^ l*\

tSk /? ^* ** ^*" ** ***

T A

o ^b

i i i i i i i

i i

70 80 90 100 110 120 130

140 150

FIG. i.

Dussumieria acuta. Body depth expressed as a percentage of standard length (ordinate)
plotted against standard length in mm. (abscissa) for various populations. Based on specimens
(see list of Study Material) labelled or recorded under the following names :

D. acuta X Iranian Gulf specimens collected by Blegvad (1944).

O Gulf of Aden specimens

(in rectangle, five Taiwan specimens, based on Liang, 1948).

All others.
D. elopsoides Hong Kong and China specimens.

^ All others.
D. productissima O Specimens from Haifa (eastern Mediterranean).

314 P. J- P. WHITEHEAD

hasselti (55 and 54 respectively). He also mentions a slight difference in dorsal
position, but I have not found this. The weight of other evidence (snout length,
body depth, gillraker and dorsal fin ray counts) strongly suggests that the number
of scales conforms to the general geographical trends shared by both D. acuta and
D. elopsoides. On the basis of the four characters mentioned, it is impossible to
recognize D. elopsoides as a separate species ; the evidence for this is presented below
under each heading.

(a) Body Depth

The measurements plotted in Text-fig. I (as percentages of standard length) refer
to specimens of Dussumieria covering almost the entire geographical range of the
genus. The Aden specimens (open circles) suggest positive allometry with standard
length, and this may explain some of the lower figures for the fishes labelled D. acuta
(black dots). The scatter-diagram shows clearly that size for size, the specimens of
D. elopsoides (triangles j cannot be distinguished from D. acuta in this character,
except in a few cases ; three of these are from China (black squares) , one is from
Amboina, and one bears no locality label. In addition, measurements for five
fishes from the Pescadores Islands, Taiwan (D. acuta of Liang, 1948) have been
placed within a rectangle. The result is a striking similarity between the far-eastern
specimens (D. elopsoides) and the Gulf of Aden specimens, with the Mediterranean
specimens (encircled dots) also giving low values.

I conclude that body depth cannot be used to separate D. elopsoides from D. acuta
and that in general the lowest values are found in both the eastern and the western
populations, with higher values in intermediate areas. The two specimens recorded
(as D. hasselti} by Schultz & Wellander (195.3) from Batavia appear to be much more
slender than any I have measured, having a body depth of only 15% of standard
length. In addition, these authors count 61 scale pockets. Unfortunately the
fishes were rather damaged, so it would be unwise to include these figures until
more specimens can be examined.

(b) Snout Length

Although the snout comprises only 8-10% of standard length and variations in
its length are barely perceptible, nonetheless, when plotted in a series of histograms
for various populations (Text-fig. 2) the results are very suggestive of a cline which
may reverse direction after reaching a minimum value in the Indian Ocean. Thus
the highest figures are those for specimens from the Gulf of Aden, from the eastern
Mediterranean (derivatives of a Red Sea population) and from China ; the lowest
are those from the coasts of India. The samples are small, but the overall picture
deserves attention because of its almost perfect correspondence with the situation
found in gillraker counts (Text-fig. 3).

In Text-fig. 2 specimens from each locality have been separated into the species
under which they were recorded or labelled. The result shows clearly that in areas
where both D. acuta and D. elopsoides are represented, there is nothing to distinguish
the two, and that D. elopsoides from China, and D. productissima from the Mediter-
ranean merely continue the trends already shown by the other populations. Snout

A REVISION OF THE RECENT ROUND HERRINGS

315

n n n n

. d

i i i i i i i I I i i i i i i i i I I i i i i i i i i I I i i i i i i i i

8 9 10

Snout length (% of S.L.)

FIG. 2.

Dussumieria acuta. Snout length (expressed as a percentage of standard length) in various
populations, showing relative frequency in each group. Based on specimens (see list of Study
Material) which have been labelled or recorded under the following names :

White D. acuta, Black D. elopsoides, Hatched D, productissima.

a. Hong Kong, Foochow, Amoy.

b. Siam, Amboina, Java, Madura, Sumatra, Borneo, Andamans, Singapore.

c. Malabar, Calicut, Bombay, Madras, Coromandel.

d. Gulf of Iran (upper), Gulf of Aden (lower).

e. Haifa (eastern Mediterranean).

length cannot serve as a basis for distinguishing the species when specimens from
the entire range are considered.

Gillraker frequency counts are presented similarly in Text-fig. 3. These counts
were made on the lower part of the first arch only, and included in the count was
the occasional raker lying exactly in the angle of the arch. Counts for the Mediter-
ranean specimens lie in the lower part of the range cited by Chabanaud (1933) for
D. productissima. It is possible that, like other meristic characters, gillrakers are
susceptible to exogenous factors (especially temperature), and that a higher count may

P. J. P. WHITEHEAD

. . }e

i i i i i i i i i

1 1 1

u - 1 J
1 1 1

19 20 21 22 23 24 25 26 27

28 29 30 31

32 33 34

Gillrakers
FIG. 3.

Dussumieria acuta. Gillraker count frequencies (lower half of first arch only) in various
populations. Based on specimens (see list of Study Material) which have been labelled or
recorded under the following names :

White D. acuta, Black D. elopsoides, Hatched D. productissima.

a. Hong Kong, Foochow, Amoy.

b. Siam, Amboina, Java, Madura, Sumatra, Borneo, Andamans, Singapore.

c. Malabar, Calicut, Bombay, Madras, Coromandel.

d. Gulf of Iran (upper), Gulf of Aden (lower).

e. Haifa eastern Mediterranean (broken line includes range given by Chabanaud (1933)

for 1 6 specimens from the Gulf and Isthmus of Suez).

occur in Red Sea populations ; this is, however, the reverse of the usual direction
in temperature-influenced meristic variations.

As in snout length and body depth, there is a hint that the general west to east
trend of decreasing gillrakers is reversed, so that the China specimens again approach
the western population in numbers.

The Gulf of Aden specimens (Text-fig. 3d lower histogram) effectively bridge
the gap between " D. productissima " (sensu Chabanaud) and the remaining popula-
tions of D. acuta. The sample from the Gulf of Iran (d-upper histogram) only
contained three fishes, but a larger sample might well provide a more effective bridge
linking the Gulf of Aden population with that of the Indian Ocean. But for the
Gulf of Aden specimens, there would be good reason to suppose the Red Sea

A REVISION OF THE RECENT ROUND HERRINGS 317

D. productissima a distinct species. If a subspecies were considered, then the Gulf of
Aden fishes should be included also, but I do not think such a subspecies could be
maintained once adequate collections have been made in the western Indian Ocean.

(d) Dorsal Rays

The first simple ray of both dorsal and anal fins is easily missed, being small, often
minute. For this reason simple and branched rays in both fins have been counted
separately, and only the latter plotted in Text-fig. 4. Here the trend, if such there
is, appears to be reversed, highest values occurring in Indian Ocean populations,
and lowest values at the extremities of the geographical range. But variations in
both dorsal and anal rays are very small, and the graph probably shows no more
than that once again no distinction can be made between D. acuta and D. elopsoides.

I have been unable to find other characters on which populations of Dussumieria
can be distinguished. The shape of the exposed portion of the suboperculum varies
somewhat, from a rectangle with an obliquely truncated posterior margin, to a more
triangular area with a rather rounded margin (as in Etrumeus}, but such differences
appear to be individual variants. Similarly, the area and shape of the palato-
pterygoid toothpad also shows some variation, and the size and number of jaw
teeth varies. Nor can any distinction be made on the sculpture patterns on the
wedge-shaped fronto-parietal surface, the shape of the operculum, or in any body
proportions.

Thus the only difference remaining between the three recognized species is in
numbers of scales in lateral series, a badly recorded and uncertain character. From
the evidence presented on other characters it seems unlikely that scale numbers
would in fact show the clear-cut differences suggested by previous descriptions.
Therefore, I do not think that three separate species of Dussumieria can be main-
tained, the populations from one area merging imperceptibly with those of the next.
There are more grounds for considering the Red Sea population a distinct subspecies,
but here again no definite limits can be drawn between the Red Sea specimens and
those for example from the Gulf of Aden. If the Red Sea form is to be separated,
so also should the far eastern populations, but the latter could only be defined in
terms which could include the Red Sea fishes, which would surely be an unrealistic
use of the concept of subspecies. It seems preferable therefore, to leave all in one
rather variable species, D. acuta, until much more work has been done.

Several authors have noted the similarity between specimens of Dussumieria from
the extreme eastern and western limits of its range. Bertin (1943) suggested
" segregation centrifuge ". Certainly there is no hydrological similarity between the
two areas, so that exogenous factors cannot be held entirely responsible for the
reversal in the east-west trend in certain characters.

Finally, mention must be made of Etrumeus (Montalbania] albulina Fowler, which
is here placed in the synonymy of D. acuta. Bertin (1943) raised this form to generic
status, apparently believing it to be intermediate between Dussumieria and Etrumeus,
but nearer to the former than the latter. From Fowler's description and figure
however, it is clear that this form cannot be referred to Etrumeus. Thus the almost

P. J. P. WHITEHEAD

> d

14 15 16 17

Branched dorsal rays

FIG. 4.

Dussumieria acuta. Dorsal finray count frequencies (branched rays only) in various popula-
tions, showing relative frequency in each group. Based on specimens (see list of Study
Material) which have been labelled or recorded under the following names :

White D. acuta, Black D. elopsoides, Hatched D. productissima.

a. Hong Kong, Foochow, Amoy.

b. Siam, Amboina, Java, Madura, Sumatra, Borneo, Andamans, Singapore.

c. Malabar, Calicut, Bombay, Madras, Coromandel.

d. Gulf of Iran (upper), Gulf of Aden (lower).

e. Haifa (eastern Mediterranean).

A REVISION OF THE RECENT ROUND HERRINGS 319

horizontal border of the operculum, and the shape of the sub-operculum are charac-
teristic of Dussumieria (see Text-fig. 3oa and b). In no case has a specimen of
Etrumeus been found with n branched anal rays (7-8, Text-fig. 6) or 25 gillrakers
(27-36, see Text-fig. 8) or a body depth exceeding 21% of standard length (Text-fig. 9),
whereas these are all within the normal range of Dussumieria. Bertin (loc. cit.)
stresses the rather advanced dorsal origin in his Montalbiana albulina, but this may
be an error ; it is not apparent in Fowler's figure. M. albulina should therefore be
included in D. acuta.

Jl^

FIG. 5.
Dussumieria acuta (from a specimen 148 mm. standard length, Gulf of Aden. Scales omitted).

DESCRIPTION : Based on the holotypes of D. elopsoides (124 mm.) and D. hasselti
(134 mm.) ; eight fishes, 112-117 mm. standard length from Calicut, S. Malabar ;
and six fishes, 121-148 mm. standard length from Shihr and Burum, Gulf of Aden.
In addition, meristic counts and proportions of snout and body depth for all speci-
mens listed under Study Material have been used (including sixteen syntypes of
D. acuta).

In percentages of standard length : body depth 18-2-29-6 (Text-fig, i),
head length 25-7-29-4, snout length 7-4-10-6 (Text-fig. 2), eye diameter 5-3-6-3, post-
orbital distance 9-2-10-1, inter-orbital width 5-6-5-8, maxilla length 8-4-9-3, lower
jaw length 12-3-12-4, pectoral length 15-8-16-2, pelvic length 8-6-8-9, pre-dorsal
distance 53-0-59-0, pre-pelvic distance 6o-o-655, pre-anal distance 79-0-81-5 (84-5
in one instance).

Body moderately compressed, more or less rounded ventrally, its depth less or
equal to head length. Snout pointed, larger than eye diameter. Lower jaw projects
beyond upper. Maxilla shorter than snout, not quite reaching anterior eye border ;
two supra-maxillae, the posterior about half width of maxilla, not expanded pos-
teriorly. Curved, conical teeth on premaxilla, maxilla (anterior two thirds) and
dentary. Post-orbital about equal to snout length. Dorsal origin nearer to caudal
base than to snout tip. Pelvic origin below middle or anterior half of dorsal fin,
nearer to caudal base than to pectorals.

Dorsal iv 14-17 (Text-fig. 4), pectoral i 11-14, pelvic i 7, anal iii 11-13.

Gillrakers on lower part of first arch 18-34 (Text-fig. 3), longest raker 3-4-3-5%
in standard length.

Scales in lateral series 42-56, transverse 11-12.

320 P. J. P. WHITEHEAD

Vertebrae 53 (i fish), 54 (4), 55 (i), 56 (4) (ten fishes from Singapore;.

Branchiostegal rays 14-16.

SYNTYPES. Coast of Coromandel (India). Paris Museum, No. 3697, 3694 and
3217.

COLOUR IN ALCOHOL. Dorsal surfaces brown, sides yellow-brown or silver. Tip
of snout strongly pigmented. Sometimes a dusky line from operculum to caudal
base. Fins pale, but first pectoral ray and tips of caudal dusky.

MAXIMUM SIZE. 216 mm. (Day).

DISTRIBUTION. Red Sea, Madagascar to northern part of Indian Ocean, Indo-
Malayan Archipelago and northwards to Hong Kong.

The presence of D. acuta in the Mediterranean was first noted by Lissner (1949),
and later Ben-Tuvia (1953) stated that these round herrings were common along the
shores of Israel and were caught by trawl or purse seine. It will be interesting to
see whether the change in environment will produce any corresponding departure
from the Red Sea form. Fowler & Steinitz (1955) placed Lissner's five fishes in
D. acuta (rather than D. productissima, as Ben-Tuvia had done for his own specimens),
but it is not clear from the text whether this determination was based solely on an
approximate scale count, or whether gillrakers, etc. were also considered.

ETRUMEUS Bleeker

Etrumeus Bleeker, 1853, Verh. Bat. Gen., 25 : 48 (Type : Clupea micropus Schlegel).

Perkins ia Eigenmann, 1891, Amer. Nat. Philad., 25 : 153 (Type : Perkinsia othonops Eigenmann) .

Halecula Jordan, 1925, Stanford Univ. Publ. Biol. Sci., 4 : 41 (Type : Haleculaacuminata Jordan).

Parahalecula Fowler, 1958, Notul. Naturae, Philad., No. 310 : 5. (Halecula Jordan, 1925, pre-
occupied) .
For notes on this synonymy, see under species.

DESCRIPTION. Body elongate, almost round, scarcely compressed. Snout pointed,
jaws equal or lower projecting slightly. A single supra-maxilla, about half length
of maxilla, tapering uniformly to point anteriorly and about a quarter as deep as
maxilla (Text-fig. 28). Maxilla thickened along whole dorsal edge and bearing a
branched sensory canal. Small conical teeth along almost entire ventral edge of
maxilla. Premaxilla toothed. A w-shaped scute surrounding base of pelvic fins
and a smaller, triangular scute immediately behind this. Dorsal 17-22, anal lo-n,
26-39 gillrakers on lower part of first arch. Branchiostegal rays 14-15. Vertebrae
48-56. Anal well behind dorsal, whose origin is a little nearer to snout than to
caudal base. Pelvic origin behind dorsal fin. Very small or no fleshy eminence at
postero- ventral angle of gill opening (Text-fig. 3oa). Ventral margin of operculum
rises at steep angle ; exposed portion of suboperculum triangular.

There are five principal populations of Etrumeus in temperate seas : North
American Atlantic and North American Pacific coasts, the coasts of Japan, of
South Africa and of southern Australia. In addition there appears to be a popula-
tion in the Red Sea, members of which have now colonized parts of the eastern
Mediterranean ; another population in the region of the Galapagos Islands ; and a
population near Hawaii. This distribution will be discussed later.

A REVISION OF THE RECENT ROUND HERRINGS 321

Etrumeus teres (DeKay)
(Text-fig, n)

Alosa teres De Kay, 1842, Nat. Hist. New York, pt. 4 Fishes: 262, pi. 40, fig. 128 (type locality

New York region).
Clupea micropus Schlegel, 1846, Faun. Japan. Poiss., pts. 10-14 : 2 3^. pi- IO 7 fig- 2 (type

locality, southeast coast of Japan).
Etrumeus micropus Bleeker, 1853, Verh. Bat. Gen., 25 : 48 ; Idem, op. cit., 26 : 5 ; Gunther, 1868,

Cat. Fish. Brit. Mus., 7 : 467 ; Jordan and Evermann, 1905, Bull. U.S. Fish Comm., 23 (i) :

58 ; Jordan and Herre, 1906, Proc. U.S. nat. Mus., 31 : 628 ; Gilchrist and Thompson, 1917,

Ann. Durban Mus., 1 (4) : 295 ; Barnard, 1925, Ann. S. Afr. Mus., 21 (i) : 108 ; Fowler, 1928,

Mem. Bernice P. Bishop Mus., 10 : 29 ; Idem, 1934, P*oc. Acad. nat. Sci. Philad. ,86 : 410;

Idem, 1941, Bull. U.S. nat. Mus., No. 100: 576 ; Chapman, 1948, Proc. Calif. Acad. Sci., 26

(2) : 25, figs. 1-3, 7-10, 12-13, I 5> 17-18 ; Svetovidov, 1952, Tabl. Anal. Faune U.R.S.S.

N.S. No. 48, 2 (i) : 102 ; Smith, 1955, Ann. Mag. nat. Hist., (12) 8 : 307 ; Fowler and Steinitz,

1956, Bull. Res. Counc. Israel, 5 B (3-4) : 261 ; Matsubara and Iwai, 1959, Fishes, biol. Res.

Jap. Antarct. res. Exped., No. 9.

Harengula teres Girard, 1859, Proc. Acad. nat. Sci. Philad., 2 : 158.
Dussumieria teres Brevoort, 1856, in Perry, Narrative of the U.S. Exped. to Japan : Washington,

2 : 279 ; Gill, 1861, Proc. Acad. nat. Sci. Philad., 12 : 21.
Etrumeus teres Gunther, 1868, Cat. Fish. Brit. Mus., 7 : 467 ; Jordan and Gilbert, 1882, Bull.

U.S. nat. Mus. Wash., 16 : 263.
Etrumeus jacksoniensis Macleay, 1879, Proc. Linn. Soc. N.S.W., 3 : 36, pi. 4, fig. I ; Ogilby, 1886,

Cat. Fishes New South Wales : 56; McCulloch, 1914, Rec. W. Aust. Mus., 1 : 211, pi. 29 ;

Waite, 1921, Rec. S. Aust. Mus., 2 (i) : 36, fig. 51 ; Blackburn, 1941, Bull. Coun.sci.industr.

Res. Aust., No. 138 : 64.
Etrumeus sadina Jordan and Evermann, 1896, Bull. U.S. nat. Mus. Wash., No. 47 : 420 ; Bertin,

1943, Bull. Inst. oceanogr. Monaco, No. 853 : 9, fig. 4.
Etrumeus acuminatus Gilbert, 1891, Proc. U.S. nat. Mus. Wash., 13 : 56.

Jenkinsia acuminata Jordan and Evermann, 1896, Bull. U.S. nat. Mus. Wash., No. 47 : 419.
Perkinsia othonops Eigenmann, 1891, Amer. Nat. Philad., 25 : 153 ; Jordan and Evermann,

1896, Bull. U.S. nat. Mus. Wash., No. 47 : 420 ; Breder, 1928, Bull. Bingham. oceanogr. Coll.

N.Y., 2 (2) : 5, figs. 2-4 ; Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 13.
Halecula acuminata Jordan, 1925, Stanford Univ. Publ. Biol. Sci., 4 : 41 ; Bertin, 1943, Bull. Inst.

oceanogr. Monaco, No. 853 : 24.

Parahalecula acuminata Fowler, 1958, Notul. Naturae, Philad., No. 310 : 5.
Stolephorus delicatulus Scale, 1940, Reps. Allan Hancock Pacific Exped. 1932-38, 9 : 3.
Etrumeus othonops Phillips, 1961, Calif. Fish Game, 37 : 512.

Notes on Synonymy

Bertin (1943) gave Clupea sadina Mitchill 1814 as the earliest name for this species,
but in fact he cited a paper published in 1815 ; it was first used by Mitchill in 1814
in a short paper entitled " Report in part of Samuel L. Mitchill, M.D., Professor of
Natural History, &c., on the fishes of New York " (pp. i-x, 1-30). Described by
Bashford Dean as " one of the rarest of American contributions to ichthyology ",
it was reprinted in 1898 by Theodore Gill. In his first description of Clupea sadina,
Mitchill follows a colour description with "scales fall off very readily ; body has a
taper, slender, and very delicate appearance. Abdomen not at all serrated, but
quite smooth ... ". Although this description fits a species of Etrumeus,
Mitchill enlarged on it in a paper read within the space of a year (8th December,

ZOOL. 10, 6 18

322

P. J. P. WHITEHEAD

1814) and published in 1815, and while this second description of Clupea sadina does
not contradict the first in any way, it adds details which cannot be reconciled with
a species of Etrumeus. Thus he places Clupea sadina, the " New York Shadine ",
under a subheading " Bellies carinated without serrae " and says that " On account
of the even connection of the false ribs, the belly is not at all serrated, but quite
smooth." More important are the discrepancies in meristic counts, and especially
that for branchiostegal rays ; he counts 7 rays, as against 14 or 15 in Etrumeus. He
records 9 pelvic rays (normally 8 in Etrumeus), and 15 anal rays (never more than
II counting the minute first, unbranched ray). He also states that the mouth is
wide and toothless and mentions " a small smutty spot behind the gill-cover ".
Finally, he states that there is " A semitransparent space in front of the eyes from
side to side."

(2)
(I)

(9)1

(9)

Red Sea
Mediterranean

Australia

U.S.A.

(Pacific)

Japan

U.S.A.

(Atlantic)

South Africa

iii7 iii8 i 14 j 15 i 16

Anal rays Pectoral rays

FIG. 6.

Anal (white) and pectoral (black) finray count frequencies in seven different populations of
Etrumeus. Branched rays only. Number in sample placed in parentheses.

A REVISION OF THE RECENT ROUND HERRINGS

323

Elsewhere (Whitehead, in press) I have dealt more fully with this problem and
have shown that Clupea sadina Mitchill need not become a nomen dubium, because
the second description strongly suggests that Mitchill was describing a species of
Sardinella and nothing in the first description contradicts this. There is also good
reason to believe that both descriptions were based on the same specimen or speci-
mens. DeKay's description of Etrumeus teres on the other hand leaves no doubt that
it is a species of Etrumeus.

Perkinsia othonops (American Pacific) is placed in the synonymy because the
corselet of scales surrounding the pectoral fin base, a supposed generic character, is
in fact found in large specimens of Etrumeus.

Etrumeus (Montalbanid) albulina Fowler has here been synonymized with Dus-
sumieria acuta (see p. 312). The advanced pelvic base and the shape of the sub-
operculum exclude it from Etrumeus, and numbers of gillrakers and anal rays are
those of Dussumieria.

Red Sea
Mediterranean

Australia

U.S.A. (Pacific)

(2)
(8)

Japan

U.S.A. (Atlantic)

South Africa

(22)

(9)

(8)

17 18 19 20 21
Total dorsal rays

FIG. 7.

Dorsal finray frequencies (branched and unbranched combined) in specimens of Etrumeus
from seven different populations. Number in sample placed in parentheses.

3 2 4 P. J. P. WHITEHEAD

Halecula acuminata is placed in the synonymy since it was based on juvenile
specimens of E. acuminatus (Hubbs, in. litt.}.

(a) Meristic Differences

Variations in anal finray counts (Text-fig. 6) are small (7 or 8 branched rays and
always 3 simple rays, the first very small and easily missed). Pectoral finray
numbers also vary little, and in my material only the Australian, Mediterranean, and
Japanese specimens occasionally have 16 branched rays, the remainder 14 or 15
(Text-fig. 6). Differences in dorsal rays are more marked. Normally there are 4
simple rays (occasionally 5), the first very small, but sometimes the last of the simple
rays is branched, although recognizable by its length. In dorsal ray numbers one
population, that of South Africa, can be separated immediately because of its low
count, but the remainder overlap and cannot be separated from each other (Text-
fig. 7). Pelvic counts are always i 7.

In numbers of gillrakers there is also some variation. The Australian and
American Pacific specimens have low counts, the South African are intermediate,
and the Japanese, Mediterranean and American Atlantic have high counts (Text-
fig. 8.) Again there is considerable overlap between the regions.

(b) Morphometric Differences

The most obvious proportional difference found between the samples is that of
body depth (Text-fig. 9) . Again the Japanese and the American Atlantic specimens
resemble each other, and are more slender than the rest. But although insufficient
numbers have been measured, Text-fig. 10 strongly suggests that body depth shows
positive allometry with standard length. Thus the Japanese form may well be
deeper-bodied in larger fishes, as is suggested by the two large Japanese specimens
examined (126 and 136 mm.). Certainly the American Atlantic and the American
Pacific specimens can be distinguished on this character, but the remainder, and
probably also the Japanese specimens, are very similar.

A second morphometric difference is found in the positions of the dorsal, pelvic
and anal fins. In the American Pacific specimens these fins are all set slightly further
from the snout than in the fishes from the American Atlantic (see Table II). The
remaining differences are small and would probably disappear in larger samples.

If the American Atlantic population is taken as the starting point, then the
American Pacific fishes can be distinguished by their deeper bodies and fewer gill-
rakers (27-33 ; cf. 34-36). The South African population can be separated from all
others by its lower dorsal count (17-18 total rays; cf. 19-22). The Japanese
population, however, cannot, on the basis of my material, be adequately separated
from the American Atlantic population, and the two Australian specimens are in all
characters within the range of the American Pacific population.

The distribution of Etrumeus is difficult to explain in zoogeographical terms. In
some ways it resembles that of Sardinops in the Southern Hemisphere, but so far no
specimens have been recorded from South American localities ; the Galapagos
population may represent a southern American form, pushed northwards by the

A REVISION OF THE RECENT ROUND HERRINGS 325

Red Sea _ (2)

Mediterranean (I)

Australia (2)

U.S.A. (Pacific) (8)

Japan ..;':-^ (22)

U.S.A. (Atlantic) ^^.J (9)

South

Africa

(9)

27 28 29 30 31 32 33 34 35 36 37 38

Gillrakers

FIG. 8.

Range and frequency in numbers of gillrakers in seven different populations of Etrumeus.
Number in sample placed in parentheses.

cold Peruvian current. I have not been able to examine specimens from the
Galapagos Islands, and cannot relate them to any of the other forms examined here.
Possibly the Hawaiian population represents a link between the former and the
Japanese population ; Jordan & Evermann (1905) stated that their Hawaiian
specimens were indistinguishable from the Japanese form.

The isolated Red Sea population is even more difficult to explain. The two
Eilat specimens, and the single fish from the eastern Mediterranean, certainly do
not belong to the South African population, their nearest neighbours ; they appear
to be most closely related to the Japanese fishes. This seems to provide further
evidence that meristic characters may coincide in populations which are not closely
related geographically. It is certainly strange that a species which elsewhere
appears to be limited to between the (approximately) 12 and 20 C. isotherms should
appear in the Red Sea, and equally remarkable that it should, under these conditions,
show so little divergence from other populations.

The isolation of each of these populations is probably complete and it would be
expected that each would have diverged at least slightly. It is possible therefore

326 P. J. P. WHITEHEAD

MM Red Sea (2)

Mediterranean (!)

Australia (2)

Japan (22)

U.S.A. (Atlantic) (9)

_J | | | | | i i

14 15 16 17 18 19 20

Body depth as a % of standard length

FIG. 9.

Range of variation in body depth (expressed as a percentage of standard length) in seven
populations of Etrumeus. Number in sample placed in parentheses.

that, since variation is principally restricted to a few meristic characters, similarities
have occurred between otherwise well isolated populations purely as a result of
parallel evolution. Thus the similarities between the two population pairs cited
above may be to a large extent coincidental. This in some ways resembles the
case of Dussumieria discussed earlier, where morphological similarities contradict
probable geographical relationships. But whereas in Dussumieria there is a series
of both geographical and morphological intermediates, in Etrumeus the populations
are well isolated.

The genus Etrumeus is at present under revision by Prof. Carl L. Hubbs and
Mr. Robert Wisner, and I have therefore made no attempt to interpret the present
data. I have here followed the example of the last reviewer (Bertin) and have
placed all the forms in a single species, rather than create new taxa on the basis of
my material alone.

A REVISION OF THE RECENT ROUND HERRINGS

327

20
19
18
17
16
15
14

k X

o o

O xx x

60 70 80 90 100 110 120 130 140 ISO 160 170

Standard length (mm.)

FIG. 10.

Body depth expressed as a percentage of standard length (ordinate) plotted against standard
length (abscissa) for specimens of Etrumeus from seven different populations.

Japan and Hong Kong.

O American Atlantic.

O American Pacific.

X South Africa.

A Australia.

+ Mediterranean.

V Eilat (Red Sea).

FIG. ii.
Etrumeus teres. From a specimen 150 mm. S.L., ex Woods Hole, Mass. Scales omitted.

328 P. J. P. WHITEHEAD

Etrumeus teres (DeKay)
(Text-fig, n)

DESCRIPTION. Based on twelve Japanese specimens (67-5-85-5 mm. standard
length) ; nine fishes from South Africa (117-5-166-0 mm.) ; two fishes from Australia
(105-5-122-0 mm.) ; nine American Atlantic fishes (115-0-130-2 mm.) ; seven
American Pacific fishes (110-0-125-0 mm.) ; one Mediterranean fish (165-0 mm.) ;
the type of Perkinsia othonops (265-0 mm.) ; and two Red Sea fishes (133-2 and
134-0 mm.). Specimens listed in Study Material (p. 374).

In percentages of standard length : body depth 14-6-21-2 (Text-figs. 9 and 10),
head length 23-2-29-2, snout length 6-9-8-8, eye diameter 7-0-9-8, post-orbital
distance 7-9-10-6, maxilla length 8-7-9-7, pectoral length 14-7-17-4, pelvic length
7-6-10-2, pre-dorsal distance 44-0-48-4 (also 49-4 and 50-0), pre-pelvic distance
62-0-7O-0, pre-anal distance 82-0-88-0.

Body rounded, little compressed, its depth less than head length. Snout pointed,
more or less equal to eye diameter. Lower jaw usually projects beyond upper.
Maxilla longer than snout, reaching vertical with anterior border of pupil. A single
supra-maxilla, not expanded posteriorly. Conical pointed teeth on pre-maxilla,
maxilla and dentary. Post-orbital exceeds snout length. Dorsal origin nearer to
snout than to caudal base. Pelvic origin behind dorsal base, nearer to caudal base
than to pectorals. Anal origin further from that of pelvics than from anal base.

Dorsal iii-v, usually iv, 14-18, pectoral i 14-16, pelvic i 7, anal iii 7-^8 (Text-figs.
6 and 7).

Gillrakers on lower part of first arch 27-36 (Text-fig. 8).

Scales in lateral series 50-56, transverse 13-14 (after Bertin).

Vertebrae 48-56 (Hubbs, in. litt.).

Branchiostegal rays 14-15.

COLOUR IN ALCOHOL. Dorsal surfaces light or dark brown, sides silver, the two
meeting at a fairly well-defined midlateral line. Tip of snout strongly pigmented.
Fins hyaline.

MAXIMUM SIZE. At least 260 mm.

DISTRIBUTION. Seven probably discrete populations : American Atlantic coast
(Cape Cod to Gulf of Mexico) ; American Pacific coast (Gulf of California and north
of Los Angeles area see Phillips, 1951) ; eastern coast of South Africa (Natal,
Zululand) ; southern coasts of Australia (New South Wales, Victoria, St. Vincent's
Gulf, and Albany in Western Australia see Blackburn, 1941) ; coasts of Japan
(Nagasaki, Wakanoura, Misaki, Aomora, Tokyo) ; Galapagos Islands (see Scale,
1940) ; and the eastern Mediterranean (? migrants from the Red Sea), and Red
Sea (Eilat) (Fowler & Steinitz, 1956).

Subfamily SPRATELLOIDINAE

Diagnosis

Dussumieriid fishes with 6-7 branchiostegal rays, the first three to five attached
to the cerato-hyal, which is excavated ventrally (Text-fig. 29). Premaxilla some-
times toothed ; maxilla with a broadly expanded posterior supra-maxilla, often as

A REVISION OF THE RECENT ROUND HERRINGS 329

deep as the maxilla at its widest point ; an anterior supra-maxilla sometimes present.
Abdominal scutes represented by either a w-shaped plate surrounding pelvic base
(Spratelloides , Jenkinsia} or a horse-shoe-shaped plate with ascending spines (tribe
Ehiravini) (Text-fig. 26) ; exceptionally a series of 6-9 such plates between pectoral
and pelvic fins (Gilchristella aestuarius}, all with ascending spines but never keeled
ventrally (Text-fig. 27).

Posterior fontanelles present, but decreasing in extent with size of fish, and in
some species (Ehirava malabaricus] absent entirely in large fishes. Posterior margin
of pre-operculum more or less vertical. Posterior margin of operculum excavated,
ventral margin horizontal or slightly inclined. A small, fleshy eminence at postero-
ventral angle of gill opening, prominent in some genera. Lower edge of inter-
operculum exposed in lateral view. Sub-operculum rectangular.

Dorsal rays 11-16 ; anal 9-20. Transverse scales on body 7-10. Vertebrae
30-46.

Adult size 50-110 mm.

Five genera are recognized here, Ehirava, Gilchristella, Sauvagella, Spratelloides
and Jenkinsia. The first three differ so much from the other two that I have thought
it advisable to split the Spratelloidinae into two tribes.

A. Pelvic scute with an ascending, pointed spine ; a single supra-maxilla ; premaxilla

toothed ; pelvic fins under anterior half of dorsal or in advance of first dorsal ray ;
fleshy eminence at postero-ventral angle of gill opening little developed ; posterior
fontanelles broadly divided anteriorly by wedge of bone .... Ehiravini

B. Pelvic scute w-shaped as in Dussumieriinae, without ascending spine ; one or two

supra-maxillae ; pre-maxilla normally edentulous ; pelvic fins under middle, or
second half of dorsal ; fleshy eminence at postero-ventral angle of gill opening
usually well developed ; posterior fontanelles narrowly divided anteriorly in most
species ........... Spratelloidini

Notes on Tribe Ehiravini

In his review of the round herrings, Bertin (1943) recognized two genera from
southern Africa and Madagascar which differed from Spratelloides in having the
pelvic base in advance of the dorsal origin, not under the dorsal. The first,
Gilchristella Fowler, contained G. aestuarius Fowler, and also Sauvage's Spratelloides
madagascariensis , into which Bertin had earlier placed his two subspecies of Sauva-
gella madagascariensis (longianalis and brevidorsalis , Bertin, 1940). The second
genus, the monotypic Sauvagella Bertin, was further distinguished by possession of
a split anal, the last two anal rays being distinctly separated from the rest of the
fin (confirmed in alizarin preparations, Bertin, 1943). Bertin (1943) felt it possible
that Sauvagella bianalis might be merely a mutant form of Gilchristella madagas-
cariensis, the two differing little except in the form of the anal fin.

Later, Angel, Bertin & Guibe (1946) proposed the nomen novum Spratellomorpha
to replace Sauvagella of Bertin, 1943 (not of Bertin, 1940, which was now included
in Gilchristella). This is discussed under the synonymy of Sauvagella.

These three South African and Malagasi species could be placed in a single genus
but for the discovery that G. aestuarius, alone of the whole Dussumieriidae, possesses
ventral scutes (Whitehead, 1^623.). While the split anal fin of 5. bianalis could

330 P. J. P. WHITEHEAD

perhaps be considered a chance mutation, the possession of abdominal scutes cannot
be lightly dismissed in view of the importance of scutes in the phylogeny of the group
as a whole. I have found these scutes in two specimens of G. aestuarius from
Durban, and also in seven further specimens from various South African localities
(see under species description).

Although G. aestuarius is thus unique, and shows supra-limital variation (in the
sense of Myers, 1960) in this one character, otherwise it closely resembles its non-
scuted geographical relatives. Therefore, I do not think G. aestuarius should be
separated from the South African species at higher than generic level. Nor does it
seem that G. aestuarius is the sole representative of an ancient line deriving from
earlier clupeids. It belongs to a group in which at least one other clupeid character
is also found, the divided anal fin of 5. bianalis (i.e. in the clupeid genus Corica).
It would appear therefore that the Ehiravini share certain potential genetic patterns
characteristic of the clupeids, but that these have shown only a partial development
in some species but not at all in others. The Ehiravini may thus be derived from
forms which lay close to the split between the round herrings and the true herrings.

The South African group, although all more closely allied to Spratelloides than to
any other dussumierid genus, differ from the latter in six important characters,
i. Pelvic scute with lateral spines (Text-fig. 26).
ii. A single supra-maxilla.

iii. Advanced pelvics.

iv. Fleshy eminence on postero-ventral angle of cleithrum little developed,
v. A toothed premaxilla.

vi. Posterior fontanelles broadly divided by wedge of bone anteriorly, in front of
which is a triangular depression (Text-fig. 32a).

These characters are also shared by Spratelloides malabaricus from the Malabar
coast of India, which should therefore be included with the South African species.
S. malabaricus is at the same time identical to Ehirava fluviatilis Deraniyagala.
Deraniyagala (1929) proposed a new family, the Ehiravidae, on the strength of this
one species, distinguishing it from the Dussumieriidae by the possession of only one
supra-maxilla ; he considered Ehirava intermediate between Spratelloides and
Dussumieria. Since its description, Ehirava has been mentioned only once (Monroe,
1955). Since 5. malabaricus cannot be retained in Spratelloides, the genus Ehirava
is available for it. At the same time I have been unable to find any but very small
differences between S. malabaricus and the Malagasi species Gilchristella madagasca-
riensis. The latter should therefore be placed in Ehirava also, but it can be separated
from the Indian form at species level, at least on the available material ; the greatest
difference is in scale numbers, but this may well prove dependent on locality when
a larger sample is examined.

The tribe Ehiravini thus contains three genera and five species.

Tribe EHIRAVINI
Diagnosis

Members of the subfamily Spratelloidinae which possess only a single supra-
maxilla and a toothed premaxilla. Eminence on postero-ventral angle of cleithrum

A REVISION OF THE RECENT ROUND HERRINGS 331

poorly developed and similar to that found in Etrumeus or Dussumieria. Pelvic
scute horse-shoe-shaped, with pointed, ascending spines. Pelvic origin just behind,
below, or just in front of dorsal origin. Posterior fontanelles broadly separated at
anterior end by wedge of bone (frontals), in front of which is shallow triangular
depression.

DISTRIBUTION. Eastern coast of South Africa, coast of Madagascar, and the
western coast of India.

KEY TO GENERA

A. Additional scutes absent between pectoral and pelvic fins

i. Anal fin entire, last two rays not separate ...... Ehirava

ii. Last two anal rays separate from rest of fin . . . . . Sauvagclla

B. Six to nine abdominal scutes between pectoral and pelvic fins ; anal fin entire

Gilchristella

Ehirava Deraniyagala

Ehirava Deraniyagala, 1929, Spolia Zeylan, 15 : 34, pi. 14 (type E. fluviatilis Deraniyagala

= Spratelloides malabaricus Day).
Sauvagella Bertin (part.), 1940 (Sauvagella madagascariensis, i.e. S. m. longianalis and S. m. brevi-

dorsalis, but non S. m. bianalis), Bull. Mus. Hist. nat. Paris, (2) 12 : 300 (type Spratelloides

madagascariensis Sauvage ex Madagascar) ;

DESCRIPTION. Body elongate, more compressed than in Spratelloides, snout
pointed, lower jaw projecting. Posterior supra-maxilla present, as deep as maxilla,
anterior absent ; maxilla toothed, with anteriorly indented lower border (Text-fig.
281). Premaxilla toothed. Small fleshy eminence on postero-ventral angle of
cleithrum (cleithral flap), not more prominent than in Dussumieria (Text-fig. 3ob).
Posterior border of operculum slightly indented, but not to the extent found in
Spratelloides ; junction between operculum and sub-operculum not horizontal (as
in Spratelloides, Text-fig. 3oc), but inclined (as in Dussumieria, Text-fig. 3ob).

A single (pelvic) scute with ascending spines (Text-fig. 26). Pelvic origin just
behind or in front of dorsal origin. Branchiostegal rays 6. Scales not strongly
deciduous.

Two species recognized, but more material may merge the differences shown here.

a. Snout equal to or smaller than eye ; pelvic origin below first dorsal or slightly

behind; scales 35-38 ........ E. malabaricus

b. Snout a little greater than eye ; pelvic origin in front of first dorsal ray ; scales 43-48

E. madagascariensis

Ehirava malabaricus (Day)
(Text-fig. 12)

Spratelloides malabaricus Day, 1873, Proc. zool. Soc. Land., 240 ; Idem, 1878, The Fishes of India :
648, pi. 161, fig. 5 (Type locality : Malabar, India) ; Bertin, 1943, Bull. Inst. oceanogr. Monaco,
No. 853 : 17.

Ehirava fluviatilis Deraniyagala, 1929, Spolia Zeylan, 15 : 35, pi. 14 (Type locality : Ceylon) ;
Monroe, 1955, Marine and freshwater fish. Ceylon : 28.

332

P. J. P. WHITEHEAD

Note on Synonymy

The five specimens of S. malabaricus in the Museum are identical to the type and
paratypes of E.fluviatilis in both meristic characters and proportional measurements.

Pellonulops Smith, purported to be based on Spratettoides madagascariensis
Sauvage, is not placed in the synonymy because the genus is clearly based in fact
on a clupeid, not a dussumieriid. Thus Smith (1949) described both pre- and post-
pelvic scutes, and the presence of the latter eliminates the chance that his specimens
were Gilchristella aestuarius.

FIG. 12.
Ehirava malabaricus. From a specimen 58 mm. standard length, Canara. Scales omitted.

DESCRIPTION. Based on the type and seven of the larger paratypes of E. fluviatilis
from Kehelvatta, Ceylon (35-1-48-6 mm. standard length) and five specimens of
S. malabaricus from Malabar (40-5-56-2 mm.).

In percentages of standard length : body depth 15-4-22-4, head length 22-4-26-7,
snout length (5-7 one fish) 6-6-8-2, eye diameter 6-8-8-5, post-orbital distance 7-9-
9-8, maxilla length 8-7-9-9, pectoral length 13-9-16-3, pelvic length 11-3-12-5, pre-
dorsal distance 50-0-52-2, pre-pelvic distance 49-1-52-5 (53-4), pre-anal distance
72-0-76-5 (80-0).

Body fairly strongly compressed, especially in larger fishes, depth less than head
length. Snout pointed, equal or a little smaller than eye. Jaws unequal, lower
projecting. Maxilla longer than snout, almost reaching vertical with anterior rim
of pupil, with excavated lower edge anteriorly (Text-fig. 281). One supra-maxilla
only, expanded posteriorly as in Spratelloides. Maxilla with a single row of conical
teeth along lower edge. Pre-maxilla also with conical pointed teeth in a single series.

Dorsal origin mid-way between snout and caudal base. Pelvic origin below first
dorsal ray or just behind, a little nearer anal origin than to pectoral base. Anal
origin almost equidistant between that of pelvics and caudal base.

Dorsal iii 11-12, pectoral i 10-12, pelvic i 7, anal ii-iii 12-15 (total 14-18).

Scales in lateral series 35-38, 9 transverse.

Branchiostegal rays 6.

HOLOTYPE. Kehelvatta, Ceylon. B.M. (N.H.) 1929.7.1.1.

A REVISION OF THE RECENT ROUND HERRINGS 333

COLOUR IN ALCOHOL. A uniform light brown with in some specimens a faint silvery
midlateral stripe not quite as broad as eye. A short oblique line of dark pigment on
lower half of caudal base and another, almost horizontal, line along upper edge of
caudal base. In smaller specimens, bases of dorsal and anal fins pigmented.

SIZE. Largest specimen examined 56*2 mm. standard length.

ALLOMETRY. Apart from the eye (negative) there is no evidence from the
specimens measured that any other body part shows allometry with standard length.

DISTRIBUTION. Ceylon and Malabar coast of India.

Ehirava madagascariensis (Sauvage)

Spratelloides madagascariensis Sauvage, 1883, Bull. Soc. philom., Paris (7) 7 : 160 ; Idem, 1891,

Hist. Nat. Madagascar, Poiss. : 496, pi. 48, fig. 2.
Sauvagella madagascariensis longianalis and S. m. brevidorsalis Bertin, 1940, Bull. Mus. Hist.

nat., Paris (2) 12 : 300.
Gilchristella madagascariensis Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 21 ; Angel,

Bertin and Guib6, 1946, Bull. Mus. Hist, nat., Paris (2) 18 : 473-4.

DESCRIPTION. Based on a specimen, 40-0 mm. standard length ex Madagascar
(paratype of Spratelloides madagascariensis} ; and eleven other fishes, 41-0-52-5 mm.
from Buffalo river, Cape Province (S. Africa).

In percentages of standard length : body depth 16-0-22-8, head length 25-0-28-2,
snout length 7-3-7-8, eye diameter 7-3-8-5, post-orbital (8-2) 10-6-11-9, maxilla
length 8-7-9-0, pre-dorsal distance 50-3-56-0, pre-pelvic distance 50-0-54-9, pre-anal
distance 66-8-72-5.

Body compressed, depth less than head length. Snout pointed, usually a little
greater than eye diameter. Jaws subequal, lower projecting. Maxilla longer than
snout, reaching vertical through anterior border of pupil. Maxilla shape as in
Gilchristella. A single supra-maxilla.

Dorsal origin slightly nearer caudal base than snout. Pelvic origin in front of
dorsal, about equidistant between snout and caudal base, nearer anal base than
pectoral base. Anal nearer pelvic base than caudal base.

Dorsal iv 10-11, pectoral i 9, pelvic i 7, anal iii 14-17, gillrakers 40-56 (lower
numbers mainly in the smaller specimens and vice versa) .

Scales in lateral series 43-48, transverse 8.

TYPE. Madagascar. Paris Museum No. 3794.

COLOUR IN ALCOHOL. Uniform grey-brown. A faint silvery mid-lateral stripe
Two pigmented lines at base of caudal, as in Sauvagella bianalis.

SIZE. 60 mm. (Bertin, 1943).

DISTRIBUTION. Madagascar and Buffalo river, King Williamstown (Cape
Province) .

GILCHRISTELLA Fowler

Gilchristella Fowler, 1935, Proc. Acad. nat. Sci. Philad., 87 : 365, fig. 4 (Genotype : Spratelloides
aestuarius Gilchrist from Swartkops river).

DESCRIPTION. Body compressed, more so than in Ehirava, its depth almost equal
to head length. Snout pointed, lower jaw projecting. Anterior supra-maxilla

334

P. J. P. WHITEHEAD

absent ; posterior supra-maxilla expanded posteriorly to almost depth of maxilla,
anterior shaft-like. Maxilla toothed along lower edge ; pre-maxilla toothed. Small
fleshy eminence on postero-ventral angle of cleithrum scarcely as developed even
as in Ehirava (i.e. nearer to the Etrumeus than to the Dussumieria condition).
Posterior border of operculum slightly indented, as in Ehirava ; junction between
operculum and suboperculum not horizontal but oblique, more steeply inclined than
in Ehirava and thus resembling that of Etrumeus (Text-fig. 3oa).

Pelvic scute with thin, pointed ascending arms. Between pectoral and pelvic
bases, a series of six to nine similar scutes (Text-fig. 27).

Dorsal origin a little further from snout than caudal base ; pelvic origin below
first dorsal ray or in front, equidistant between snout and caudal base. Branchio-
stegal rays 6-7. Scales moderately deciduous.

A single species recognized here.

Gilchristella aestuarius (Gilchrist)
(Text-fig. 13)

Spratelloides aestuarius Gilchrist, 1914, Mar. Biol. Rep. S. Afr., No. i : 55 (Type material from :
Swartkops river, Port Elizabeth) ; Regan, 1916, Ann. Durban Mus., 1 : 167 ; Gilchrist and
Thompson, 1917, Ann. Durban Mus., I (pt. 4) : 296; Barnard, 1925, Ann. S. Afr. Mus. ,21
(pt. i) : 109.

Gilchristella aestuarius Fowler, 1935, Proc. Acad. nat. Sci. Philad., 87 : 365, fig. 4 ; Bertin, 1943
Bull. Inst. oceanogr. Monaco, No. 853 : 21.

FIG. 13.
Gilchristella aestuarius. From a specimen 60 mm. standard length, Durban. Scales omitted.

DESCRIPTION. Based on two fishes 51-0 and 53-0 mm. standard length from
Durban, and seven fishes 35-9-51-5 nun. from other South African localities (East
London, Knysa, St. Lucia, Keimouth, Milnerton and the type locality Swartkops
river-on loan from Professor J. L. B. Smith).

In percentages of standard length : body depth (17-5 one fish) 19-2-24-5, head
length 25-2-27-5, snout length 6-3-7-5 (8-1), eye diameter 6-8-8-1, post-orbital 8-1-
9-5, maxilla length 10-2, pectoral length 14-2, pelvic length 11-7, pre-dorsal distance
52-5-57-7, pre-pelvic distance 50-0-54-4, pre-anal distance 66-7-70-0.

A REVISION OF THE RECENT ROUND HERRINGS 335

Body strongly compressed, especially in larger fishes, its depth just less than head
length. Snout pointed, usually a little less than eye diameter. Jaws sub-equal,
lower projecting slightly. Maxilla longer than snout, reaching vertical through
anterior border of pupil, anterior excavation not as pronounced as in Ehirava (see
Text-fig. 281). One supra-maxilla only, shape and proportions as in Ehirava.
Maxilla and premaxilla with single row of conical teeth.

Dorsal origin a little further from snout than caudal base. Pelvic origin below
first dorsal ray or more usually in front, equidistant between snout and caudal base
or a little nearer the latter, and nearer to anal than to pectoral base. Anal origin
nearer that of pel vies than to caudal base. Six to nine pre-pelvic scutes.

Dorsal iii 11-12, pectoral i 10-11, pelvic i 7, anal iii 17 (total 20).

Scales in lateral series 40, transverse 9-10 (Bertin, 1943).

Branchiostegal rays 6-7, gillrakers on lower part of first arch 39-45 ; inner series
on first two arches absent, and lower part of inner series on third arch also absent.

TYPES. Swartkops river, near Port Elizabeth. South African Museum. (Nos.
10822 4).

COLOUR IN ALCOHOL. A uniform light brown with a faint midlateral silvery band.
Individual bases of dorsal and anal rays pigmented (black). Two short dark pig-
ment lines at base of anal, one almost horizontal along upper border, the other
oblique on lower border.

SIZE. Largest fish examined 53-5 mm. Barnard (1925) gives maximum size
70 mm.

ALLOMETRY. No indication except with eye measurement (negative allometry
with standard length).

DISTRIBUTION. Estuaries and lagoons of the eastern coast of South Africa.

SAUVAGELLA Bertin

Sauvagella (part.) Bertin, 1940 (Sauvagella madagascariensis bianalis only), Bull. Mus. Hist. nat.

Paris (2) 12 : 300 ; Berlin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 22.
Spratellomorpha Bertin, 1946, in Angel, Bertin and Guib6, 1946, Bull. Mus. Hist. nat. Paris (2)

18 : 473-4 (type Sauvagella madagascariensis bianalis Bertin ex Madagascar).

Note on Synonymy

Spratellomorpha was proposed (Angel, Bertin & Guibe, 1946) as a nomen novum in
order to overcome the confusion arising from the splitting of Spratelloides madagasca-
riensis into three subspecies (Bertin, 1940), first placed all together in the genus
Sauvagella and later separated, two being united and placed in Gilchristella. Because
of this the meaning of Sauvagella became obscured. However, as a result of placing
G. madagascariensis (in the sense of Bertin, 1943) in Ehirava, there is no reason why
Sauvagella cannot again be applied to bianalis ; thus Sauvagella Bertin, 1940 and
Sauvagella Bertin, 1943 both included bianalis, the latter description being the more
definitive since the two other subspecies had by that time been transferred to Gil-
christella.

DESCRIPTION. Body elongate, compressed, as in Gilchristella. Snout pointed,
lower jaw projecting. Anterior supra-maxilla absent ; posterior supra-maxilla

336

P. J. P. WHITEHEAD

expanded posteriorly to almost maxilla depth, anterior shaft-like. Pre-maxilla and
maxilla with a single row of conical teeth. Fleshy eminence on postero- ventral
angle of cleithrum very small, similar to that in Etrumeus (Text-fig. 3oa). Posterior
border of operculum slightly excavated ; junction between operculum and sub-
operculum not as steeply inclined as in Gilchristella but resembling that in Ehimva
and thus Dussumieria (Text-fig, sob).

Pelvic scute with thin, pointed ascending arms. No abdominal scutes.

Dorsal origin further from snout than caudal base. Pelvic origin in front of dorsal.

Branchiostegal rays 6.

A single species.

Sauvagella bianalis Bertin
(Text-fig. 14)

Sauvagella madagascariensis bianalis Bertin, 1940, Bull. Mus. Hist. nat. Paris, (2) 12 : 300.
Sauvagella bianalis Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 22, fig. 8.
Spratellomorpha bianalis Bertin, 1946, in Angel, Bertin and Guibe, 1946, Bull. Mus. Hist. nat.
Paris (2) 18 : 473-4.

Sauvagella bianalis.
Museum No. A 5174).

FIG. 14.

From a specimen 47 mm. standard length, Madagascar (syntype, Paris
Scales omitted.

DESCRIPTION. Based on four fishes, 44 '0-45 '5 mm., ex Madagascar (types of
Sauvagella bianalis).

In percentages of standard length : body depth 17-1-18-0, head length 25-2-26-5,
snout length 6-9-7-1, eye diameter 7-2-7-8, post-orbital 8-4-8-6, maxilla length 9-7-
10-0, pre-dorsal distance 53-4-56-5 (62-8 one fish), pre-pelvic distance 49-4-52-2, pre-
anal distance 69-0-72-5.

Body compressed, but body depth less than head length. Snout pointed, a little
less than eye diameter. Jaws subequal, lower projecting slightly. Maxilla longer
than snout, reaching vertical through anterior border of pupil. Maxilla as in
Gilchristella. A single supra-maxilla.

Dorsal origin a little nearer caudal base than snout. Pelvic origin in front of
dorsal, nearer to anal base than to pectoral base. Anal origin nearer that of pelvics
than to caudal.

A REVISION OF THE RECENT ROUND HERRINGS 337

Dorsal iii 12-13, pectoral i 12, pelvic i 7, anal iii 11-12 -}- 2.

Last two rays of anal fin separated from others by a gap equal to three rays.
After examining an alizarin stained specimen, Bertin (1943, p. 23) stated that the
gap between the two parts of the fin is due, not to a complete separation of the two,
but the fact that " /' actiniophore du premier rayon de I'anale posterieure est trois plus
allonge, dans sa partie horizontale, que les autres actiniophores."

Scales in lateral series 42-45, transverse 8-9. Vertebrae 45-46 (16-17 caudal)
(after Bertin).

SYNTYPES. Madagascar. Paris Museum No. A.5I74.

COLOUR IN ALCOHOL. A uniform grey-brown. A faint silvery midlateral stripe.
Caudal base with two dark pigmented lines, as in GilchristeUa aestuarius.

SIZE. 60 mm. (Bertin, 1943).

ALLOMETRY. Except in eye size (negative) no allometry found in other body parts
with standard length.

DISTRIBUTION. Madagascar.

Tribe SPRATELLOIDINI

Diagnosis

Members of the subfamily Spratelloidinae which possess a w-shaped pelvic scute
(Text-fig. 25), two supra-maxillae (except in Jenkinsia}, well-developed posterior
fontanelles usually narrowly divided anteriorly, and pelvic fins below the middle of
the dorsal.

DISTRIBUTION. Ranging from Japan and Australia to the Red Sea and South
Africa, with a genus in the Caribbean area (Venezuela to Bermuda).

Two genera :

A. Two supra-maxillae ; premaxilla edentulous ; cleithral flap well developed ;

posterior border of suboperculum evenly rounded ; posterior fontanelles always
narrowly divided anteriorly ; Indo-pacific region .... Spratelloid.es

B. A single supra-maxilla ; premaxillary teeth sometimes present ; cleithral flap little

developed ; posterior and ventral margins of suboperculum meeting at well-
defined angle ; posterior fontanelles sometimes broadly divided anteriorly ;
Caribbean region Jenkinsia

In many ways Jenkinsia stands between Spratettoides and the genera of the
Ehiravini, but the evolution of the spined pelvic scute seems to be such an important
step taken by the Ehiravini in the direction of the Clupeidae that Jenkinsia must be
placed closer to Spratettoides. In addition, within the genus Jenkinsia, one species
differs from Spratettoides principally in lacking the anterior supra-maxilla, whereas
the other species is much nearer the Ehiravini, possessing premaxillary teeth, and a
formation of the posterior fontanelles which differs from that in Spratettoides. This
is further discussed under the generic descriptions.

Bertin (1943) included the Caribbean species in Spratettoides, but the two are well
separated by the characters listed in the key above, geographical isolation reinforcing
this distinction.

ZOOL. 10, 6 19

338 P. J. P. WHITEHEAD

SPRATELLOIDES Bleeker

Spratelloides Bleeker, 1852, Verh. Bat. Gen., 24 : 29 (type Clupea argyrotaeniata'Bleeker = Clupea

gracilis Schlegel).
Stolephorus (non Lacepede) Fowler, 1941, Bull. U.S. nat. Mus., 13 (No. 100) : 561.

Note on Synonymy

Some recent authors have substituted Stolephorus for Bleeker 's Spratelloides, but
this is quite wrong and is discussed fully after the synonymy for Spratelloides gracilis.

DESCRIPTION. Body elongate, slightly compressed, rounded ventrally ; snout
pointed, jaws equal or lower very slightly projecting. Two supramaxillae, the
second (posterior) bone paddle-shaped with a slender anterior shaft ; maxilla
toothed, with evenly rounded lower border (Text-fig. 28). Pre-maxilla edentulous.
Fleshy eminence on postero-ventral angle of cleithrum (cleithral flap) well-developed,
indented anteriorly (Text-fig. 3oc), gill filaments of first arch also indented to accom-
modate cleithral flap. Posterior border of operculum strongly indented, more so
than in Ehirava. Junction between operculum and sub-operculum horizontal, the
latter bone subrectangular, its posterior margin rounded. Interoperculum exposed
and three or four branchiostegal rays visible externally. Posterior border of pre-
operculum vertical. Posterior border of gill opening s-shaped.

Two posterior fontanelles with a narrow median division (see Text-figs. 31 and 32).

A single, w-shaped pelvic scute. No abdominal scutes. Dorsal equidistant
between snout and caudal base or a little nearer snout. Pelvic origin below mid-
dorsal or below second half of dorsal.

Two species of Spratelloides recognized here, each with a subspecies.

a. A bright and prominent silver mid-lateral band ; total anal rays 11-14 scales in

lateral series 41-49 .......... S* gracilis

b. No silver band, but whole lower flank silver ; total anal rays 9-1 1 ; scales in lateral

series 32-46 .......... 5. delicatulus

These two species also differ in body depth, head length, post-orbital distance and
pectoral length.

Spratelloides gracilis (Schlegel)
(Text-fig. 18)

Clupea gracilis Schlegel, 1846, Faun. Japon. Poiss., pts. 10-14 : 2 3^, pi. 108, fig. 2 (type locality :
southeast coasts of Nagasaki) .

Clupea argyrotaeniata Bleeker, 1849, Journ. Ind. Arch., 3 : 72 (type locality : Macassar, south-
west Celebes.

Spratelloides argyrotaenia Bleeker, 1851, Natuurk. Tijdschr. Ned. Ind., 2 : 214 ; Idem, 1852, Verh.
Bat. Gen., Batavia, 24 : 29 ; Idem, 1852, Natuurk. Tijdschr. Ned. Ind., 3 : 775 ; Schultz and
Wellander, 1953, Bull. U.S. nat. Mus., 1 (202) : 23-24 (S. argyrotaeniata) .

Spratelloides gracilis Bleeker, 1853, Verh. Bat. Gen., Batavia, 25: 18 ; Idem, 1892, 6: 96, pi. (8)
266, fig. 2 ; Gunther, 1868, Cat. Fishes Brit. Mus., 6 : 465 (type of Clupea argyrotaenia} ;
Klunzinger, 1871, Verh. zool.-bot. Ges. Wien, 21 : 601 ; Weber and Beaufort, 1913, Fishes Indo-
Aust. Arch., 2 : 20, fig. 12 ; Hardenburg, 1933, Treubia, 14 (2) : 215 ; Bertin, 1943, Bull. Inst.
oceanogr. Monaco, No. 853 : 15-16.

A REVISION OF THE RECENT ROUND HERRINGS 339

Stolephorus japonicus (non Lac6pede) Jordan and Seale, 1905, Proc. U.S. nat. Mus.,28 : 770;
Jordan and Herre, 1906, Proc. U.S. nat. Mus., 31 : 629 ; Fowler, 1928, Mem. Bernice P. Bishop
Mus., 10 : 30 ; Herre, 1936, Field Mus. nat. Hist. Zool., 21 : 33 ; Fowler, 1941, Bull. U.S. nat
Mus., 13 (100) : 567 (full synonymy) ; Idem, 1956, Fishes of the Red Sea and Southern Arabia,
Jerusalem, p. 61, fig. 23.

Stolephorus gracilis Evermann and Seale, 1907, Bull. Bur. Fisher., 26 : 53 ; Whitley, 1953, Aust.
Zool., 11 : 332.

Spratelloides japonicus Mori, 1928, Journ. Pan Pacific Res. Inst., 3:3; Tanaka, 1933, Jap. Fish.
Life Colours, No. 47 ; Marshall, 1952, Bull. Brit. Mus. nat .Hist. (Zool.), 1 : 22 ; Schultz and
Wellander, 1953, Bull. U.S. nat. Mus., No. 202 : 24; Morrow, 1954, Ann. Mag. nat. Hist., (12)
7: 804; Okada, 1955, Fishes of Japan, Tokyo: 41; Monroe, 1955, Marine and freshwater fish.
Ceylon : 28 ; Jones, 1961, /. Mar. Biol. Assn. India, 2 (2), 267-8.

Spratelloides atrofasciatus Schultz, 1943, Bull. U.S. nat. Mus., No. 180 : 8, fig. i; Schultz and
Wellander, 1953, op. cit. No. 202 : 27, fig. 7.

Note on Synonymy

Only the most important references prior to 1940 are cited here ; full synonymies
are given by Fowler (1941) under Stolephorus japonicus.

The synonymy reveals two main issues. The first is whether the Japanese and
other far eastern populations should be separated from those of the Red Sea, Indian
Ocean and Indo-Malayan Archipelago. Marshall (1952) showed that his (admittedly
few) specimens of 5. gracilis from the Red Sea had lower pectoral and anal counts
than did specimens from Japan (the type locality of Schlegel's Clupea gracilis}.
Schultz & Wellander (1953) were however more emphatic, stating that gracilis, as
understood by Bertin (1943) and Weber & de Beaufort (1913), actually represented
" at least two species " the " japonicus " of Houttuyn, and Bleekers " argyrotaenia "
and their finray, scale and gillraker counts seemed to support their conclusion.
Counts made on specimens in the British Museum, and supplemented by those of
Schultz & Wellander (loc. cit.} are given in Text-figs. 15-17. Certainly the counts are
higher in the Japanese specimens, but not only do the counts from the extreme
boundaries of the geographical range overlap (i.e. Red Sea and Japan), but the
specimens of intermediate provenance (i.e. from the Indo-Malayan Archipelago)
show intermediate values.

The situation is similar to that found in the Museum specimens of Dussumieria (see
p. 312) and the same conclusion must be drawn. The populations of 5. gracilis from
any one region cannot be distinguished sufficiently clearly to merit specific distinc-
tion. The fact that in numbers of dorsal, pectoral and anal rays and in gillrakers
there appears to be the same gradual shift to higher numbers as one proceeds east-
wards suggests that this is merely a phenotypic response to some environmental
factor, possibly temperature (but cf. Dussumieria, p. 316).

A second eastern population which has been separated from 5. gracilis is S.
atrofasciatus Schultz, 1943, described from Samoa and distinguished from 5. gracilis
by its lower gillraker and scale counts (Schultz, & Wellander, 1953). In numbers
of dorsal, anal and pectoral rays it overlaps S. gracilis, but in gillrakers it is rather
lower (19-23 on the lower part of the first arch ; cf. 26 given as the lowest for
S. argyrotaeniata by Schultz & Wellander (loc. cit}}. This, and its geographical
isolation from other populations, suggests that separation from S. gracilis would be

340

P. J. P. WHITEHEAD

(34)
(10)

(18)

9 10 II 12

Branched dorsal rays

FIG. 15.

Spratelloides gracilis, branched dorsal rays. Dorsal finray frequencies in populations from:

A. Japan and Formosa.

B. Indo-Malayan archipelago and Philippines.

C. Red Sea.

Based on specimens in the British Museum and supplemented by figures from Schultz and
Wellander (1953). Numbers in sample placed in parentheses.

justified. But since only a single character is involved (i.e. gillrakers), and since in
this character as well as in other meristic counts 5. atrofasciatus consistently lies
at the lower end of the range for S. gracilis, there seems good reason to suppose
that it represents another ecophenotypic variation of S. gracilis. Therefore, I
do not believe the Samoan population differs specifically from S. gracilis, and I
have here given 5. atrofasciatus subspecific status only in order to emphasize its
place amongst the forms included in 5. gracilis.

The second issue raised by the synonymy is the question of the use of Stolephorus
Lacepede for a genus of round herring, and the citing of Atherina japonica Houttuyn
as the genotype of Stolephorus.

Some confusion has occurred over the application of the specific name " japonica "
Houttuyn to a species of Spratelloides. Houttuyn (1782) gave a poor description
of a Japanese fish, Atherina japonica, and Lacepede (1803) placed this fish, together

A REVISION OF THE RECENT ROUND HERRINGS

341

(24)

(7)
(3)

(17)
(46)

(23)

(20)

(49)

II
Total

12
anal rays

FIG. 1 6.

Anal finray frequencies in populations of Spratelloides from :

A. Japan and Formosa.

B. Indo-Malayan archipelago, Philippines and Australia.

C. Seychelles and Maldives.

D. Red Sea and Gulf of Aden.

Black S. gracilis, White S. delicatulus, Hatched S. argyrotaenia, Stippled S. g. robustus
(Australia) .

Based on specimens in the British Museum and supplemented by figures from Schultz &
Wellander (1953). Numbers in each sample placed in parentheses. N.B. Both branched and
simple rays included in counts.

with an anchovy described but not named by Commerson, in his genus Stolephorus.
The generic description and figure were evidently based on the latter species,
S. commersonii, and Opinion 93 given by the International Commission for Zoological
Nomenclature directed that 5. commersonianus (i.e. S. commersonii see footnote,

342

P. J. P. WHITEHEAD

p. 309) should be the designated genotype of Stolephorus and not Atherina japonica
(Jordan & Gilbert (1883, p. 272) had unfortunately designated the latter previously).
The confusion is aggravated because some authors have ignored Opinion 93 and
continue to call the round herring genus Stolephorus rather than Spratelloides (e.g.
Fowler, 1941 and 1958, Smith, 1955). At the same time the European anchovy

(17)

(1 5) (4)
(34) (35)

(22)

("3)
(33)

12 13

FIG. 17.

Pectoral finray frequencies in populations of Spratelloides from :

A. Japan and Formosa.

B. Indo-Malayan archipelago and Philippines.

C. The Seychelles and Maldives.

D. Red Sea.

Black S. gracilis, White S. delicatulus , Hatched S. argyrotaenia, Stippled S. robustus (Australia).

Based on specimens in the British Museum and supplemented by figures from Schultz &
Wellander (1953). Numbers in each sample placed in parentheses. The first, unbranched ray
included in the counts.

A REVISION OF THE RECENT ROUND HERRINGS 343

(Engmulis encrasicholus} has sometimes been placed in Stolephorus (Poll, 1947) ; its
very close Japanese relative is the Engraulis japonicus of Schlegel. Thus in Fowler
(loc. cit.) Atherina japonica Houttuyn is cited as the type both of a dussumieriid and
an engraulid species. Stolephorus should in fact refer only to those Indo-Pacific
engraulids which have at least some abdominal scutes (cf. the encrasicholus forms
which do not, and which should be placed in Engraulis Cuvier), normal pectoral rays
(cf. the species referred to Setipinna), and a maxilla not extending beyond the gill
opening (cf. Thrissocles] . Stolephorus is generally a synonym for those Indo-Pacific
anchovies otherwise placed in Anchoviella Fowler.

If Houttuyn's fish is assumed not to be an Engraulis, then E. japonicus Schlegel
is perfectly correct for the Japanese encrasicholus-type, anchovy. This assumption
is based on the supposition that Houttuyn was familiar with the European
E. encrasicholus and could have located the similar Japanese form in the Systema
Naturae (to which he refers in his description). Also, of the three possible Japanese
fishes with broad silvery lateral stripes (a character stressed by Houttuyn) only the
round herring has 8 pelvic rays, the scuted anchovy and non-scuted (encrasicholus-
type) anchovy each having only seven. Again, placing his fish in Linnaeus' Atherina,
Houttuyn may have counted (but not recorded) six branchiostegal rays, which also
accords with a sprat elloidine. In addition, Houttuyn described a fish with 14
pectoral rays. In 24 specimens of Spratelloides from Japan, I counted 14 rays in
5 fishes, 13 rays in 15 fishes, and 12 rays in 4 fishes (see Text-fig. 17) . Five specimens
of Engraulis from Japan had 16-18 rays, but in seven specimens of Stolephorus from
China there were 13 rays. The rays are however, extremely difficult to count in the
smaller dussumieriids and perhaps too much reliance should not be placed on
Houttuyn's count.

The remaining details of Houttuyn's description could fit all three Japanese fishes
(mouth toothless, head scaleless), but the dorsal count of 5 is surely a mistake. He
mentions size 4 inches but says " Ook heb ik 'er een van drie Duimen " (which
suggests that Jordan & Evermann (1917) were perhaps wrong in thinking that the
fish was described from rough notes or memory).

To resolve the matter, two courses are open. Either japonica Houttuyn 1782 is
considered a nomen dubium, there being no type specimen nor adequate description ;
or one of the three Japanese species with a broad silvery lateral stripe is accepted as
Houttuyn's fish. I favour the first course, for although Houttuyn's description
probably fits a dussumieriid fish of the genus Spratelloides better than it does an
anchovy, there is no further evidence that can be produced which will confirm the
identity of this fish. Spratelloides japonica Houttuyn should therefore be suppressed
in favour of 5. gracilis (Schlegel) as the first recognizable description of this species.
Application has been made to the International Commission for Zoological Nomen-
clature to this effect.

DESCRIPTION. Based on nineteen fishes, 59-0-93-0 mm. standard length, including
the lectotype (a specimen 66-7 mm. S.L. believed by Giinther (1868) to be Bleeker's
type from the East Indian Archipelago), and from Japan (15) and Formosa (3). In
addition, all other specimens listed under Study Material (p. 375) used for meristic
counts.

344

P. J. P. WHITEHEAD

Spratelloides gracilis.

FIG. 18.

From a specimen 90 mm. standard length, Wakanoura, Japan.
Scales omitted.

In percentages of standard length : body depth (under dorsal origin) 13-2-17-3,
head length 1 22-0-24-5, snout length 6-9-8-2 ; eye diameter 5*3-6-3, post-orbital
distance 2 7-4-8-5, maxilla length 8-1-9-0, pectoral length 11-6-12-2, pelvic length
9-4-9-8, pre-dorsal distance 47-0-50-0 (one fish 52-3), pre-pelvic distance 53-5-58-7,
pre-anal distance 79-0-84-7.

Body slightly compressed, its depth less than head length. Snout pointed, a
little larger than eye diameter. Maxilla longer than snout, passing front border of
eye but not reaching pupil. Post-orbital a little larger than snout length. Dorsal
origin equidistant, or usually just nearer snout than caudal base. Pelvic origin
under middle or second half of dorsal, a little nearer to caudal base than to snout.

Dorsal ii 9-12 (see Text-fig. 15), pectoral i 10-15 ( see Text-fig. 17), pelvic i 7,
anal 11-14, f which ii or iii are branched (see Text-fig. 16).

Gillrakers 20-37 on the lower part of the first arch (including one at angle) and
7-12 above angle.

Scales in lateral series 41-49 (based partly on Schultz & Wellander (loc. cit.} and
Bertin, 1943) ; 8-9 transverse rows.

Vertebrae 46 (3 specimens, Bertin (loc. cit.}}.

TYPE. East Indian Archipelago. Lectotype, B.M. (N.H.), 1867.11.28.17.

COLOUR IN ALCOHOL. Upper and lower surfaces brown, divided by a broad lateral
silver stripe, whose greatest width just exceeds eye diameter ; in some specimens the
lateral band is dark brown or black, much darker along its upper margin ; and in
all Japanese specimens examined the silvery stripe is outlined above by a thin,
dark brown line. In some specimens a dark brown or black line dorsally from nape
to caudal. Glandular scales on caudal not, or but faintly pigmented (cf. 5. deli-
catulus}.

DISTRIBUTION. Indo-Pacific region, from Red Sea to Japan ; southwards along
African coast to Pemba (Morrow, 1954) ; in Pacific, southwards to Samoa ; Indian
Ocean, Ceylon and Laccadive Sea (Jones, 1961).

1 The longest measurement, i.e. premaxillary symphysis to posterior border of operculum below the
indentation in the latter characteristic of Spratelloides. The measurement is thus not along a horizontal
line.

* The shortest distance, i.e. from posterior eye border to centre of indentation in operculum.

A REVISION OF THE RECENT ROUND HERRINGS 345

Two subspecies recognized here.

a. Gillrakers on lower part of first arch 24-37 ; scales in lateral series 44-49 ; total anal

rays 11-14 ; Indo-Pacific region excluding Samoa . Spratelloides gracilis gracilis

b. Gillrakers on lower part of first arch 19-23 ; scales in lateral series 41-42 ; total anal

rays lo-n ; restricted to Samoa . . . Spratelloides gracilis atrofasciatus

Spratelloides gracilis gracilis Schlegel

Spratelloides gracilis Schlegel, 1846, Faun. Japan. Poiss., pts. 10-14 ; 238, pi. 108, fig. 2 (for
full synonymy, see under species).

DESCRIPTION AND DIAGNOSIS. Distinguished from the Samoan subspecies by a
higher gillraker count (24-37) an d more scales in lateral series (44-49). Dorsal ii 9-
12, pectoral i 10-14, pelvic i 7, anal 11-14 (including ii or usually iii simple rays).

DISTRIBUTION. As for species, but not found in Samoa.

TYPE. 5. gracilis Schlegel.

Spratelloides gracilis atrofasciatus Schultz

Spratelloides atrofasciatus Schultz, 1943, Bull. U.S. nat. Mus., No. 180 : 8, fig. i ; Idem, 1953,
op. cit., No. 202 : 24.

DESCRIPTION AND DIAGNOSIS. A Samoan population distinguished from 5. g.
gracilis by its lower gillraker count (19-23) and fewer scales in lateral series (41-42).
Dorsal ii 9-10, pectoral i 10-11, pelvic i 7, anal 10-11 (including ii or usually iii
simple rays). Description based on Schultz.

DISTRIBUTION. Samoa only.

TYPE. 5. atrofasciatus Schultz.

Spratelloides delicatulus (Bennett)
(Text-fig. 19)

Clupea delicatula Bennett, 1831, Proc. zool. Soc. London, 1 : 168 (Type locality : Mauritius).

Clupea macassariensis, Bleeker, 1849, Journ Indian Arch., 3 : 72.

Clupeoides macassariensis Bleeker, 1851, Natuurk. Tijdschr. Ned. Ind., 2 : 214 ; Idem, 1852, Verh.
Bat. Gen., Batavia, 24 : 17 ; Idem, 1852, Natuurk. Tijdschr. Ned. Ind., 3 : 772.

Alausa alburnus, Kner, 1867, Sitzb. K. Akad. Wiss. Wien, 54 : 387, pi. i, fig. 16.

Spratelloides alburnus Giinther, 1868, Cat. Fishes Brit. Mus., 7 : 464.

Stolephorus alburnus Fowler, 1941, Bull. U.S. nat. Mus., No. 100 : 565.

Spratelloides delicatulus Giinther, 1868, Cat. Fishes Brit. Mus., 7 : 464 ; Bleeker, 1872, Atlas.
Ichth. Ind. Neerland., 6 : 96, pi. 264, fig. 3 ; Giinther, 1910, /. Mus. Goddefroy, Hamburg, 6 :
383 ; Weber and Beaufort, 1913, Fishes Indo-Aust. Arch., 2 : 20 ; Gilchrist and Thompson,
1917, Ann. Durban Mus., 1 : 296 ; Barnard, 1925, Ann. S. Afr. Mus., 21 (i) : no ; Harden-
berg, 1933, Treubia, 14 (2) : 216 ; Roxas, 1934, Philipp. J. Sci., 55 : 249 ; Bertin, 1943, Bull.
Inst. ocdanogr. Monaco, No. 853 : 18 ; Marshall, 1950, Bull. Raffles Mus., No. 22 : 168 ; Idem,
1952, Bull. Brit. Mus. nat. Hist. (Zool.}, 1 : 222 ; Schultz and Wellander, 1953, Bull. U.S.
nat. Mus., No. 202 : 26 ; Morrow, 1954, Ann. Mag. nat. Hist., (12) 7 : 804 ; Randall, 1955,
Atoll Res. Bull., No. 47 : 6 ; Fowler and Steinitz, 1956, Bull. Res. Counc. Israel, 5, 13 (3-4) :
262 ; Rofen, 1958, Nat. Hist. Rennell Is., Brit. Solomon Is. 1, Copenhagen : 151 ; Jones,
1960, /. Mar. biol. Ass. India, 2 (i) : 103 ; Idem, 1961, op. cit., 2 (2) : 267.

346 P. J. P. WHITEHEAD

Stolephorus delicatulus Jordan and Scale, 1906, Bull. Bur. Fisher., 25 : 186 ; Evermann and Scale,
1906, Bull. Bur. Fisher., 26 : 53 ; Fowler, 1928, Mem. Bernice P. Bishop. Mus., 10 : 29 ; Whit-
ley, 1929, Proc. Linn. Soc. N.S.W., 54 : 92 ; Herre, 1936, Field Mus. nat.Hist. Zool.,21 : 32 ;
Fowler, 1941, Bull. U.S. nat. Mus., No. 100 : 562 ; Smith, 1955, Ann. Mag. nat. Hist., (12) 8 :

3<>7-
Spratelloides robustus Ogilby, 1897, Proc. Linn. Soc. N.S.W., 22 : 64 (Type locality: coast of

New South Wales) ; Bertin, 1943, Bull. Inst. oceanogr. Monaco, No. 853 : 16.
Stolephorus robustus Waite, 1904, Mem. New South Wales Nat. Club, No. 2 : 12 ; McCulloch,

1920, Rec. Austral. Mus., 13 (2) : 42, pi. n, fig. i ; McCulloch and Whitley, 1925, Mem.

Queensland Mus., 8 (2) : 131 ; McCulloch, 1927, Fishes of New South Wales, ed. 2 : 16, pi. 4,

fig. 5ia ; Blackburn, 1941, Bull. Counc. sci. ind. Res. Aust., No. 138 : 59.

A fuller synonymy is given by Fowler (1941) under Stolephorus delicatulus,
S. robustus and S. alburnus.

Notes on Synonymy

The problem of Stolephorus has already been discussed. Spratelloides alburnus
(Kner), erroneously described from " Valparaiso, Chile " in fact from Samoa is
almost certainly 5. delicatulus, which has been recorded from the Marshall Is. and
may well occur to the south (Schultz & Wellander, 1953). Bertin (1943) reached the
same conclusion.

The only major change from previous synonymies is the inclusion of 5. robustus.
In proportional measurements and in all meristic counts except scales, the specimens
of 5. robustus in the Museum (all from New South Wales) fall within the ranges of
5. delicatulus (see Table III), and I have been unable to find any colour differences.
Unfortunately insufficient specimens of either species have had a full series of scales,
and the descriptions of Fowler (1941), Bertin (1943) and Schultz & Wellander (1953)
vary somewhat (respectively 32-36, 35-38, 40-42 in lateral series for 5. robustus,
and 36-40, 43-45, 45-46 for S. delicatulus). All are agreed however that the New
South Wales population has the lower count, and on this basis I believe that it
should be considered a subspecies of S. delicatulus. There is also a tendency for
the New South Wales population to have a slightly longer pelvic fin (Table III)
and slightly more rays in the pectoral (Text-fig. 17) but the differences are small.

FIG. 19.

Spratelloides delicatulus. From a specimen 69 mm. standard length, Hasler Collection

(locality not stated). Scales omitted.

A REVISION OF THE RECENT ROUND HERRINGS 347

DESCRIPTION. Based on thirteen fishes (S. delicatulus), 39-0-52-0 mm. standard
length from the Maldives, Thousand Is., Bonham Is., and the Seychelles ; and ten
fishes (S. robustus) 50-3-71-0 mm. standard length from the coast of New South
Wales. Additional meristic counts made on other specimens listed under Study
Material (p. 376). Single measurements outside normal range are placed in
parenthesis.

In percentages of standard length : body depth (16-7) 17-9-21-0, head length 1
24-0-28-2, snout length 6-6-7-7, e Y e diameter 6-5-7-7, post-orbital distance 8-8-10-3,
maxilla length 8-4-9-6, pectoral length (11-9) 14-0-15-8, pelvic length 9-2-12-8,
pre-dorsal distance 45-6-49-2, pre-pelvic distance (50-5) 53-0-57-5, pre-anal distance
(75-0) 76-2-83-5.

Body slightly compressed, rounder than in 5. gracilis, its depth less than head
length. Snout pointed, a little larger than eye diameter. Maxilla longer than
snout, reaching almost to pupil of eye. Post-orbital a little longer than snout.
Dorsal origin equidistant, or usually nearer snout than caudal base. Pelvic origin
under mid-dorsal or under second half of dorsal, a little nearer to caudal base than
to snout.

Dorsal ii 9-11, pectoral i 10-13 ( see Text-fig. 17), pelvic i 7, anal 9-11 of which ii
or iii are branched (see Text-fig. 16).

Gillrakers 26-33 on the lower part of the first arch (including one at angle), 9-11
on upper part.

Scales in lateral series 32-46, 7-9 transverse.

Vertebrae 42-44 (Bertin, 1943).

TYPE. Mauritius.

COLOUR IN ALCOHOL. Upper surfaces grey-blue or brown, sides and ventral
surfaces white or silvery, the two areas meeting at an abrupt line dorso-laterally.
Top of head, tip of snout, lower jaw and tongue dark brown and a small black spot
in front of eye. Two black streaks along glandular scales on each lobe of caudal.

DISTRIBUTION. Eastern coast of Africa, from Natal, northwards to Gulf of Aden
and Red Sea ; India, East Indies, Philippines, Cocos-Keeling Is., Hawaii ; New
South Wales, Queensland, Tasmania. S. delicatulus apparently does not penetrate
as far north as S. gracilis, but reaches further south (i.e. to Australia).

Two subspecies recognized here.

a. Scales in lateral series 36-46 ; pelvic fins 9-2-11-5 % of standard length ; pectoral

finrays i 10-12 ; Indo-Pacific region excluding Australia

Spratelloides delicatulus delicatulus

b. Scales in lateral series 32-42 ; pelvic fins in-i2'8% of standard length ; pectoral

finrays i 11-13 ; confined to Australian coasts

Spratelloides delicatulus robustus

Spratelloides delicatulus delicatulus Bennett

Spratelloides delicatulus Bennett, 1831, Proc. Comm. zool. Soc. London, 1 : 168 (for full synonymy,
see species) .

1 Measurements as in S. gracilis, footnote, p. 344.

348 P. J. P. WHITEHEAD

DESCRIPTION AND DIAGNOSIS. Distinguished from the Australian subspecies by
its greater number of scales in lateral series, although authors are not agreed on the
precise range in this character (see p. 346). It also has slightly shorter pelvic fins
(9'2-ii'5% of standard length) and fewer pectoral rays (i 10-12).

DISTRIBUTION. As for species but excluding the coasts of Australia.

TYPE. Spratelloides delicatulus (Bennett.)

Spratelloides delicatulus robustus Ogilby

Spratelloides robustus Ogilby, 1897, Proc. Linn. Soc. N.S.W., 22 : 64 (see under species for full
synonymy).

DESCRIPTION AND DIAGNOSIS. An Australian population differing from 5. d.
delicatulus in having fewer scales in lateral series (32-42), slightly longer pelvic fins
(11-1-12-8% of standard length), and slightly more rays in the pectoral (i 11-13).

DISTRIBUTION. Queensland, New South Wales, Victoria, Tasmania (after Fowler,
1941).

TYPE. Coast of New South Wales, Australia. Australian Museum, Sydney,
No. i. 3668.

JENKINSIA Jordan & Everman

Jenkinsia Jordan and Evermann, 1896, Bull. U.S. nat. Mus., 50 : 418 (genotype Dussumieria
stolifera Jordan and Gilbert).

DESCRIPTION. Body elongate, slightly compressed, rounded ventrally ; snout
pointed, jaws equal or lower slightly projecting. A single (posterior) supra-maxilla,
paddle-shaped with a slender anterior shaft ; maxilla toothed. Premaxilla with or
without a single series of fine conical teeth. Little or no development of fleshy
eminence (cleithral flap) at postero-ventral angle of gill-opening. Posterior border
of operculum strongly indented, more so than in Ehirava. Sub-operculum rec-
tangular, posterior border not rounded but forming a well-defined and slightly obtuse
angle. Inter-operculum exposed and three or four branchiostegal rays visible
externally. Posterior border of pre-operculum vertical.

A pair of posterior fontanelles becoming reduced in size in adults. Anterior
frontal fontanelle present in smaller fishes, sometimes in adults.

A single w-shaped pelvic scute. No abdominal scutes. Dorsal a little nearer
snout than caudal base. Pelvic origin below mid-dorsal.

A bright silvery midlateral stripe along flanks.

The Species of Jenkinsia

Authors have recognized variations in finray and gillraker counts amongst speci-
mens of Jenkinsia from different parts of the Caribbean region, and several species
have been based on these. But Parr (1930), Beebe & Tee- Van (1933) and Longley
& Hildebrand (1941) all believed that the overlap between finray counts in such
populations was too great for any one population to be separated at specific level.

A REVISION OF THE RECENT ROUND HERRINGS

349

20
19

18
17

16
15
14
13
12

o
o

A A

25 30 35 40 45 50 55 60
Standard length (mm.)

FIG. 20.

Body depth expressed as percentage of standard length (ordinate) plotted against standard
length (abscissa) in various populations of Jenkinsia.

J. lamprotaenia

X Jamaica (types of /. lamprotaenia).

Cat Cay, Bahamas.

Key West (types of Dussumieria stolifera) .
O St. John, Virgin Islands.

+ Venezuela (figures from Martin, 1955).

J. majua

J^ Campeche Banks.

A Swan Island.

V Gun Cay, Bahamas.

I have found however, that the populations in this region can be split into two
natural groups on the basis of two correlated non-meristic characters. The first is
the presence or absence of teeth on the premaxilla ; the second is in the form of the
frontal bones at the anterior end of the posterior fontanelles. In the group lacking
premaxillary teeth the fontanelles are very narrowly divided even at their anterior
end (and this is especially the case in juveniles), and in front of these two fontanelles

P. J. P. WHITEHEAD

(27) Florida

(45) Cat Cay

(2) Jamaica

(21) Virgin Is.

(6) Venezuela

Quita Sueno Bank
Gun Cay

Swan I. area

Campeche Banks

ii9 ii 10 ii

Dorsal rays

13 14 15
Total anal rays

FIG. 21.

Dorsal and anal finray count frequencies in populations of Jenkinsia.
Black, J. majua ; White, /. lamprotaenia.

Numbers in sample placed in parentheses. Range only for Bermuda specimens (based on
published records see text).

the part of the head immediately behind the junction of the two inner arms of the
transverse frontal sensory canal is domed or flat (again, especially in juveniles)
(Text-fig. 320). In the other group however, the posterior fontanelles are divided
anteriorly by a broad wedge, and within this wedge is a triangular depression in
which lies a median branch of the post-frontal canal (Text-fig. 32a).

There is also a difference in the size of the fish at which the frontals start to extend
posteriorly to close the fontanelles. In the first group this occurs at a larger size
than in the second, and in this the former resemble Spratelloides , while the second
group resemble rather the genera of the Ehiravini.

These two differences are important because they also help to separate the
Ehiravini from Spratelloides. But for the geographical isolation of Jenkinsia, a
direct evolution of the Ehiravini from those members of Jenkinsia which have
premaxillary teeth and a wedge of bone separating the posterior fontanelles would

A REVISION OF THE RECENT ROUND HERRINGS 351

seem clear. As it is, Jenkinsia may have become isolated at a time when Spratel-
loides had only just begun to diverge (through loss of premaxillary teeth, develop-
ment of the cleithral flap and retention of the second supra-maxilla) .

A third character which can be correlated with the tooth and fontanelle characters
is the higher number of rays in the anal fin in those fishes with a toothed premaxillary.
In addition these fishes also tend to be deeper bodied (Text-fig. 20) although body
depth shows positive allometry with standard length and the difference is difficult
to define when all size groups are considered.

A final difference between populations of Jenkinsia involves gillraker numbers.
This is here interpreted as a subspecific character.

a. Premaxilla toothed ; posterior fontanelles divided anteriorly by wedge of bone, with

triangular depression in front ; anal rays 13-16 ; body deeper, its depth usually

over 15% of standard length ......./. lamprotaenia

b. Premaxilla edentulous ; posterior fontanelles narrowly divided anteriorly, area in

front domed, becoming flat in adults ; anal rays 11-13 > body more slender, its
depth usually under 15% of standard length ..... /. majua sp. nov.

The geographical distribution of these two species overlaps in the Bahamas (where
I have examined specimens of the first from Cat Cay, and of the second from the
nearby Gun Cay) and in the Gulf of Campeche. Further distributional records are
required, but on the basis of the present specimens, /. lamprotaenia seems to occur
mainly along the outer boundary of the area (i.e. Bermuda, the islands of the West
Indies and Antilles, Venezuela), while /. majua appears to be a more western species
(Gulf of Campeche, Swan Island, British Honduras).

Jenkinsia lamprotaenia (Gosse)
(Text-fig. 23)

Clupea lamprotaenia Gosse, 1851, Naturalist's Sojourn in Jamaica : 291, pi. I, fig. 2 (Type
locality: Jamaica).

(Spratelloides) lamprotaenia Giinther, 1868, Cat. Fish Brit. Mus., 7 : 465.

Dussumieria stolifera Jordan and Gilbert, 1884, Proc. U.S. nat. Mus., 7 : 25 (Type locality : Key
West, Florida).

Jenkinsia stolifera Jordan and Evermann, 1896, Butt. U.S. nat. Mus., No. 47 (pt. i):4i8;
Fowler, 1930, Proc. biol. Soc. Wash., 43 : 145.

Jenkinsia lamprotaenia Jordan and Evermann, 1896, Bull. U.S. nat. Mus., No. 47 : 419 ; Parr,

1930, Bull. Bingham oceanogr. Coll., 3 (4) : 3 ; Beebe and Tee-Van, 1928, Zoologica, 10 (i) : 43 ;

Idem, 1933, Zoologica, 13 (7) : 136 ; Longley and Hildebrand, 1941, Pap. Dep. mar. Biol.

Carnegie, 34 (No. 535) : 12 ; Fowler, 1944, Monagr. Acad. nat. Sci. Philad., No. 6 : 123 ;

Martin, 1955, Mem. Soc. Cienc. nat. La Salle, 15 : 185.

Stolephorus viridis Bean, 1912, Proc. biol. Soc. Wash., 25 : 122 (Type locality : Bermuda).
Jenkinsia bermudana Rivas, 1946, Smithson. Misc. Coll., 106 (14) : 2, fig. I, pi. I (Type locality:

Bermuda).
Jenkinsia viridis Collette, 1962, Copeia, No. 3 : 659.

352

(7)
(22)

(10)

(32)

(27)
(32)
(II)

(33)

(15)
(6)

P. J. P. WHITEHEAD
Gun Cay

Swan I. area

Quita Suerio Bank

Campeche Gulf
^ Bermuda

Florida

Cat Cay

Jamaica, Cuba

' 1 G. of Campeche, Yucatan, Brit. Honduras

Virgin Is.

Venezuela

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Gillraker counts
FIG. 22.

Gillraker count frequencies in populations of Jenkinsia.
Black, /. majua ; White, /. lamprotaenia.

Numbers in sample placed in parentheses. Rakers counted on lower part of first arch,
including one at angle when present.

Note on Synonymy

As already outlined under the generic notes, two principal non-meristic characters
are here used to define the species. I have examined the types of Clupea lampro-
taenia and Dussumieria stolifera, and in both cases premaxillary teeth are present,
and the posterior fontanelles are divided anteriorly by a wedge of bone, in front of
which is a triangular depression. I have also examined specimens from Cat Cay
(Bahamas), St. John (Virgin Islands), Monroe County (Florida), the Cayman Islands,
Cuba, British Honduras, and the Gulf of Campeche, and these agree with the
types. Recently Collette (1962) has re-examined Bean's four type specimens of
Stolephorus viridis and has shown that in fact two species are represented. The
first species, which has a gillraker count of 37 and 39, he believes to be the only
Bermuda species ; the second, represented by two rather damaged specimens with
gillraker counts of 30 and 32, he places with /. lamprotaenia and, since there is no
other evidence of this species in Bermuda, believes the specimens to have been
included in error. He places /. bermudana Rivas in the synonymy of /. viridis,
and informs me (in litt.) that the types of both these two species have premaxillary

A REVISION OF THE RECENT ROUND HERRINGS 353

teeth. Since the presence of premaxillary teeth is so exactly correlated with the
fontanelle character there seems little reason to separate the Bermuda species from
/. lamprotaenia at specific level ; I agree however that the specimens with low gill-
raker counts from Bermuda may well have come from another locality.

The pre-maxillary teeth are usually obvious and I have seen them in the smallest
specimen available, a fish of only 18 mm. Although the teeth may be deciduous,
the complete series rarely occurring, some can always be found.

On the basis of published descriptions, the fishes from Bermuda and from Venezuela
can clearly be distinguished on gillraker counts (see Text-fig. 22), both having similar
and much higher counts than the rest. In numbers of dorsal and anal rays however,
there is no evidence (Text-fig. 21) that any one population of /. lamprotaenia differs
significantly from the rest, although no fishes with only ii 9 dorsal rays have been
recorded from Jamaica or Cat Cay. The latter two populations also agree in being
more slender (Text-fig. 20 small dots and crosses) but more specimens are required
before they can be positively separated from the Florida and other populations. In
turn, the Florida fishes have slightly larger eyes (9-1-11-5% of standard length in
the size group 31-36 mm. ; compared with 7-7-8-9 in specimens from Jamaica and
Cat Cay of 31-55 mm.), but fishes from the Virgin Islands are intermediate and
closely resemble the Jamaica specimens (8-5-9-0% in fishes of 37-43 mm.).

Thus on the basis of the present collections there is reason to believe that only the
Bermuda and Venezuela populations have diverged sufficiently to be recognized as
a distinct subspecies.

xgtfSrgfi' '?"$$?&?

FIG. 23.

Jenkinsia lamprotaenia. From a specimen 51 mm. standard length, Cat Cay, Bahamas.

Scales omitted.

DESCRIPTION. Based on the types of Clupea lamprotaenia (54-0 and 55-8 mm.)
from Jamaica, ten fishes (46-7-52-4 mm.) from Cat Cay (Bahamas), five of the types
of Dussumieria stolifera (31-0-35-7 mm.) from Key West (Florida), ten fishes
(33-9-36-2 mm.) from Monroe County (Florida), and five fishes (37'7~43'3 mm.)
from St. John (Virgin Islands). Meristic counts include 101 specimens (see Text-
figs. 20-22).

In percentages of standard length : body depth 14-4-19-1, head length 1 23-5-28-7,
snout length 7-3-9-1, eye diameter 7-7-11-5, post-orbital distance 7-9-9-6, pectoral

1 Measured to indentation of posterior margin of operculum, i.e. a horizontal measurement ; cf.
Spratelloides, p. 344.

ZOOL. 10, 6 20

354 P. J- P. WHITEHEAD

length 12-8-16-8, pelvic length 10-7-13-9, pre-dorsal distance 44-8-50-1, pre-pelvic
distance 52-2-58-5, pre-anal distance 71-0-82-0.

Body moderately compressed, its depth much less than head length. Snout
pointed, equal to eye (except in Florida and Virgin Islands specimens). Maxilla
reaching beyond front border of eye, but not to pupil. Post-orbital equal or larger
than snout, larger than eye (except in Florida and Virgin Islands specimens).

Posterior fontanelles broadly separated by wedge of bone (frontals) anteriorly, as
shown in Text-fig. 32a. A shallow triangular depression in front of fontanelles
containing a branched posterior portion of the transverse frontal sensory canal.

Premaxilla with a single series of conical pointed teeth, often deciduous. Dorsal
origin a little nearer snout than caudal base or equidistant. Pelvic origin under
middle of dorsal, nearer to caudal base than to snout. Anal origin equidistant
between pelvic origin and caudal base.

Dorsal ii 9-11, pectoral i 11-13, pelvic i 7, anal ii-iii 10-13 (total 13-16).

Gillrakers 20-31 on the lower part of the anterior arch (including one at angle).

Scales in lateral series approximately 35-40.

Vertebrae 38 (in a type specimen of Dussumieria stolifera) .

SYNTYPES. Jamaica. B.M. (N.H.), 1962.7.19.3-4.

COLOUR IN ALCOHOL. Back and sides light brown, with a double line of melano-
phores down back and extending forwards along frontals to premaxillary symphysis.
A broad silvery band along flanks, almost as wide as eye diameter. Lower flanks
and belly light brown or cream. Pigment at base of caudal fin.

DISTRIBUTION. Bermuda, Bahamas, Florida ; Cuba, Jamaica to Puerto Rico and
Virgin Islands ; Gulf of Campeche and British Honduras ; Venezuela (see Martin,

1955).

Two subspecies are described here, the Bermuda and Venezuela fishes being
separated from the rest by their high gillraker count. Although the latter two
populations are also well separated geographically from each other (as in the case of
the Red Sea and China populations of Dussumieria), they both differ so sharply
from the remainder that I am forced to recognize them as distinct ; future work
may well show differences between the two.

Jenkinsia lamprotaenia lamprotaenia (Gosse)

Clupea lamprotaenia Gosse, 1851, Naturalists Sojourn in Jamaica : 291, pi. i, fig. 2 (Type
locality : Jamaica) . (For remainder of synonymy, see under species for Dussumieria stolifera
and Jenkinsia lamprotaenia) .

DIAGNOSIS. Distinguished from Bermuda and Venezuela populations by its lower
gillraker count (20-24) as shown in Text-fig. 22.

TYPE. J. lamprotaenia (Gosse), Jamaica (syntypes in Brit. Mus. (Nat. Hist.)).

Jenkinsia lamprotaenia viridis (Bean)

Stolephorus viridis Bean, 1912, Proc. biol. Soc. Wash., 25 : 122 (Type locality : Bermuda).
Jenkinsia lamprotaenia Hollister, 1936, Zoologica, 21 (4) : 276, figs. 40-44 (Caudal skeleton).
Jenkinsia bermudana Rivas, 1946, Smithson. Misc. Coll., 106 (14) : 2, fig. I, pi. i (Type locality :
Bermuda).

A REVISION OF THE RECENT ROUND HERRINGS

355

DIAGNOSIS. Distinguished from above by its higher gillraker count (26-31) as
shown in Text-fig. 23.

TYPE. Stolephoms viridis Bean, Bermuda (U.S.N.M. 74084).

FIG. 24.

Jenkinsia majua sp. nov. Holotype, 53-8 mm. standard length.
(Drawn by Miss Carolyn Reusch.)

Jenkinsia majua sp. nov.
(Text-fig. 24)

DESCRIPTION. Based on the type and seventeen other fishes (457-53-8 mm.)
from Areas Cay, Campeche Bank, six fishes (29-3, 36-1-48-4 mm.) from Swan Island,
and five fishes (30-1-38-2 mm.) from Gun Cay, Bahamas.

In percentages of standard length : body depth (10-6) 11-3-14-9, head length
20-5-24-4, snout length (6-6) 7-1-8-3, eye diameter 5-3-7-6, post-orbital distance 7-2-
9-2, pectoral length 10-3-12-3, pelvic length 8-4-11-8, pre-dorsal distance 45-5-51-0,
pre-pelvic distance 49-0-57-5, pre-anal distance 73-0-81-3.

Body moderately compressed, its depth much less than head length. Snout
pointed, greater than eye diameter. Maxilla reaching to just beyond anterior eye
border, but not to pupil. Post-orbital roughly equal to snout length.

Posterior fontanelles narrowly divided by arm of supra-occipital ; a small wedge
of bone only between anterior ends of fontanelles (see Text-fig. 32b) ; f rentals domed
or flat in front of fontanelles.

Pre-maxillary teeth either absent, or so deciduous that none have been found.

Dorsal origin a little nearer snout than caudal base or just equidistant. Pelvic
origin below middle of dorsal, nearer to caudal base than to snout, and nearer anal
origin than pectoral base. Anal origin equidistant between pelvic origin and caudal
base.

Dorsal ii 9-11, pectoral i 11-12, pelvic i 7, anal ii-iii 9-11 (total 11-13).

Gillrakers 21-28 on lower part of anterior arch (including one at angle when
present).

Scales in lateral series approximately 35-40.

COLOUR IN ALCOHOL. Uniform light brown with broad lateral silvery streak a
little narrower than eye. Two lines of melanophores down back. Base of caudal
pigmented.

356 P. J. P. WHITEHEAD

DISTRIBUTION. Gulf of Campeche ; Swan Island and Gun Cay (Bahamas) ; and
NW. coast of Caribbean Sea.

I would like to record here my gratitude to Dr. Loren P. Woods of the Chicago
Natural History Museum for his generosity in allowing me to describe this new
species after he had himself begun a preliminary description, and also for so will-
ingly making his material available to me. I have retained the specific name majua
chosen by him, which is the Cuban name for Jenkinsia spp.

Specimens of J. majua from Gun Cay and from localities in the Caribbean have a
lower gillraker count than do those from the type locality, Campeche Bank (see
Text-fig. 22). The difference is one of modal numbers, but the overlap between
the two groups is slight, three specimens from Campeche Bank having a low count.
The Campeche population seems to be sufficiently isolated from the rest (at least as
far as the available material can indicate) for this difference to be given sub-specific
rank. No other differences have been found.

a. Gillrakers on lower part of anterior arch 24-28, mode 26 . J- vnajua majua

b. Gillrakers on lower part of anterior arch 21-24, mode 23 . . /. majua woodsi

Jenkinsia majua majua

Distinguished by a slightly higher range and modal gillraker count. Apparently
confined to the Gulf of Campeche and an area to the north of Yucatan.
TYPE. J. majua (see list of Study Material).

Jenkinsia majua woodsi subsp. nov.

Distinguished from the above by a slightly lower range and modal gillraker count.
Distributed in the Bahamas (Gun Cay), and the N.W. coast of the Caribbean (British
Honduras, Swan I., Quita Suefio and Serrano Cays).

Named for Dr. Loren P. Woods.

HOLOTYPE AND PARATYPES. See list of Study Material.

SYSTEMATIC CHARACTERS

In the following section are discussed certain dussumieriid features which are of
value in defining subfamilies, tribes, genera and species. Some, such as scutes,
have not been utilized before ; others have either been missed or have received
passing mention only.

Scutes

Although the round herrings generally lack scutes, all possess in one form or
another a single scute lying immediately in front of the pelvic fin. This scute
was figured by Chapman (1948) for Etrumeus. It is essentially similar to the w-
shaped scute shown here for Spratelloides delicatulus (Text-fig. 25) and it is found
also in Dussumieria and in all members of the tribe Sprat elloidini. Chapman (loc.
cit.} shows a second, triangular scute just between the pelvic fin bases in Etrumeus ;

A REVISION OF THE RECENT ROUND HERRINGS

357

I have not found this in Dussumieria nor in the Spratelloidini. The anterior scute
has been referred to as a " modified " scute (Chapman loc. cit., Whitehead ig62a), but
this may imply an erroneous direction to scute evolution (see p. 367 below).

A second type of pelvic scute is found in the Ehiravini. This scute more closely
resembles the typical clupeid scute, having lateral ascending arms rising just in front
of the pelvic fins, but the central portion of the scute is rounded, not keeled (Text-
fig. 26). The ascending arms are very thin, and will conform to the body contours
in wrinkled specimens. This pelvic scute was noticed by Bertin (1943) in Gilchristella
and Sauvagella and figured in the latter, but he referred to it as " I'ecaille verticale qui
precede cette nagoire ". The arms of the scute lie below the scales and appear to be
embedded in the skin.

In Gilchristella aestuarius alone there are also six to nine similarly spined, but
slightly shorter scutes lying between the pectoral and pelvic fin bases (Text-fig. 27).
All the specimens examined lacked a scute immediately in front of the pelvic scute.
In no case were post-pelvic scutes present. These abdominal scutes are also rounded
ventrally, not keeled. They are again very thin and are easily overlooked unless
the specimen is stripped of scales and thoroughly dried.

FIG. 25.

The w-shaped pelvic scute (stippled) in Spratelloides delicatulus.
shown by broken lines.

Ventral view, pelvic finrays

Previous records of abdominal scutes in the round herrings have all proved
erroneous. Thus Fowler (1941) was correct to presume that the 9 pre-pelvic and
7 post-pelvic scutes shown by Sauvage (1891) for Gilchristella madagascariensis were
an artist's error : the entire figure is a poor one and there are no such scutes in the
types (see Bertin, 1943). The " hard sharp keel on the thoracic region in front of
the pectorals " described by Barnard (1925) for G. aestuarius is not due to scutes,

358

P. J. P. WHITEHEAD

but is the sharp keel formed by the ventrally apposed edges of the two coracoids ;
it occurs in other species also. The scuted Pellonulops Smith was almost certainly
based on a clupeid.

Taken in conjunction with other characters (especially numbers of branchiostegal
rays) the " modified " w-shaped pelvic scute may perhaps be the primitive form
from which fully scuted species have arisen. To some extent the variation in scute
development in the round herrings resembles that of the anchovies, where, however,
ventral scutes are more common. In the most widespread (and perhaps most

FIG. 26.
Left pelvic scute in Ehirava malabaricus (stippled) showing ascending lateral arm.

primitive) anchovy, Engraulis encrasicholus (and its allied forms in Australia,
America and Japan) there is a single pelvic scute only, which is of the spined,
5. malabaricus type. This is the case also in all the American anchovies. Of
the more specialized anchovies of the Indo-Pacific region, species of Stolephorus have
up to six or seven keeled scutes, each with a backwardly directed spine between the
pectoral and pelvic bases ; Thrissocles baelama has both pre- and post-pelvic scutes ;
and other species of Thrissocles, as well as Setipinna, have in addition a few scutes
in front of the pectorals also. In the Clupeidae scutes always appear to be present
both in front of and behind the pelvic fins. Engraulidae and Clupeidae both have
the spined pelvic scute and it seems clear that the evolution of this pelvic scute
preceded the evolution of the other abdominal scutes. The pelvic scute may perhaps
be the most important functionally, possibly lending support to the pelvic fin ; thus
in specimens from all three families I have found a small ligament arising from a
point about half way along the posterior edge of the spine of the pelvic scute. This
ligament is attached near the base of the outer (or upper) half of the first pelvic ray.
It does not occur in those round herrings which have a w-shaped pelvic scute. The

FIG. 27.

Pelvic, and seven pre-pelvic scutes (stippled) in Gilchristella aestuarius.
Pectoral and pelvic fins indicated by broken lines.

Left lateral view.

A REVISION OF THE RECENT ROUND HERRINGS 359

presence of the ligament suggests that the spined scute plays some role in the move-
ments of the pelvic fins. The abdominal scutes, however, may have had a very
different evolutionary history, although closely resembling the pelvic scute in their
modern form.

Mention should be made of Chapman's (1944) belief that the abdominal scutes in
the Engraulidae are derived from the ventral myorhabdoi (which occur from anus
to caudal base in engraulids ; and occur anterior to the anus but not behind it in
clupeids). This author argued that the possession of ventral scutes is a primitive
character in the clupeids " because of the theoretical necessity of the possession of
ventral myorhabdoi by antecedent clupeoids " ; he notes the absence of these
myorhabdoi in the Dussumieriidae (i.e. in Etrumeus}. Ontogenetic studies might
support this view, but such a direction to scute evolution contradicts the evidence
of several other characters, and especially that of branchiostegal ray number.

Upper and Lower Jaws

The lower jaw in the Dussumieriidae very closely resembles the lower jaw of the
Clupeidae arid there appear to be no consistent differences which could serve to
separate the two on this character. It is worth noting that the lower jaw in the
cretaceous Clupavus (i.e. C. neocomiensis (Bassani) figured by Arambourg, 1954) is
also identical to the normal clupeid condition.

The upper jaw elements, although very similar in the two families, are generally
narrower in the round herrings. In the Dussumieriinae in particular, the posterior
supra-maxilla is not more than half as deep as the maxilla and in Etrumeus it is
only about a quarter as deep (see Text-fig. 28d and e). In the Spratelloidinae the
posterior supra-maxilla is more paddle-shaped, the expanded portion being equal to
or a little deeper than the deepest part of the maxilla, but it is still a narrower bone
than in the clupeids, being at least five times as long as deep (see Text-fig. 28a, b and f
and compare with a fairly typical clupeid, Harengula ovalis, Text-fig. 28c).

A second, anterior supra-maxilla is present in some round herrings but not in
others. In the Dussumieriinae, Dussumieria has a second supra-maxilla, but
Etrumeus does not ; in the Spratelloidinae this bone is present in SprateUoides
gracilis and 5. delicatulus, but is absent in the Ehiravini as well as in Jenkinsia.
The importance of this little bone in the phylogeny of the group is questionable,
but it is worth noting that amongst the Spratelloidinae it is consistently absent in
the species with spined pelvic scutes, but present in fishes with w-shaped pelvic
scutes (except Jenkinsia}. When present, the second supra-maxilla is a thin and
plate-like bone lying just below the anterior shaft of the first supra-maxilla. Its
shape is somewhat variable.

The lower edge of the maxilla is in most dussumieriids fairly evenly rounded but
in E. malabaricus it is excavated anteriorly as in some clupeids. The lower edge
bears a single row of fine teeth along most of its length.

The premaxilla is a short and rather narrow bone in the Dussumieriidae. It bears
teeth in both Etrumeus and Dussumieria but of the Spratelloidinae only Jenkinsia
lamprotaenia and members of the Ehiravini have toothed premaxillae (species again

P. J. P. WHITEHEAD

FIG, 28.

Upper jaw elements in species of round herring. Maxilla, plain ; ist (anterior) supra-maxilla,
black ; 2nd (posterior) supra-maxilla, stippled.

a. Spratelloides delicatulus 41 mm.

b. Jenkinsia lamprotaenia 34 mm.

c. Harengula ovalis 112 mm. (Clupeidae).

d. Etrumeus teres 88 mm.

e. Dussumieria acuta 100 mm.

f. Ehirava malabaricus 58 mm.

which lack a second supra-maxilla). The premaxillary teeth and the anterior supra-
maxilla are probably primitive characters whose independent loss in some genera
but not others is to be expected.

Branchiostegal Rays

A principle division between the Spratelloidinae and the Dussumieriinae can be
made on numbers of branchiostegal rays (6-7 in the former, 14-20 in the latter). In
both groups the rays are borne on the epi- and the ceratohyal. In the Spratelloidinae
the ceratohyal is excavated ventrally at the head of each ray, as is the case in some
(if not all) Clupeids (see Chapman loc. cit.}, but in both Etrumeus and Dussumieria
the lower edge of the ceratohyal is smooth (Text-fig. 2Qa-d). The individual rays,
and especially the posterior ones, are broader in the Spratelloidinae than in the
Dussumieriinae.

There seems little doubt that generally speaking, reduction in the number of
branchiostegal rays in the clupeoid fishes has been a progressive evolutionary trend.
It is therefore tempting to consider all other differences between the Dussumieriinae

A REVISION OF THE RECENT ROUND HERRINGS

361

cer.h.

FIG. 29.

Hyoid bones and attachment of branchiostegal rays in species of round herring.

a. Spratelloides delicatulus 41 mm.

b. Jenkinsia lamprotaenia 34 mm.

c. Dussumieria acuta 100 mm.

d. Gilchristella aestuarius 53 mm.
ep.h. epihyal.

cer.h. ceratohyal.

br.st.r. branchiostegal rays.

and the Spratelloidinae as evidence of the primitiveness of the former, especially
where a series (such as the development of the " cleithral flap ") is concerned.
Certainly the evidence of branchiostegal ray numbers suggests the direction of scute
evolution, from a simple w-shaped pelvic scute to the spined scute and thus to the
fully keeled scutes of the Clupeidae.

Shape of Gill Opening and Bones of Opercular Series

In the series Etrumeus-Dussumieria-Spratelloides there is a steady development of
a flap of skin on the postero-ventral angle of the cleithrum (the cleithral flap) as
shown in Text-fig. 3oa-c. In Etrumem the posterior angle of the gill opening is an
approximate right-angle and there is usually no flap of skin. In Dussumieria the
posterior angle is more acute and individuals show varying degrees of development
of the flap. These two genera also differ in the posterior outline of the gill opening,
from almost straight and vertical in Etrumeus, to curved in Dussumieria. In
Spratelloides the flap of skin is raised markedly above the ventral line of the gill
opening and is indented anteriorly, while the posterior outline of the gill opening
is rather irregularly shaped, conforming to the indentation in the posterior margin
of the operculum, especially in juvenile fishes.

It is rather difficult to see what function the " cleithral flap " serves. Where it
is well developed, the ventral outline of the inter- and sub-opercular bones and the
branchiostegal rays is indented ; obviously, in order to seal the gill opening, one
development necessitates the other, but it is difficult to determine which developed

362

P. J. P. WHITEHEAD

FIG. 30.

The shape of the posterior outline of the gill opening, and of the bones of the opercular series
in the species of round herring. Gill cover slightly raised.

a. Etrumeus teres.

b. Dussumieria acuta.

c. Spratelloides delicatulus .
op. operculum.

s.op. sub-operculum.
p.op. pre-operculum.
cl. " cleithral flap ".

A REVISION OF THE RECENT ROUND HERRINGS 363

first. In Etrumeus the interoperculum and branchiostegal rays are barely visible
from the outside ; in Dussumieria they are more obvious, at least one branchiostegal
ray being visible ; in Spratelloides three or more branchiostegal rays can be seen,
and the penultimate ray is expanded posteriorly. In addition, in Spratelloides the
gill filaments of the first arch are indented round the cleithral flap.

In members of the tribe Ehiravini and in species of Jenkinsia the cleithral flap is
very poorly developed, usually approximating to a condition intermediate between
Etrumeus and Dussumieria. This is perhaps surprising since the Ehiravini are in
other characters closer to Spratelloides. The shape of the operculum and suboper-
culum in the Ehiravini are also of the dussumieriine type, the angle between these
two bones rising posteriorly, and not lying horizontally as in Spratelloides. The
cleithral flap is well-developed in the clupeid genera Harengula and Sardinella among
others and this is a further reason for expecting it to be large in the Ehiravini. As
suggested later (p. 368), the Ehiravini must be considered a group which, although
in several ways nearer to the Clupeidae than are the Spratelloidini, have none the
less retained some primitive features reminiscent of the Dussumieriinae.

As in the development of the cleithral flap, so in the shape of the operculum and
suboperculum a progressive change occurs between Etrumeus and Spratelloides, with
Dussumieria occupying an intermediate position. The most obvious trend is in the
angle of the line formed between the operculum and suboperculum. In Etrumeus
(Text-fig. 3oa) the lower edge of the operculum rises at a steep angle which, if
projected, would meet the dorsal body profile some distance before the dorsal fin.
The suboperculum is thus roughly triangular (with its apex pointing forwards). In
Dussumieria (Text-fig. 3ob) the lower opercular margin is nearer the horizontal and
the suboperculum more rectangular. This trend is continued in Spratelloides, with
a horizontal opercular margin and rectangular suboperculum (Text-fig. 3oc).

Together with this change is a progressive deepening of the operculum, and its
anterior border (i.e. junction with the preoperculum) becomes more vertical. The
ascending limb of the preoperculum thus becomes more upright.

To some extent the Etrumeus condition can be equated with that found in the
engraulids, and the Spratelloides condition to that found in the Clupeids. The
resemblance is even more marked when the posterior margin of the operculum is
also considered, for in Harengula at least, there is a tendency for the upper part of
the margin to be excavated, with a corresponding bulge in the posterior gill opening
profile. In Spratelloides, and especially in juveniles, this excavation of the operculum
is most marked, but it does not occur in Dussumieria or Etrumeus.

Other Features of the Skull

The skull of Dussumieria acuta was described and figured by Ridewood (1904),
who compared it with several other clupeoids. Chapman (1948) compared the
osteology of Etrumeus teres with that of the clupeid Sardinops caerulea and found
six principal differences, which were however shared with the Engraulidae. Some
of these differences disappear if the Spratelloidinae are included (numerous branchio-
stegal rays, smooth lower edge of ceratohyal and lack of ventral scutes). Again,
while 5. caerulea has a single articular head to the hyomandibular, the Dussumieriidae

364

P. J. P. WHITEHEAD

cannot be absolutely separated from the Clupeidae on this character since in Alosa
finta at least there are two articular facets (Ridewood, loc. cit., fig. 124).

A major difference between the Dussumieriinae and the Spratelloidinae is the
retention in adults of the latter of a pair of fontanelles immediately in front of the
supra-occipital. These have been referred to here as the posterior fontanelles, and
are shown in Text-figs. 31 and 32. They are bounded anteriorly and laterally by
the frontals, and are divided in the midline by a narrow median extension of the
supra-occipital, which also forms the posterior margin of the fontanelles.

ant. font,
s. orb. can.
trans, front, can.

temp. for.
p. fr.
post. font.

FIG. 31.

Spratelloides delicatulus, dorsal view of cranium showing the anterior and posterior fontanelles
and the temporal foramen. From an alizarin stained specimen of 40 mm.

ant. font, anterior fontanelle.

5. orb. can. supra-orbital canal.

trans, front, can. transverse frontal canal.

temp. for. temporal foramen.

p.fr. posterior extension of the frontals.

post. font, posterior fontanelle.

No such fontanelles are present in the adults of Dussumieria or Etrumeus, but in
a juvenile of Dussumieria of 49 mm. the fontanelles were present and together
comprised an area equal to the pupil of the eye.

As has already been pointed out (see p. 349) the form of the fontanelles and of the
posterior part of the two frontals is different in the two species of Jenkinsia, one
resembling the condition found in Spratelloides and the other that in the Ehiravini.

A REVISION OF THE RECENT ROUND HERRINGS

365

Thus in J. majua and in species of Spratelloides, the posterior median portion of the
frontals, while forming a small wedge between the two fontanelles, is not depressed
into a shallow triangular hollow, but in juveniles is domed.

In both types of fontanelle formation, the length of the fontanelles decreases with
the size of the fish, but more so in /. lamprotaenia and the Ehiravini. In large
specimens of Ehirava malabaricus (of about 60 mm.) the fontanelles are barely
apparent. This suggests that the second type of fontanelle formation (i.e. those
which are broadly divided anteriorly) is the more primitive of the two. In this
respect /. lamprotaenia and the Ehiravini show greater affinities with the Dussu-
mieriinae than do species of Spratelloides or J. majua.

In both the Spratelloidinae and the Dussumieriinae there is a transverse commis-
sure linking the supra-orbital sensory canals a little behind the centre of the orbits
(see Text-fig. 32). This has been referred to here as the transverse frontal canal

s.orb.con.

br.cut.can.

s.occ.

FIG. 32.

trons.fr.can.

Two types of posterior fontanelle (black) found in members of the Spratelloidinae. Dorsal
view of head with skin removed to expose fontanelles.

a. Jenkinsia lamprotaenia (at 35 mm.) showing large wedge of bone (part of frontals) dividing
anterior part of fontanelles. Within the depressed area of this wedge lies a branched, cutaneous
sensory canal.

b. Jenkinsia majua (at 40 mm.) showing posterior fontanelles only narrowly divided and with
no cutaneous sensory canal.

5. orb. can. supra-orbital canal.

p. fr. posterior extension of the frontals.

trans, fr. can. transverse frontal canal.

5. occ. supra-occipital.

br. cut. can. branched cutaneous canal lying in triangular depression.

366 P. J. P. WHITEHEAD

(the epiphysial canal of Weitzman, 1962) . The central part of this canal is cutaneous
and not roofed over with bone. Extending posteriorly from this canal there is, in
certain dussumieriids, a branched cutaneous canal which reaches to the apex of the
triangular depression mentioned earlier. This is found in the Dussumieriinae, in
the Ehiravini, and in Jenkinsia lamprotaenia, i.e. in those fishes which have a
pronounced triangular depression behind the transverse frontal canal. In
J. lamprotaenia and the Ehiravini this triangular depression is most marked ; it also
closely resembles the condition found in juvenile D. acuta. In adult dussumieriines
the triangular depression reaches to the posterior border of the skull, the fontanelles
having been occluded by posterior growth of the frontals.

Although the posterior branch leading from the transverse frontal canal appears
to be usually absent in those fishes in which the triangular depression is either
poorly developed or absent, there is a small canal in large specimens of Spratelloides.
The triangular depression and its relationship to the cephalic canals was recently
mentioned by Gunter & Demoran (1961) in dealing with the Gulf Menhaden
(Brevoortia patronus] ; they comment on the absence in the literature of any descrip-
tion of this part of the cephalic sensory system, and they refer to it as the " cephalic
spongy sensory area ".

The lateral edges of the triangular depression are usually well-defined. In some
dorosomatids and clupeids these lateral edges are wedge-shaped and bear a few or
many longitudinal striae (see Whitehead, ig62b) but this is only faintly apparent in
the dussumieriids (i.e. in Dussumieria).

SPECIATION IN THE DUSSUMIERIIDAE

The round herrings are an old group, with a lineage extending back at least to
Cretaceous times (if Clupavus is to be accepted as a dussumieriid). It is not surpris-
ing therefore to find a number of fairly well defined genera which can be clearly
separated on osteological grounds. But intrageneric variation is usually small and
I have here recognized one or at most two species only per genus, although in some
cases the species can then be subdivided into subspecies. The situation is however,
one in which the museum worker must decide in a rather arbitrary manner where
the boundary between species and subspecies and demes should lie.

If subspecies are defined as geographically isolated members of an interbreeding
population which differ consistently in at least one character from the rest of the
population, then there is a case for subdividing the forms of Dussumieria. But, as
already pointed out, the geographically and morphologically intermediate forms so
effectively bridge the gap between the extremes that no practical boundaries can be
drawn. There is thus reason to suppose that populations of D. acuta in China and
the Red Sea are reproductively linked, however tenuous that link may be.

In the case of forms which are distinct from each other but whose geographical
ranges overlap (e.g. in the two species of Spratelloides or Jenkinsia} there seems good
reason to believe the two forms to be separate species. But where the geographical
ranges do not overlap (as in Etrumeus) it would seem that the probable value (or
weight) of each particular morphological character must be assessed. Thus in
Etrumeus the differences between populations are small and are of the same order

A REVISION OF THE RECENT ROUND HERRINGS 367

as those found in populations of Dussumieria. In only a single case (the South
African population of Etrumeus) is there a clear-cut difference. But since this rests
solely on a single dorsal ray, it would be unrealistic to consider this of specific value
in view of the degree of variation in the rest of the genus.

Since there is as yet no generally applicable criterion of a bio-species other than
that it is reproductively isolated from its nearest relatives, the conclusions reached
in the museum must contain an element of guesswork. Thus the present classifica-
tion provides a framework, the details of which will no doubt be altered as larger
collections become available.

ORIGIN AND RELATIONSHIPS OF THE ROUND HERRINGS

In attempting to work out a probable phylogeny of the Dussumieriidae and their
relationship to the Clupeidae, it is possible to suggest morphological series for several
characters. The most obvious and perhaps most reliable trend is the reduction in
numbers of branchiostegal rays, with the Dussumieriinae the most primitive and the
Spratelloidinae the most advanced. If this is accepted, then it is difficult not to
correlate with it scute development, the w-shaped pelvic scute being the most
primitive, and the scutes of Gilchristella aestuarius being the most advanced form
in the Dussumieriidae as a whole. From G. aestuarius a fully scuted and keeled
clupeid is a logical development. In the shape of the gill opening, the development
of the cleithral flap, and the shape of the bones in the opercular series, there is a
good transition from Etrumeus, through Dussumieria to Spratelloides (see Text-fig.
3oa, b, c), and since these trends also coincide with those of branchiostegal rays and
scutes, it would seem at first sight as if the Ehiravini need only be added to the
series to make a perfect development through to the Clupeidae. But although the
Ehiravini complete the trend in scute development (and compression of the body),
and in addition complete also the trends for advance of the pelvics and increase in
length of anal, yet in other respects they more closely resemble the Dussumieriinae.
Thus the cleithral flap is small, the suboperculum is more nearly triangular, the
lower edge of the operculum is less horizontal and the premaxillae bear teeth.
But if scute development is a progressive trend, then the place of the Ehiravini seems
to be between the Spratelloidinae and the Clupeidae.

Also anomalous in some respects is Jenkinsia. Whereas J. majua could be placed
with Spratelloides but for the absence of an anterior supra-maxilla, /. lamprotaenia
shares with the Ehiravini a toothed premaxilla and a similar shape and development
of the posterior fontanelles. Both species however lack the large cleithral flap of
Spratelloides and thus resemble the Ehiravini. Most likely Jenkinsia represents a
form which existed before both Spratelloides and the Ehiravini diverged fully.
Because of its isolation, Jenkinsia may still retain early characters which have
become lost or modified in Spratelloides and the Ehiravini.

The scutes of G. aestuarius and the split anal of S. bianalis foreshadow trends
which become more fully elaborated in the clupeids. In the Ehiravini they probably
represent tentative developments which have no adaptive value but which have
remained and indicate the type of experiment prevalent amongst ancestors of the
modern Clupeidae.

368

P. J. P. WHITEHEAD

If the phylogeny of the Dussumieriidae and the Clupeidae were to be represented
diagrammatically, it would probably be correct in a general way to show the
Clupeidae branching from near the split between the Ehiravini and the Spratelloidini,
and the latter diverging initially from the Dussumieriinae (Text-fig. 33). Since all
these forms have a lower jaw so very different from the long jaw of the Engraulidae
(and in addition lack the prominent mesethmoid typical of the anchovies), it seems

DUSSUMIERIIDAE

Gilchristella

Ehirava

Spratelloides

Etrumeus

CLUPEIDAE

ENGRAULIDAE

Post-pelvic scutes.
Scutes often
keeled.

Abdominal pre-pelvic
scutes.

Pelvic scute 'spined'. Pelvic fins advanced.
Body more compressed.
Anterior supra-max. lost.

Reduction of branchiostegal rays (6-7).

Expansion of post, supra-max.

Adults small; posterior fontanelles retained.

Pelvic scute W-shaped.

Branchiostegal rays 14-20.

Posterior fontanelles closed in adults.

Prominent mesethmoid.
Underslung' lower jaw. Spined
pelvic scute. Abdominal scutes
\ in genera of Indo-Pacific Region.

FIG. 33.

A possible phylogeny for the round herrings, indicating stages in the development of certain
clupeid-like characters. For further explanation, see text.

reasonable to suggest that, not only did the Engraulidae diverge from the main stem
(primitive round herrings) before the Clupeidae did, but that in the Engraulidae the
initial trend at least was one of scute development, not degeneration (cf. Jordan &
Scale, 1926). Against the proposition that the Engraulidae evolved from a scuted
clupeid-like ancestor is the high number of branchiostegal rays in the Engraulidae
and the other osteological resemblances which Chapman (1948) found between the
anchovies and the Dussumieriinae.

Jordan & Scale (1926) thought an independent development of scutes in both
anchovies and herrings unlikely. But in Engraulis encrasicholus there is a single
pelvic scute only, of the spined type, and as the Ehiravini show, a fully scuted form
can be derived from this. Early engraulids may have had a w-shaped pelvic scute,

A REVISION OF THE RECENT ROUND HERRINGS 369

although it is possible that in the Dussumieriinae and Spratelloidini this type of
scute is a degeneration from the spined scute. This possibility, and the implication
that the Spratelloidini might have branched off from the Dussumieriinae after the
Ehiravini had also diverged from the Dussumieriinae, would help to explain
the similarities shown by the Ehiravini to the Dussumieriinae rather than to the
Spratelloidini. But it also implies the independent reduction of branchiostegal rays
in both Ehiravini and Spratelloidini and the parallel evolution of other similarities
between these two tribes (small adult size, retention of post-frontal fontanelles,
broad posterior supra-maxilla) . It seems more likely that the w-shaped pelvic
scute marks a step in the evolution of scutes, not a modification, but that it is no
longer found in the Engraulidae.

Two further aspects of scute evolution must be mentioned. A striking feature in
the present study has been the very constant presence of the pelvic scute, not only
in the Dussumieriidae, but in otherwise non-scuted engraulid genera (i.e. in Eng-
raulis), as well as in the clupeids. There remains the possibility that this scute has
evolved from the pelvic splint bones which are found in some fishes (see Gosline,
1961). Such an evolution demands the expansion of the base of the splint to form
either a w-shaped plate or an ascending arm, and at the same time the degeneration
of the splint itself (which normally extends along the outer face of the first pelvic
ray). Such a development is at least a possibility since splint bones are not present
in the Dussumieriidae, Clupeidae or Engraulidae, but occur in certain other lower
teleosts (Tarpon, Chanos, Salmo see Gosline, 1961).

The second aspect of scute evolution is the question whether scutes could have
evolved quite simply from the median series of abdominal scales without the prior
evolution of the pelvic scute. Thus in Denticeps clupeoides, the only extant member
of the family Denticipitidae (Clausen, 1959), pre- and post-pelvic abdominal scutes
are present, but there are no pelvic scutes comparable to the pelvic scutes of the
clupeids, engraulids or dussumieriids. However, the scutes in Denticeps are clearly
the median series of scales which have become sharply folded in the midline and
have developed posterior spines. In appearance they strongly resemble the spiny
scutes of the engraulids (e.g. Stolephorus) but they differ in one important respect :
they have no ascending arms, but are rounded laterally like a scale. The presence
of ascending arms in the three clupeoid families discussed here and the constancy of
the pelvic scute, seem to indicate that the evolution of the clupeoid scutes was in
some manner linked to the evolution of the pelvic scute, and that the evolution of
the scutes in Denticeps has followed a rather different course ; it may be significant
that the latter has no pelvic splint bones.

The fossil evidence throws a certain amount of light on the problem, but branchio-
stegal counts have never been certain and the descriptions are not always adequate
to decide whether a dussumieriine or a spratelloidine is involved.

The earliest record of a round herring is probably Clupavus neocomiensis (Bassani) .
Arambourg (1954) placed specimens from the Cretaceous of Morocco in the Dus-
sumieriidae because of their absence of scutes. His figures of the head (figs. 14 and
16) very much resemble Etrumeus, especially in the shape of the oper^ular series,
but vertebral numbers (39 or 40) are more those of a modem spratelloidine. Un-

ZOOL. io, 6 21

370 P. J. P. WHITEHEAD

fortunately exact branchiostegal counts were not possible, but there were probably
not more than twelve.

The importance of Clupavus lies in the links which Arambourg (1950 and 1954)
believed that it showed with the primitive actinopterygians (palaeoniscids, holosteans
and halecostom.es). These were principally in the large parietals, which meet in
the midline, and the presence of a pair of posteriorly directed canals branching
from the supra-orbital canals and converging across the posterior parts of the
f rentals (but not meeting each other). But from the position and size of the large
parietals, it is tempting to suppose that they are, at least in part, the posterior
fontanelles typically found in small round herrings. This would explain the rather
abrupt termination of the frontal sensory canal at the border between the frontals
and the parietals. In the specimen figured by Arambourg (1954, fig. 15 and pi. 3,
fig. 4) much of the median part of both parietals is missing, as also is an area between
the anterior ends of the frontals. The latter almost exactly corresponds in shape
and position to the normal anterior fontanelle of round herrings. Clupavus neoco-
miensis, as Arambourg recognized, is in many respects very clupeid-like, and such
large parietals meeting in the midline seem unlikely. In the round herrings the
parietals are fairly small and are well separated (see Text-fig. 31).

With regard to the sensory canal system, in the modern round herrings there is
no frontal branch similar to that found in Clupavus. However, Arambourg (1950
and 1954) was mistaken in believing that Clupavus is unique amongst the teleosts
in possessing such a frontal branch. While it may well be absent in the higher
teleosts, it is certainly present in the characid genus Brycon (the parietal canal in
Weitzman, 1962), and Dr. P. H. Greenwood has shown me a short but well-formed
canal in the problematical Denticeps.

Arambourg (1954) also makes the important observation that differentiation within
the clupeoid fishes was well-established by the beginning of the Cretaceous and he
mentions the coexistence amongst others of Parachanos, Clupavus and Diplomystus.
Presumably the Engraulidae had evolved by this time.

In Text-fig. 33 is shown a possible phylogeny of the round herrings with indications
of the points at which particular characters first appear. As discussed above, the
Ehiravini are difficult to place in any such scheme because they share characters
both with the Spratelloidini and with the Dussumieriinae as well as with the Clupei-
dae. Thus the diagram presents one of several possible schemes, but the one which
appears to involve the least number of contradictions. Whatever scheme is adopted,
the independent loss of one or more characters in genera which are not closely related
must be assumed.

It is also important to decide which features are primitive and which specialized.
Thus Dussumieria differs from Etrumeus in having two supra-maxillae and more
advanced pelvics, but Etrumeus only is represented in the fossil record (see Bertin,
1943, Arambourg, 1945). On the other hand, Leptolepis as well as the more primitive
isospondyls such as Elops or Albula, all have rather advanced pelvics and two supra-
maxillae (except Albula with one). Thus Dussumieria may in fact be the more
primitive of the two. Unfortunately other fossil genera referred to the Dussumierii-
dae (Lygisma, Quaesita and Sternbergia] are either inadequately described or too

A REVISION OF THE RECENT ROUND HERRINGS 371

poorly preserved to be of any value in determining which characters are more
primitive. But if the Dussumieriinae as a whole are considered more ancient than
the Spratelloidinae, then the following characters must be considered primitive :
pre-maxillary teeth, two supra-maxillae (the posterior rather narrow), a small
cleithral flap, pelvics set below or perhaps in front of the dorsal (the evidence here
is however equivocal), post-frontal fontanelles not retained in the adult (but present
possibly in Clupavus), suboperculum triangular, scutes absent or represented only
by a w-shaped pelvic scute, anal fin fairly short.

The possibility that primitive forms were scuted has been discussed on pp. 356-359.
This depends largely on whether the ventral myorhabdoi are in fact homologous
with the ventral scutes. Although such an interpretation would certainly explain
the resemblance shown by the Ehiravini to the Dussumieriinae, it would presuppose
the independent degeneration of the pelvic spined scute to a w-shaped scute in the
Spratelloidinae as well as the Dussumieriinae, and would involve several other
parallel specializations or losses, as already mentioned. It seems more likely on
present evidence that the round herrings are modern representatives of an early
non-scuted herring.

ACKNOWLEDGMENTS

It gives me great pleasure to record the assistance which I have received from
other museums and institutions, and their willingness to lend or donate specimens ;
to the individuals concerned go my sincere thanks. I also wish to thank Professor
Carl L. Hubbs for stimulating the search for the correct name for Etrumeus teres
and for allowing me to use some of his unpublished data ; Mr. N. B. Marshall for
commenting on the text ; and Dr. P. H. Greenwood for his most generous help with
certain systematic problems, much useful discussion, and many valuable criticisms
of the text.

REFERENCES

ANGEL, F., BERTIN, L. & GUIBE, J. 1946. Note relative a la nomenclature d'un amphibien et

d'un posson. Bull. Mus. Hist. nat. Paris, (2) 18 : 473-474.
ARAMBOURG, C. 1945. Note preliminaire sur quelque poissons fossiles nouveaux. Bull. Soc.

geol. France, (5) 13 : 281-288.
1950. Nouvelles observations sur les hale'costomes et 1'origine des Clupeidae. C.R.

Acad. Sci. Paris, 231 : 416-418.
1954- Les poissons cretaces du Jebel Tselfat (Maroc). Not. Mem. serv. geol. Maroc., 118 :

1 88 pp.
BARNARD, K. H. 1925. A monograph of the marine fishes of South Africa. Ann. S. Afr. Mus.,

21 : 108-110.
BEEBE, W. & TEE-VAN, J. 1933. Nomenclatural notes on the shore fishes of Bermuda.

Zoologica, 13 (7) : 133-158.
BEN-TUVIA, A. 1953. New Erythrean fishes from the Mediterranean coast of Israel. Nature,

172 : 464.
BERTIN, L. 1940. Catalogue des types de poissons du Museum National d'Histoire Naturelle.

Bull. Mus. Hist. nat. Paris, 12 (2) : p. 300.
1943- Revue critique des Dussumierides actuels et fossiles. Description d'un genre

nouveau. Bull. Inst. oceanogr. Monaco, No. 853 : 1-32.

ZOOL. 10, 6 si

372 P. J. P. WHITEHEAD

BLEGVAD, H. 1944. Fishes of the Iranian Gulf. Danish Scientific Investigations in Iran,

Pt. 3 : 1-247.
CHABANAUD, P. 1933. Sur divers poissons de la Mer Rouge et du Canal Suez. Descriptions

de deux especes nouvelles. Bull. Inst. oceanogr. Monaco, No. 627 : 4-8.
CHAPMAN, W. L. 1944. The osteology of the Pacific deep-bodied anchovy, Anchoa compressa.

J. Morph., 74 (2) : 311-329.
1948. The osteology and relationships of the round herring Etrumeus micropus Temminck

and Schlegel. Proc. Calif. Acad. nat. Sci., 26 (2) : 25-41.
CLAUSEN, H. S. 1959. Denticipitidae, a new family of primitive isospondylous teleosts from

West African freshwater. Vidensk. Medd. naturh. For en. Kbh., 121 : 141-151.
COLLETTE, B. 1962. The status of the Bermuda population of Jenkinsia. Copeia, No. 3 : 659.
DEKAY (see KAY, DE).

DELSMAN, H. C. 1925. Fish eggs and larvae from the Java Sea. Treubia, 6 : 297-306.
DERANIYAGALA, P. E. P. 1929. Ceylon sardines. Spolia Zeylan, 15 : 31-35.
FOWLER, H. W. 1941. Contributions to the biology of the Philippine archipelago and adjacent

regions. Bull. U.S. nat. Mus., No. 100 (12) : 1-879.
1958. Some new taxonomic names of fish-like vertebrates. Notul. Naturae, Philad., No.

310 : i-io.
FOWLER, H. W. & STEINITZ, H. 1956. Fishes from Cyprus, Iran, Iraq, Israel and Oman.

Bull. Res. Counc. Israel, 5 B : 260-292.
GOSLINE, W. A. 1951. The scientific name of the nehu, an engraulid fish of the Hawaiian

Islands. Pacif. Sci., 5 (3) : 272.
1961. Some osteological features of modern lower teleostean fishes. Smithson. Misc.

Coll., 142 (3) : 1-42.
GUNTER, G. & DEMORAN, W. J. 1961. Lateral lines and an undescribed sensory area on the

head of the Gulf Menhaden, Brevoortia patronus. Copeia, No. i : 39-42.

GUNTHER, A. 1868. Catalogue of the fishes in the British Museum, London. 7 B : 273-512.
HERRE, A. 1953. Check list of Philippine fishes. Res. Rep. Fish Wildl. Sew., No. 20 : 1-977.
HOLLISTER, G. 1936. The caudal anatomy of Bermuda shallow water fishes. Order Iso-

spondyli. Zoologica, 21 (4) : 257-290.
HOUTTUYN, M. 1782. Beschryving van eenige Japanse visschen en andere Zee-Schepzelen.

Verh. Holland. Maatsch. Haarlem, 20 : 340-341.
JONES, S. 1961. Further notes on Spratelloides delicatulus (Bennett) as a tuna live-bait fish,

with a record of S. japonicus (Houttuyn) from the Laccadive Sea. /. mar. Biol. Assn.

Ind., 2 (2) : 267-268.
JORDAN, D. S. 1925. Fossil fishes from Los Angeles county, California. Stanford Univ.

Publ.Biol. Sci. ,4 : 41.
JORDAN, D. S. & EVERMANN, B. W. 1896. Fishes of North America. Bull. U.S. nat. Mus.,

No. 47 : 419-421.

1905. The shore fishes of the Hawaiian Islands. Bull. U.S. Fish. Comm., 23 (i) : 1-587.

1917. The genera of fishes, from Linnaeus to Cuvier, 1758-1833, with the accepted type

of each. Stanford Univ. Publ., 1-161.
JORDAN, D. S. & GILBERT, C. H., 1883. Synopsis of the fishes of North America. Bull. U.S. nat.

Mus. 16 : 1-1018.
JORDAN, D. S. & SEALE, A. 1926. Review of the Engraulidae, with descriptions of new and

rare species. Bull. Mus. comp. Zoo/., 67 : 355-418.

KAY, DE J. E. 1842. Zoology of New York or the New York Fauna, 4 ; Fishes, pp. 262-263.
LACEPEDE. 1803. Histoire naturelle des poissons, 5, Paris : p. 383.
LIANG, Y-S. 1948. Notes on a collection of fishes from Pescadores Islands, Taiwan. Quart. J.

Taiwan Mus., 1 (2) : 120.
LISSNER, H. 1949. Sardine fishing in Israel. Sea Fish. Res. Sta. Sci. Tech. Inform., No. 2:

1-25-

MARSHALL, N. B. 1952. " The Manahine " Expedition to the Gulf of Aqaba, 1948-49.
Fishes. Bull, Brit, Mus, nat. Hist. (Zool.), 1 : 221-252,

A REVISION OF THE RECENT ROUND HERRINGS 373

MARTIN, S. F.. 1955. Une nueva familia de peces para Venezuela : Dussumieriidae. Mem.

Soc. Cienc. nat. La Salle, 15 : 184-188.
MITCHILL, S. L. 1815. The fishes of New York, described and arranged. Trans, lit. phil. Soc.,

New York, 1 : 355-492.
MORROW, J. E. 1954. Fishes from East Africa, with new records and descriptions of two new

species. Ann. Mag. nat. Hist., (12) 7 : 797-820.
MYERS, G. S. 1960. The endemic fish fauna of Lake Lanao, and the evolution of higher

taxonomic categories. Evolution, 14 (3) : 323-333.
PARR, A. E. 1930. Teleostean shore and shallow-water fishes from the Bahamas and Tarks

Island. Bull. Bingham. oceanogr. Coll., 3 (4) : 1-148.

PHILLIPS, J. B. 1951. Round Herring off Central California. Calif. Fish. Game, 37 : 512.
RIDEWOOD, W. G. 1904. On the cranial osteology of the clupeoid fishes. Proc. zool. Soc.

Lond., 2 : 448-493.

RIVAS, L. R. 1946. A new dussumieriid fish of the genus Jenkinsia from Bermuda. Smith-
son. Misc. Coll., 106 (14) : 1-4.
SAUVAGE, H. 1891. Poissons de Madagascar, in Grandidier, Histoire physique, naturelle et

politique de Madagascar, Paris, p. 496.
1883. Description de quelques poissons de la collection du Mus6um d'Histoire Naturelle.

Bull. Soc. Philom., Paris, (7) 7 : 160.
SCHULTZ, L. P. & WELLANDER, A. D. 1953. Fishes of the Marshall and Marianas Islands.

Bull. U.S. nat. Mus., No. 202 : 1-685.
SMITH, J. L. B. 1949. Forty-two fishes new to South Africa, with notes on others. Ann. Mag.

nat. Hist., (12) 2 : 97-111.
SVETOVIDOV, A. N. 1952. (Clupeidae) Tabl. anal. Faune U.R.S.S. N.S. No. 48 2 (No. i) :

I-33I-
WEBER, M. & DE BEAUFORT, L. F. 1913. The fishes of the Indo- Australian Archipelago, 2 :

xx + 404 pp. Leiden.
WEITZMAN, S. H. 1962. The osteology of Brycon meeki, a generalized characid fish, with an

osteological definition of the family. Stanford Ichthyol. Bull., 8 (i) : 1-77.
WHITEHEAD, P. J. P. 1962^. Abdominal scutes in the round herrings (Dussumieriidae).

Nature, 195 : 511-512.
19626. A review of the Indo-Pacific gizzard shad genera Nematalosa, Clupanodon and

Konosirus (Pisces : Dorosomatidae) . Bull. Brit. Mus. nat. Hist. (Zool.), 9 (2) : 87-102.
(In press.) The true identity of Clupea sadina Mitchill, 1814. Copeia.

STUDY MATERIAL

The first name is that which is used in this paper ; the second (in parenthesis),
that under which the specimens have been labelled hitherto.

Dussumieria acuta

(D. acuta)

6 fishes, 121-148 mm., Shihr & Burum, Gulf of Aden (I962-3.26.2II-2I6 1 ).

(D. acuta)

5 fishes, 105-114 mm., Mukalla, Gulf of Aden (1962.3.26.217-221).
14 fishes, 105-120 mm., Singapore (1935.4.12.3-12).
i fish, 114 mm., Andamans (1889.2.1.2047).
i fish, 122 mm., Bombay (1889.2.1.2037).

1 British Museum registration numbers unless stated otherwise.

374 P- J- P. WHITEHEAD

(D. acw/a)-contd.

4 fishes, 93-141 mm., Madras (1889.2.1.2042-5).
i fish, 118 mm., Borneo (1868.1.28.81).

i fish, 98 mm., Port Blair (1889.2.1.2046).

1 fish, 94 mm., locality unknown (1860.3.19.956).

2 fishes, 106-107 mm., Djambi, Sumatra (1934.8.18.12.13).

3 fishes, 69-84 mm., Malabar (1889.2.1.2039-41).

5 fishes, 82-119 mm., Batavia (1934.8.18.17-19).

5 fishes, 92-100 mm., Panci Sumatra (1934.8.18.14-16).

5 fishes, 89-100 mm., Batavia (1934.8.18.20-22).

5 fishes, 64-114 mm., Haifa, Israel (1962.6.13.4-8).

i fish, 48 mm., Malabar (1889.2.1.2049).

20 fishes, 77-107 mm., Haifa, Israel (Sea Fisheries Research Station (Israel) collection).

(D. elopsoides)

HOLOTYPE, 124 mm., Madura (1867.11.28.17).
i fish, 116 mm., Zool. Soc. Coll. (1852.9.13).

1 fish, 88 mm., Amboina (1858.4.21.360).

2 fishes, 114-115 mm., Siam (1920.7.13.4-5).
i fish, 101 mm., Amoy, China (1860.7.20.88).

12 fishes, 107-115 mm., Calicut (1935.9.20.1-8).
i fish, 118 mm., Foochow (1936.10.7.21).
i fish, 113 mm., Hong Kong (1939.3.23.5).

(D. hasselti)

HOLOTYPE, 134 mm., Java 1867.11.28.21).

3 fishes, 85-88 mm., Coromandel (types of D. acuta, Paris Museum, No. 3694).
9 fishes, 68-120 mm., Malabar (types of D. acuta, Paris Museum, No. 3217).

4 fishes, 112-131 mm., Coromandel (types of D. acuta, Paris Museum, No. 3697).

3 fishes, 89-110 mm., Persian Gulf (coll. Blegvad. Univ. Zool. Mus., Denmark, No.
CN 5-7).

Etrumeus teres

(E. teres]

9 fishes, 115-130 mm., Woods Hole, U.S.A. (1934.11.30.1-9).

i fish, 120 mm. approx. (skeleton), Woods Hole, U.S.A. (1934.11.30.10).

(E. micropus)

87 fishes, 67-78 mm., Nagasaki (1923.2.26.73-82).
i fish, 136 mm., Kobe, Japan (1907.12.23.91).
i fish, 85 mm., Hong Kong (1939.3.23.6).
i fish, 126 mm., Japan (1867.11.28.265).

1 fish, 165 mm., Haifa (Sea Fish. Res. Sta. (Israel) collection).
8 fishes, 117-149 mm., Natal (1903.2.6.11-17).

2 fishes, 133-134 mm., Eilat, Red Sea (Sea Fish. Res. Sta. (Israel) collection).

(E. jacksoniensis)

2 fishes, 105-122 mm., Botany Bay (1897.10.27.49-50).

(E. acuminatus)

i fish (HOLOTYPE of Perkinsia othonops), 265 mm., California (1891.5.19.210).

7 fishes, 109-125 mm., California (1962.6.13.9-15).

i fish, 120 mm. approx. (skeleton), California (1962.6.13.16).

A REVISION OF THE RECENT ROUND HERRINGS 375

Ehirava tnalabaricus

(E. fluviatilis)

i fish (HOLOTYPE of E. fluviatilis}, 48 mm., Ceylon (1929.7. i . i).
13 fishes, 30-46 mm., Ceylon (1929.7.1.2-9).

(Spratelloides wialabaricus)

4 fishes, 46-56 mm., Canara (1889.2.1.2052-5).

i fish, 40 mm., Malabar (1889.2.1.2048).

i fish, 53 mm. (skeleton), Malabar (1889.2.1.2050).

Ehirava madagascariensis

i fish (TYPE of Spratelloides madagascariensis), 40 mm., Madagascar. (On loan from

Paris Museum, No. 3794).

10 fishes, 41.0-52-5 mm., Buffalo river, Cape Province (1878.1.22.33-43).
i fish, 52 mm., Buffalo river, Cape Province (1878.1.22.25).

Gilchristella aestuarius

(Spratelloides aestuarius)

i fish, 51 mm., Durban (1915.7.6.3).

i fish, 53 mm., Durban (1919.9.12.3).

i fish, 53 mm., Milnerton lagoon, S. Africa. 1

i fish, 45 mm., Knyssa, S. Africa. 1

i fish, 40 mm., St. Lucia, S. Africa. 1

1 fish, 35 mm., Swartkops river, S. Africa. 1

2 fishes, 42-44 mm., East London, S. Africa (1962.6. 13.2 and 3). 2
i fish, 39 mm., Kei mouth, S. Africa (1962.6. 13. i). 2

i fish, 58 mm., Buffalo river, Cape Province (1878.1.22.24).

(Spratelloides sp.)

4 fishes, 37-41 mm., Swartkops river, S. Africa (1905.1.7.1-4).

(Clupea sagax)

3 fishes, 28-45 mm., Swartkops river, S. Africa (1898.12.17.9-11).

Sauvagella bianalis

(Spratellomorpha bianalis)

5 fishes, 44-45 mm., Madagascar (TYPES of Sauvagella bianalis Bertin, Paris Museum,

No. A 5174).

Spratelloides gracilis gracilis

(Spratelloides gracilis)

1 fish (TYPE of Clupea argyrotaenia Bleeker), 66 mm., E. Ind. Arch. (1867. n .28. 17).

3 fishes, 68-72 mm., Japan (4.46.8134).

4 fishes, 66-79 mm., Goto I., Japan (1907.12.23.96-9).
Many fishes, 52-83 mm., Nagasaki (1923.2.26.31-40).

Many fishes, 88-93 nim., Wakanoura, Japan (1923.2.26.41-50).
3 fishes, 56-66 mm., Formosa (1934.8.18.1-3).

2 fishes, 45-47 mm., Ghardaqa, Red Sea (1935.9.30.10-11).

Many fishes, post -larvae to 37 mm., Senafir, Gulf Aqaba (1951 . i . 16.36-60).

3 fishes, 47-48 mm., Sumatra (1934.8.18.9-11).

1 On loan from Prof. J. L. B. Smith.
* Donated by Prof. J. L. B. Smith.

376 P. J. P. WHITEHEAD

Spratelloides delicatulus delicatulus

(Spratelloides delicatulus)

5 fishes, 53-70 mm., Hasler collection (1855.9.19.1153-58).
i fish, 51 mm., Bonham I. (1874.11.19.65).
i fish, 65 mm., purchd. of Dr. Bleeker (1869.11.28.34).
i fish, 77 mm., Australia (1851.2.20.14).

1 fish, 28 mm., Kosi Bay, Zululand (1906.11.19.27).
29 fishes, 37-44 mm., Seychelles (1927.4.14.13-22).

5 fishes, 50-53 mm., Thousand Is. (1934.8.18.4-8).

15 fishes, 28-35 nim., Sulu Prov., Philippines (1933.3.11.25-34).
10 fishes, 21-49 mm., Firaun I., Gulf of Aqaba (1951 .1.16. 14-23).
12 fishes, 42-47 mm., Senafir, Gulf of Aqaba (1951.1.16.24-35).
22 fishes, 27-35 mm., Cocos-Keeling (1949.1.29.5-26).

1 6 fishes, 16-45 mm., Marsa Halaib, Red Sea (1960.3.15.16-31).

8 fishes, 36-39 mm., Kad Eidwid reefs, Red Sea (1960.3.15.33-39).

2 fishes, 49-52 mm., G'an, Maldives (1962.1.22.2-3).
16 fishes, 39-42 mm., Gulf of Aden (1962.6.19.1-16).

i fish, 40 mm., alizarin preparation, Marsa Halaib, Red Sea (1960.3.15.32).

Spratelloides delicatulus robustus

(Spratelloides robustus)

5 fishes, 33-68 mm., Port Jackson (1897.10.27.43-47).
81 fishes, 46-74 mm., coast of N.S.W. (1924.4 .30.1-10).

Jenkinsia lamprotaenia lamprotaenia

(Dussumieria stolifera)

SYNTYPES, 15 fishes, 30-36 mm., Key West, Florida (1884.7.7.47-56).

(Jenkinsia lamprotaenia)

SYNTYPES/ 2 fishes, 53-56 mm., Jamaica (1962.7.19.3-4).

(Dussumieria lamprotaenia)

5 fishes, 19-22 mm., Cayman Island (1939.5.12.6-9).

(Jenkinsia lamprotaenia)

52 fishes, 48-53 mm., Cat Cay, Bahamas (1962.7.21.1-50; ex 2 U.M.M.L. 2317).
15 fishes, 25-36 mm., Cat Cay, Bahamas (1962.7.21.101-115 ; ex z U.M.M.L. 5066).
25 fishes, 25-30 mm., St. John, Virgin Islands (1962.7.21 . 121-145 ; ex* U.M.M.L. 5280).

5 fishes, 38-44 mm., St. John, Virgin Islands (1962.7.21 .116-120 ; ex 2 U.M.M.L. 5278).
50 fishes, 33-38 mm., Monroe County, Florida (1962.7.21 .51-100 ; ex* U.M.M.L. 6742).
19 fishes, 39-59 mm., Caribbean (20 12' N ; 91 59' W) (1962.8.1.1-19 ; ex 3 C.N.H.M.

66009) .

6 fishes, 33-39 mm., Tom Owen Cay, British Honduras (1962 .8.1. 20-25 ' exZ C.N.H.M.

50028).

22 fishes, 33-45 mm., Belize, British Honduras (1962.8.1.26-47 ; ex 3 C.N.H.M. 9831).
25 fishes, 24-32 mm., Golfo Batabano, Cuba (1962.8.1.48-72 ; ex 3 C.N.H.M. 61363).

8 fishes, 42-46 mm., off Dry Tortugas (1962.8.1.73-80 ; ex 3 C.N.H.M. 61365).
21 fishes, 23-40 mm., Serrana Rock, Caribbean (1962.8. i .81-101 ; ex 3 C.N.H.M. 66012).
6 fishes, 23-51 mm., Alacran Rock, Gulf of Campeche (1962 .8.1. 102-107 ; ex 3 C.N.H.M.

66010).

19 fishes, 27-45 mm., Alacran Rock, Gulf of Campeche (1962 .8.1. 108-126 ; ex 3 C.N.H.M.
61366).

1 Stated by Giinther (1868) to be two typical specimens from Gosse's collection.

2 Donated by Prof. C. Robins from the collections of the University of Miami Marine Laboratory.

3 Donated by Dr. Loren Woods from the collections of the Chicago Natural History Museum.

A REVISION OF THE RECENT ROUND HERRINGS 377

Jenkinsia tnajua majua

1 fish, HOLOTYPE, 53-8 mm., Areas Cay, Campeche Bank (U.S. Nat. Mus., No. 197412).
52 fishes, 25-53 mm., PARATYPES, Areas Cay, Campeche Bank (Chicago Nat. Hist. Mus.

No. 61362).

20 fishes, 45-55 mm., PARATYPES, Areas Cay, Campeche Bank (1962.8.1.152-171 ; ex 1
C.N.H.M., No. 66007).

2 fishes, 54 mm., PARATYPES (alizarin) Areas Cay, Campeche Bank (Chicago Nat. Hist.

Mus. No. 66008).

10 fishes, 4654 mm., PARATYPES, Campeche Banks (1962.7.21.146-155).
25 fishes, 40-44 mm., PARATYPES, Areas Cay, Campeche Bank (1962.8.1.127-151 ; ex*
C.N.H.M. 46265).

Jenkinsia majua woodsi

I fish, 48 mm., HOLOTYPE, Nr. Quita Sueno Bank, Caribbean (U.S. Nat. Mus. No. 197413 ;

ex* C.N.H.M., No. 66011).
16 fishes, 40-46 mm., PARATYPES, Nr. Quita Sueno Bank, Caribbean (1962 .8.1. 173-188 ;

ex* C.N.H.M. 66011).

10 fishes, 36-46 mm., PARATYPES, Nr. Quita Sueno Bank, Caribbean (C.N.H.M. 66011).
25 fishes, 31-43 mm., PARATYPES, Nr. Swan I., Caribbean (U.S. Nat. Mus. No. 197413 ;

ex* C.N.H.M., No. 66013).
6 fishes, 29-48 mm., PARATYPES, Swan I., Caribbean (1962.7.21.181-186; ex*

C.N.H.M. no number).
25 fishes, 28-39 mm., PARATYPES, Gun Cay, Bahamas (1962.7.21.156-180; ex*

U.M.M.L., No. 2102).

4 Donated by Dr. Loren Woods from the collections of the Chicago Natural History Museum.

P. J. P. WHITEHEAD

TABLE I. A Comparison of Proportional and other Measurements of the Holotypes of
Dussumieria hasselti and D. elopsoides

D. hasselti D. elopsoides

Standard length .... 134-0 mm. . 124-0 mm.

In % of S.L.

Body depth . . . . 20-9 . 20-8

Head length . . . . 26-4 . 26-9

Snout length . . . . 9-0 . 8-8

Eye diameter . . . . 6-7 . 7-1

Post-orbital . . . . 10-1 . 9-2

Inter-orbital . . . . 5-8 . 5-6
(above eye centre)

Snout to maxillary tip . . 8-9 . 8-7

Lower jaw length ... 12-3 . 12-4

Operculum height . . . 11-2 . 1 1 3

Peduncle depth .... 8-2 . 8-8

Pre-dorsal distance . . . 59 -o . 55 -o

Pre-pelvic distance . . . 65-0 . 64-5

Longest gillraker on ist arch . 3-4 . 3-5

Gillraker count .... 22 . 23

Dorsal rays : simple ... iv . iv

branched ... 16 . 17

Anal rays : simple ... iii . iii

branched ... 13 . 13

Branchiostegal rays ... 17 . 15

A REVISION OF THE RECENT ROUND HERRINGS 379

* O

.d "*

ON O

fl) ^ Q Q\

S^ ^5"

'tj o3 * ^^ 04

o o M" o HI c^i

<* N ON iO O ^

JH 'C O ON 00

^* O t^ 00 t^ ro ^O ^

^^ ^ J, 4. J ^^ 4 J,

-* HH -^- O OO ro M M oo
5 N 6 N ' ' ' '

"53 t^ i>-Tt~-r^oooo'

O 00 iO vO rn . . c^

- S M M f< 1

H Tj" *O 00 ^ '^H *^X(

It M

"So

"o3 ~~^

:3 o

-2

*^S

<j- r^-

rt N O O

^ w M N r^ moo

^f 1C iO N 00 O

1
V)

^j uj ^^ ^f- Q oO ^"
CO ...

3 10 00 >o Ir^ oo

v O iO *^-<

TfO 00 <* ^ ^^

<j o M N

> --H . 0-(

H 1

Q g

s -S

"o"

-. 55/5
sj ^y

^ Q

^'l

tf -NO or

iO N vO t*N
O (N M t^ O VJ

* >O O

3 g

C<J3 M <SiNOOOOi-iONt.

io - o ^pxs;

HOOrt-vOOO rOMM

"T 111 ill

""^5 d

<^N

CO ^

to J*

< t/2

I-H HI M , h

*Z M

q" C^ t^ VO ^O CO

i <* o co > ' :fl 2
1

^_- rt
tt) S3

"f?

J? 2"

to -f-*

1 w

^ **->

> "3
1

_. W) 10 P^- <N fi

s g oo oo vo t~x

HO if) vO ON v
tC NOO HIMOOONONON*-
tf M 1 1 1 1 I 1 \ \
J^tuO >O NN>OONrnOC

5 ^00 8
. f n3

> 00 00 00 O 00 O -^

1 f jf 9 1 7 7 ?

^ CO O O C^ N *O ^ t> bjO

to g

9*O r^ ^^-r^oooooN 1 '

hoon-NN K^.^.^ rt

SH S

42,W H, ,N __ K

ON iO

*7 :a -g

^ ^

o t-^

*"^ 2

^ CT)

- -

S ^

1"*

CJ <N
rt . 00 10 H

oo O <* oo oo <*
'u '-5 ^ o r^ . u

10 O O

t^ g

l-iOrfTj- vOt^>O o

flol ^^ T7? t T'??7

C42 O NNNOONONOC

sr^i 25 11 i

jj >O iOvOO-l>.rOoOT

55 S

1-00 ** N N
"*-O 00 .> JJH %

t/3

o ^

1 .

o o -3,

< .ri -rt * ^

-M tf c/5

^^5 ,-c r- a3' 4J b

5^ .2 .22 -2 *

S gjftt!|f!

OJD 12 to

S *~3 O ^ b>,
O) rt OT rt3

O '~ H CO-^^o3 r ^'n

' (H ^ 03 E ST 1 S nj

i i^tiitlii

oO!DCj i_i QJ | gf

^ ' ' 1 i^ ~J (-1 L.

"TjCj^cu cto3o

S CIj bid r/^ ' ' * ^ CL

CU QJ QJ QJ *""^ *H "o o3

-S ri

r73 O a) G

55 !/3 HH

P ft <

3 8o

P. J. P. WHITEHEAD

TABLE III. Proportional Measurements, Expressed as Percentages of Standard Length
for Specimens of Spratelloid.es delicatulus /row Australia (S. d. robustus) compared
with Specimens from the Indo-Malayan Archipelago and the Seychelles (S. d.
delicatulus)

Number of specimens
Standard lengths
In % of S.L.

Body depth .

Head length .

Snout .

Eye diameter

Post-orbital distance

Maxilla length

Pectoral length

Pelvic length

Pre-dorsal distance

Pre-pelvic distance

Pre-anal distance .

Dorsal rays
Pectoral rays .
Anal rays

Single measurements outside normal range placed in parentheses.

Australia

Other areas

S. d. robustus

S. d. delicatulus

50-3-71-0

39-0-52-0

18-2-21-0

(16-7)* 17-9-20-5

25-2-28-2

24-0-26-7

7-2-7-7

6-6-7-4

6-5-7-7

9-3-10-3

8-8-io-r

9 2-9 6

8-4-9-5

14-2-15-8

(11-9) 14-0-15-0

II-I-I2-8

9-2-11-5

46-3-48-0 (49-1)* -

46-5-47-1 (49

(50-5) 53-0-57-5

53-2-56-0

76-2-83-5

(75-o) 79-2-82-0

ii 10-11

ii 9-1 1

i 11-12

i 10-12

ii-iii 7-9

/T7. If

THE POLYCHAETE FAUNA OF

SOUTH AFRICA

PART 8: NEW SPECIES AND

RECORDS FROM GRAB SAMPLES

AND DREDGINGS

J. H. DAY

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 7

LONDON: 1963

THE POLYCHAETE FAUNA OF SOUTH AFRICA

PART 8: NEW SPECIES AND RECORDS FROM

GRAB SAMPLES AND DREDGINGS

j. H. DAY

Professor of Zoology, University of Cape Town

Pp. 381-445 ; 12 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 7

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949,
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.

This paper is Vol. 10, No. 7 of the Zoological series.

Trustees of the British Museum, 1963

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued July, 1963 Price Twenty One Shillings

THE POLYCHAETE FAUNA OF SOUTH AFRICA

PART 8: NEW SPECIES AND RECORDS FROM

GRAB SAMPLES AND DREDGINGS

By J. H. DAY

CONTENTS

Page
INTRODUCTION ........... 383

STATION LISTS ........... 384

SYSTEMATIC ACCOUNT . . . / . . . . . . 386

REFERENCES ........... 443

SYNOPSIS

Recent marine collections by the University of Cape Town have revealed 27 new species of
Polychaeta and 32 new records for South Africa. These are described. A new family is erected for
the genus Cossura and three new genera are defined. The family Paraonidae is revised and a key to
17 species is drawn up.

INTRODUCTION

THE marine fauna of South Africa is at present being surveyed along a number of
transect lines extending from the coast out to the edge of the continental shelf.
Seven lines have been worked between Lamberts Bay on the Atlantic Coast and
East London on the Indian Ocean side. An analysis of earlier polychaete samples
has been presented in two reports (Day, 1960 and 1961). New species and new
records for South Africa obtained from later samples are described below.

An analysis of grab samples has shown that the Polychaeta are the dominant
group of benthonic animals on soft bottoms around the Cape both as regards numbers
of species and numbers of individuals. Surprisingly enough there is no indication
that the number of species decrease down to 200 metres ; rather the reverse since
there are indications that polychaetes become more numerous as the substratum
changes from sand to soft mud with increasing depth. Even at a depth of 1,250
metres a o-i metre square van Veen grab brought up 17 species. Of course most of
these species are small, many of them less than 20 mm. in length and some as small as
4 mm. These escape when a dredge is drawn up from deep water and can only be
obtained by grab samples which are carefully washed through a fine sieve of I mm.
mesh. The Polychaeta obtained by earlier dredgings around the Cape contained
171 errant species and 166 sedentary ones. The new grab samples contain repre-
sentatives of many of these and in addition 32 new records for the area and 27 new
species.

No attempt will be made here to list new locality records of well known species
since these and earlier records will shortly be summarised in a monograph. Only

ZOOL. 10, 7 22

3 8 4 J.H.DAY

the new species and those which show major extensions of range are described.
Among them are two new genera of the family Syllidae and one new genus of the
family Nereidae. The discovery of five species of the family Paraonidae, in addition
to the seven already reported, has provided an opportunity for revising the genera
of this family and a key covering 17 species is given. In addition a new family is
erected to contain the genus Cossura which differs widely from other genera of the
family Cirratulidae.

The material upon which the present report is based was obtained with the help
of funds from the South African Council of Scientific and Industrial Research to
which I am very much indebted. Further, the Ernest Oppenheimer Memorial
Trust made a generous grant which allowed me to work at the British Museum
(Natural History) for the best part of a year. In this way I was able to compare
South African material with the types or at least with identified specimens from
the type locality. As a result certain types have been redescribed, several erro-
neous records have been removed from the South African list and many doubtful
records have been confirmed.

I wish to thank my many friends at the British Museum (Natural History) for
their help and courtesy during my visit, particularly Dr. F. Fraser, the Keeper of
Zoology and Mr. R. Sims of the Annelid Section.

The types of all new species will be deposited in the British Museum (Natural
History).

STATION LISTS

Collection data are given in the station lists below. In the systematic section the
collection data for each species are given in a summarised form. Thus the records for
the first species Harmothoe antilopis are shown as SCD 207 (i), 237 (i). This means
that one specimen was found at station SCD 207 and one at station SCD 237.
Reference to the station lists shows that SCD stands for south coast dredging and
that station 207 was made on 30.11.60 at 34 51' 8/23 41' E in 183 metres on a
bottom of khaki coloured sand. Full details of other records can be obtained in
the same way.

DREDGE AND VAN VEEN GRAB STATIONS OFF THE WEST COAST OF THE CAPE PROVINCE

SYMBOL WCD

Depth
Station No. Date Position (metres) Bottom

WCD 41 . 2.5.60 . 33 06-5' S./I7 56-7' E. . 18-33 . R.

WCD 53 . 15.7.60 . 34 ii'S./i8i3'E. . 148 . kh. S. M.

WCD 60 . 21.9.60 . 32 04-7'S./i8 i2'E. . 96 . gn. M.

WCD 61 . 21.9.60 . 3205'S./i8o6'E. . 108 . gn. M.

WCD 63 . 21.9.60 . 3205'S./i8o6'E. . 108 . gn. M.

WCD 64 . 21.9.60 . 3205'S./i756 / E. . 128 . gn. M.

WCD 65 . 21.9.60 . 32 05'S./i7 56'E. . 128 . gn. M.

WCD 66 . 21.9.60 . 3205 / S./i8i4'E. 83 . ?

WCD 67 . 22.9.60 . 3244 / S./i8oi / E. . n . bl. M.

WCD 73 . 8.10.60 . 3425'S./i7 36'E. . 1240 . kh. M.

WCD 79 . 8.10.60 . 34i7'S./i7 / 53 / E. . 320 . gn. S. M.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

385

Depth

Station No.

Date

Position

(metres)

Bottom

WCD 82

21.9.60

3 205'S./I756'E.

128

gn. M.

WCD 83

8.10.60

34 i7'S./i 7 53'E.

320

gn. S. M.

WCD 85

3.7.61

33 06-4' S./I7 44-9' E.

I 4 6

gn. bl. M.

WCD 86

3.7.61

33 06-4' S./i 7 44-9' E.

I 4 6

gn. bl. M.

WCD 87

3.7.61

33o6-2 / S./i749 / E.

d. gn. M. R.

WCD 90

2.7.61

32o5'S./i8i6-7'E.

co. wh. S. & Sh.

WCD 91

2.7.61

32 05-5' S./i8 17-3' E.

co. wh. S. & Sh.

WCD 92

2.7.61

3205'S./i8 16-7' E.

co. wh. S. & Sh.

WCD 94

2.7.61

32 05-3' S./i8 15-8' E.

WCD 97

3.7.61

33o 4 -2'S./i7 49'E.

d. gn. M. & R.

WCD 103

2.7.61

32 05-3' S./i8 15-8' E.,

kh. M.

WCD 106

3.7.61

33 06-5' S./i 7 32-9' E.

183

d. gn. M.

WCD 108

2.7.61

32 o8'S./i 7 39 / E.

172

f.d. gn. M.

WCD 109

2.7.61

3 2 o8'S./i7 39 / E.

172

f. d. gn. M.

WCD no

2.7.61

3 2o8 / S./i7 39 / E.

172

f. d. gn. M.

WCD 112

3.7-6i

33 06-4' S./i 7 47-2' E.

I 4 i

d. gn. M.

WCD 115

3.7.61

33 06-5' S./i 7 32-9' E.

183

d. gn. M.

WCD 116

3.7.61

33 06-5' S./i 7 32-9' E.

183

d. gn. M.

WCD 118

3.7.61

33 06-4' S./i 7 44-9' E.

I 4 6

gn. bl. M.

WCD 121

2.7.61

3 2 o8'S./i 7 39'E.

172

f. d. gn. M.

WCD 122

2.7.61

32o8'S./i 7 39'E.

172

f. d. gn. M.

DREDGE AND VAN

VEEN GRAB STATIONS IN FALSE BAY SYMBOL

FAL

Depth

Station No.

Date

Postion

(metres)

Bottom

FAL 284

23.9.54

Roman Rock (diving)

12-14

S. Sh.

FAL 342

31.1.59

3 4 ii / S./i8 35-5'E.

f. br. S.

FAL 375

25.2.59

34 i6-8'S./i842-8'E.

gn. S. Sh.

FAL 390

15.11.60 .

34 12-6' S./i8 29-1' E.

kh. M.

FAL 397

15.5.61

3 4 i2-5'S./i837'E.

kh. S. & Sh.

FAL 406

16.5.61

34 08-8' S./i8 33-5' E.

f. kh. S.

FAL 413

15.5.61

34 i2-5'S./i8 37'E.

kh. S. & Sh.

FAL 419

15.5.61

34i2-5'S./i8 37'E.

kh. S. & Sh.

FAL 422

15.5-61

34i2. 5 'S./i837'E.

kh. S. & Sh.

FAL 424

15.5.61

34i2-5'S./i837'E.

kh. S. & Sh.

DREDGE AND VAN VEEN GRAB SAMPLES OFF THE SOUTH

COAST OF THE CAPE PROVINCE

SYMBOL SCD

Depth

Station No,

Date

Position

(metres)

Bottom

SCD 130

3 . 6 . 60

3 4 4 8'S./22 o6'E.

IOO

co. kh. S.

SCD 136

28 . 8 . 60

3435'S./2i56'E.

co. & f. Sh.

SCD 154

25 . i i . 60

3 4 03'S./25 5 9'E.

SCD 156

25.11.60

34 03'S./25 59'E.

SCD 170

24 . i i . 60 .

3358-9'S./2 5 4 i. 4 'E.

4-ii

SCD 182

30 . i i . 60

342o'S./233i / E.

R. kh. S.

SCD 185

25.11.60

34i3 / S./26o 4 / E.

124

gr. gn. S.

SCD 187

30 . i i . 60 .

34io'S./23 32'E.

gn. M.

SCD 193

29 . i i . 60

34 04-3' S./23 25-8' E.

gn. M.

SCD 197

29.11.60

34 07-5' S./23 31-7' E.

IS.

SCD 200

30 . i i . 60 .

3 4 io'S./23 32'E.

gn. M.

SCD 207

30 . i i . 60

3 4 5i / S./23 4 i'E.

183

kh. S.

SCD 212

24 . i i . 60

33 5 8-8'S./25 4 2-2'E.

co. S. br. Sh.

386 J. H. DAY

SCD 214 . 25.11.60 . 34 03 / S./2558'E. . 78 . co. gr. S. br. Sh.

SCD 215 . 25.11.60 . 3403'S./25 58'E. . 78 . co. gr. S. br. Sh.

SCD 218 . 25.11.60 . 34o3 / S./2558 / E. . 78 . co. gr. S. br! Sh.

SCD 220 . 29.11.60 . 34 02'S./23 28-4'E. . 49 . S.M.R.

SCD 223 . 25.11.60 . 34 i3'S./2604'E. . 124 . gr. gn. S.

SCD 227 . 29.11.60 . 34 07-5' S./23 31-7' E. . 79 . f. S.

SCD 228 . 5.12.60 . 35 43-5'S./203i'E. . 143 . gn. M.

SCD 232 . 4.12.60 . 3628-5 / S./2i n'E. . 183 . kh. S.

SCD 233 . 4.12.60 . 3628-5'S./2i n'E. . 183 . kh. S.

SCD 236 . 30.11.60 . 345i'S./234i / E. . 183 . kh. S.

SCD 237 . 30.11.60 . 345i'S./23 4 i / E. . 183 . kh. S.

SCD 245 . 29.11.60 . 34o2'S./2328-4'E. . 49 . S.M.R.

SCD 246 . 29.11.60 . 3402'S./23 28-4'E. . 49 . S.M.R.

SCD 247 . 29.11.60 . 34 04-3' S./23 25-8' E. . 45 . gn. M.

SCD 251 . 30.11.60 . 3448'S./23 39'E. . 148 . R.

SCD 275 . 30.11.60 . 345i'S./234i / E. . 183 . kh. S.

SCD 284 . 6.2.62 . 33oi'S./2755 / E. . 7 . f. wh. S.

SCD 288 . 6.2.62 . 3304'S./27 57'E. . 84 . co. S. Sh.

SCD 298 . 6.2.62 . 3309 / S./2802'E. 84 . ?

SYSTEMATIC ACCOUNT

Family APHRODITIDAE

Sub-family POLYNOINAE

Harmothoe antilopis Mclntosh, 1876

(Text-fig. la-f)

Harmothoe antilopis Mclntosh, 1876: 383, pi. 69, figs. 4-6 ; Mclntosh, 1900 ; 334, pi. 28, fig. 13 ;
pi. 32, fig. 16; pi. 38, figs. 24-25 ; Fauvel, 1923 : 56, fig. 19 a-/.

RECORDS. SCD 207 (i), 237 (i).

NOTES. These South African specimens have been checked as identical with the
types in the British Museum (Natural History) (No. 1921-5-1-437-438) obtained
in 358-567 fathoms off Scotland by the " Porcupine ".

The resemblance is in fact even closer than the published descriptions of the types
would suggest. The type material consists of two broken specimens each about
20 mm. long. The prostomium (Text-fig. la) is roughly square, not tapered ante-
riorly, the anterior pair of eyes is large and their position is characteristic for they
are directly below the small but well marked prostomial peaks. The median an-
tenna is missing both from the types and the South African specimens but the laterals
are present ; they are small, tapered and, like the dorsal cirri, have numerous long
papillae. The elytra (Text-fig, ib) are oval with fringed margins and numerous
chitinous tubercles (Text-fig, ic) which are best described as tall crowns for they are
cylindrical with 4-5 divergent projections at the truncate end. In the South
African specimens (but not in the types) some of the tubercles near the posterior
margin are twice the size of the rest and instead of blunt projections each has three
saucer-like depressions the edges of which form prominent flanges (see Text-fig, i c').

THE POLYCHAETE FAUNA OF SOUTH AFRICA

387

FIG. i. Harmothoe antilopis : (a) head ; (b) elytron ; (c, c'} details of elytrous tubercles
near posterior margin ; (d) posterior view of parapodiurn ; (e, e'} notoseta and details
of tip ; (/, /') neuroseta and details of tip. Harmothoe lagiscoides serrata : (g) head ; (h)
elytron ; (_;', j') notoseta and details of tip ; (k, k f ) neuroseta and details of tip ; (/)
posterior view of parapodiurn ; (m) details of elytrous tubercles near posterior margin.

388 J. H. DAY

The dorsal cirri (Text-fig, i d] are tapered and just shorter than the neurosetae.
The notopodium is well developed and the neuropodium has a triangular presetal
lip. The notosetae (Text-fig. I e) are much more closely serrated than the figures of
Mclntosh and Fauvel would suggest ; the tips (Text-fig. i e'} are blunt and in the
South African forms they are flanged. The neurosetae (Text-fig. I /) have about
15 rows of spinules and long naked and bidentate tips (Text-fig. I /') with the secon-
dary tooth half the length of the terminal one.

The two most characteristic features are the position of the eyes and the tall
crowns on the elytra. Both Harmothoe aequiseta and Antinoe epitoca have the first
pair of eyes antero-ventral in position but there are obvious differences in the elytra
and setae.

H. antilopis seems to be a deep water species. It has long been known from the
Atlantic coasts of Europe but when more samples have been obtained at the edge of
the continental shelf its distribution may prove to be much wider than is known at
present.

Harmothoe lagiscoides serrata subsp. n. (Text-fig, i g-m)

RECORDS. WCD 112 (i).

DIAGNOSIS. Differs from H. lagiscoides (Willey, 1902) in having strongly serrated
notosetae, neurosetae with long knife-bladed tips and only a single row of giant
curved spines on the posterior margin of the elytra.

DESCRIPTION. The type locality is 33 06-4' 8/17 47-2' E in 141 metres. The
single specimen is an ovigerous female broken into four pieces, but if complete it
would measure about 20 mm. The body is well tapered and creamy brown in
alcohol.

The prostomium (Text-fig i g) has well marked frontal peaks. The eyes are
well developed and the anterior pair are about half-way back on the sides of the
prostomium. The median antenna is missing but the laterals are present. They
are obviously ventral in origin with stout ceratophores and small tapered papillose
ceratostyles two-thirds the length of the prostomium. The palps and tentacular
cirri are long and tapered and the latter are beset with long slender papillae.

The dorsal cirri (Text-fig, i /) are similar to the tentacular cirri and exceed the
neurosetae in length. Anterior elytra are missing but those in the middle of the body
are large, delicate and faintly reniform in shape (see Text-fig, i h). The lateral
and posterior margins (Text-fig, i m) bear fringes of very long slender papillae with
faintly knobbed tips. Similar papillae are scattered over the elytron surface but
are difficult to see since the elytron is mottled with pale brown and bears numerous
well marked chitinous tubercles. The tubercles near the anterior margin are
minute and hemispherical but the great majority are stout thorns with faintly
curved tips. They are scattered over the rest of the elytron and are uniform in
size except for 4-5 in a row along the posterior margin which are much larger than
the others.

The notopodia are normally developed and bear numerous notosetae. The neuro-
podia (Text-fig, i /) have pointed presetal lobes and shorter, more triangular post-

THE POLYCHAETE FAUNA OF SOUTH AFRICA 389

setal lips. The ventral cirri are very small, markedly tapered and do not extend
to the base of the neurosetae. Nephridial papillae are well developed from setiger
5 onwards.

Notosetae vary in length and are slightly stouter than the neurosetae. An aver-
age notoseta (Text-fig. I _;') has numerous strongly developed rows of serrations
extending over more than half its length and the tip (Text-fig. I /) is pointed and
faintly flanged. The neurosetae also vary in length and development of the secon-
dary tooth. An average neuroseta from the middle of the series (Text-fig. I k)
has over 20 rows of long spinules and a long blade-like naked tip (Text-fig. I k')
with a small secondary tooth. As usual, superior neurosetae are finer with a better
marked secondary tooth and inferior ones are shorter and stouter with a finer
secondary tooth.

Willey (1902) described Harmothoe spinosa var. lagiscoides from Antarctica.
Bergstrom (1916) raised H. lagiscoides to species rank and Hartman (1953) agreed.
According to Bergstrom the development of giant spines on the elytra is variable but
Bergstrom, Hartman and Monro (1936) all agree that the notosetae are either
smooth or only faintly serrated. It is in this character particularly that the South
African form differs from the Antarctic species. Holotype : B.M.(N.H.), Reg. No.
1963.1.2.

Antinoe epitoca Monro, 1930

Antinoe epitoca Monro, 1930 : 67, fig. 19 a-f.

RECORDS. FAL 406(1).

NOTES. As far as I am aware this species has not been recorded since Monro's
original description of the type from dredgings off Angola. The present record
extends its range to False Bay. Monro's specimen lacked elytra but these are
present on the Cape specimens and are large and oval with smooth surfaces and
entire margins.

A. epitoca is easily distinguished from A. lactea Day by the position of the eyes
under the prostomial peaks and the very much fainter serrations on the notosetae.

Subfamily SIGALIONINAE
Pholoe dorsipapillata Marenzeller, 1893
Pholoe dorsipapillata Marenzeller, 1893 : 30; Fauvel, 1923, fig. 23 n-q.

RECORDS. SCD 298 (i).

NOTES, The single specimen is well preserved and permits me to modify Fauvel's
description.

This species is broader and tougher than P. minuta and the elytra are very hard,
almost brittle with successive layers of chitin forming marked concentric rings on
the elytra. The median antenna arises from a short ceratophore and there is a
papillose subterminal swelling preceding the filiform tip. It is usually stated that
in all species of Pholoe there are no lateral antennae, but in this specimen there are

ZOOL. 10, 7 22

390 J. H. DAY

indications that minute lateral antennae are present fused to the dorsal surface
of the bases of the tentacular cirri. Other points of difference from P. minuta
are that P. dorsipapillata has a single tentacular cirrus, not two, and that there are
numerous fine setae at the base. The neuropodium moreover has a long papilla at
its apex and the shaft-heads of the neurosetae are more finely serrated than those
of P. minuta.

Family PISIONIDAE
Pisione africana sp. n.

(Text-fig. 2 a-e)

RECORDS. SCD 136 (i juv.) ; 170 (i-type) ; 212 (6 juvs.). FAL 419 (2 juvs.) ;
424 (2 juvs.).

DESCRIPTION. All specimens except one are juveniles 2-5 mm. long. The single
adult from station SCD 170, dredged at 33 58-9' S./25 41-4' E. is designated the
holotype. It is 15 mm. long with 75 segments and is uniformly pale in alcohol.

The prostomium (Text-fig. 2 a) is small, indistinct and almost completely fused
to the enveloping peristome. A pair of cephalic eyes probably formed by the fusion
of two pairs is visible through the skin at the level of setiger i. The two pairs of
jaws when dissected proved to be curved and pointed with flanged bases but have
no teeth. The large cirriform palps arise from basal sheaths fused to the peristome.
The latter is very large and extends forward completely embracing the prostomium
and bears two pairs of tentacular cirri above the bases of the palps. The superior
pair is long and tapered, simulating antennae, but the inferior pair is small and
papilliform. Both pairs are essentially similar in structure to the cirri of setiger I
and differ from the palps. The peristomial segment also has a pair of very large
acicula whose expanded ends project through the skin in front of the mouth and
probably assist in feeding. Their posterior ends diverge outwards and almost reach
the bases of the first pair of parapodia. The first setigerous segment is relatively
short and not very distinct from the peristomium. The parapodium consists of a
papilliform dorsal cirrus similar to those of later segments, a conical setigerous lobe
and a long tapering ventral cirrus which is directed forwards and just reaches the
bases of the palps. Setiger 2 is essentially similar except that in the adult type
specimen the dorsal cirrus is markedly larger than that of the first or third setiger.
In juveniles this is not obvious. Later parapodia (Text-fig. 2 e) are uniform in
structure. Each has a flask-shaped dorsal and ventral cirrus which is indistinctly
jointed near the ciliated tip and a long setigerous lobe with small presetal and post-
setal lips. No genital papillae were seen.

Each parapodium has two pale internal acicula and about 5 setae. The superior
one (Text-fig. 2 b) is simple and stouter than the rest with an expanded, obliquely
truncate end, sometimes incised to form indistinct teeth. The other setae are all
compound. One is rather slender (Text-fig. 2 c) with a tapering sword-like blade
and the other three (Text-fig. 2 d) have short, straight falcigerous blades.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

391

FIG. 2. Pisione africana : (a) anterior end ; (b) superior simple seta ; (c) superior compound
seta ; (d) inferior compound seta ; (e) parapodium. Phyllodoce schmardaei : (/) anterior
end ; (g) parapodium ; (h) seta.

Pisione is a rare genus and Hartman (1959) listed only 6 species. P. africana
differs from P. remota (Southern, 1914) in the character of the setae and the posses-
sion of an enlarged dorsal cirrus on setiger 2. In this it is closer to P. oerstedi Grube
1857 from Peru as redescribed by Ehlers (1901) but again the setae differ. In par-
ticular the possession of a compound seta with a long sword-like blade is
characteristic.

392 J. H. DAY

There has been much discussion as to the homology of the large cirriform ap-
pendages with sheathed bases at the front of the head. Southern (1914) suggested
that they are palps which have become attached to the large, forwardly produced
peristome. This interpretation which is accepted here makes them homologous to
the palps of Pisionidens indica. In the latter the prostomium is well developed
and anterior to the peristome. It bears a pair of antennae and a pair of palps which
also have sheathed bases. The peristome bears two normal pairs of tentacular
cirri. In Pisione the tentacular cirri are dorsal to the palps and the inferior pair
are greatly reduced. Holotype : B.M.(N.H.), Reg. No. 1963.1.12.

Family PHYLLODOCIDAE
Phyllodoce schmardaei sp. n.

(Text-fig. 2 f-h)

Phyllodoce macrophthalma (non Phyllodoce macrophthalmos Grube, 1857 : 169.)? Schmarda,
1861 : 84, pi. 30, fig. 234 ;? Ehlers, 1913 : 454 ;? Day, 1960 : 297.

RECORDS. FAL 422 (i).

DESCRIPTION. Hartman (1959) showed that the name Phyllodoce macrophthalma
was first used by Grube 1856 and that Schmarda's name is a homonym. There is also
some doubt as to whether the specimen described by Schmarda " in Atlantischen
Meer " is identical to those from France described by Saint Joseph (1888), Fauvel
(1923) and! Day (1960). For these reasons the South African species is given a
new name and described briefly below.

The holotype dredged in False Bay at station FAL 422 is 30 mm. long with 60
segments and when fresh it was uniformly green in colour.

The prostomium (Text-fig. 2 /) is cordate with rather long frontal antennae, a
pair of large eyes and a small occipital papilla. The basal half of the proboscis is
lightly papillose and the distal half has 6 rugose ridges. The first tentacular seg-
ment is fused to the prostomium and not visible dorsally, but the second and third
are distinct and separate. All tentacular cirri are cylindrical and fairly long.
Setae are present on tentacular segments 2 and 3 giving the formula I + S - T -f
5 ^ . Anterior dorsal cirri (Text-fig. 2 g) are cordate and about as broad as
long but become distinctly longer than broad posteriorly. The setigerous lobe
is blunt apically with subequal and rounded presetal lips. The ventral cirrus is
oval with a pointed tip rather shorter than the setigerous lobe. The setae (Text-fig.
2 h) have oval shaft-heads which are denticulate distally and blades of normal
length. Holotype : B.M.(N.H.), Reg. No. 1963.1.13.

Phyllodoce tubicola sp. n.

(Text-fig. 3 a-c)

RECORDS. SCD 200 (i), 232 (i) ; FAL 397 (i).

DESCRIPTION. The holotype is from station SCD 200 dredged on the Agulhas

THE POLYCHAETE FAUNA OF SOUTH AFRICA

393

Bank at 34 10' 8/23 32' E. in 97 metres. It is encased in a delicate and transparent
tube which is closely ringed. The body is brownish in alcohol without any obvious
pattern. It is 30 mm. long and very slender with about 80 segments.

FIG. 3. Phyllodoce tubicola : (a) anterior end ; (b) parapodium ; (c) seta. Phyllodoce longipes :
(d) parapodium ; (<?) anterior end ; (/) seta. Ancistrosyllis parva : (g) anterior end ; (h)
parapodium ; (;') long central neuroseta ; (k) short outer neuroseta.

394 J.H.DAY

The prostomium (Text-fig. 3 a) is oval and longer than broad with a small posterior
occipital notch and a minute occipital papilla. The frontal antennae are rather
long but the eyes are not enlarged. The proboscis on dissection proved to have small
irregular papillae on the basal part, a brown band near the middle and then six
rugose ridges distally.

The first tentacular segment is fused to the prostomium and not visible dorsally
but the second and third are distinct and separate. The tentacular cirri are all
rounded in section and very long, the longest being D 2 which reaches back to
setiger 8. Setae are present only on the third tentacular segment, the formula
thus being i + Oj + S^.

The dorsal cirri (Text-fig. 3 b] are oval to cordate anteriorly but become longer,
almost hastate posteriorly. The setigerous lobe is bluntly rounded distally with
a notched presetal lip. The ventral cirrus is pointed and obviously longer than the
setigerous lobe. Each parapodium bears about 12 setae with oval, striated shaft-
heads and long tapering blades (see Text-fig. 3 c).

The only other South African species with setae restricted to the third tentacular
segment is P. (Anaites) capensis, but here the prostomium is partly covered by the
fused first and second tentacular segments. Several European species listed by
Fauvel, 1923 have the same tentacular formula but differ either in the shape of the
prostomium or the character of the proboscis. The genus Phyllodoce does not norm-
ally live in a tube and it is suspected that the delicate tube in which the holotype
(but not the paratypes) was encased may have been a temporary mucous secretion
which hardened around the worm during preservation. Holotype : B.M.(N.H.),
Reg. No. 1963 . i . 14.

Phyllodoce (Anaitides) longipes Kinberg, 1865
(Text-fig. 3 d-f)

Phyllodoce longipes Kinberg, 1865 : 241 ; Ehlers, 1901 : 72, pi. 7, figs. 1-4. (non Phyllodoce
longipes : Monro, 1930 : 73, fig. 21 ; nee Monro, 1936 : 101)

RECORDS. FAL 342 (3).

DESCRIPTION. The largest specimen is 25 mm. long but the last few segments are
missing. The anterior end is shown in Text-fig. 3 e. The first three segments
following those that bear tentacular cirri are dusky, the later ones have three rows
of blurred spots. The prostomium is longer than broad with a small posterior
notch but no occipital papilla. The proximal half of proboscis has 6 long regular
rows of papillae at the sides, and about 14 papillae per row. The distal half has
6 rugose ridges. The first tentacular segment is fused to the prostomium and is not
visible dorsally ; the second segment is indistinct, but the third segment is distinct
and separate. Tentacular cirri are all tapered and rounded in section, the longest
being D 2 which is very long. No setae are present on any tentacular segment the
formula being i -f \ + -^. The dorsal cirri (Text-fig. 3 d) are large
and cordate anteriorly but become broader, almost reniform posteriorly. They are
dusky, often with a central darker spot. The setigerous lobe has the superior

THE POLYCHAETE FAUNA OF SOUTH AFRICA 395

division of the presetal lip long and pointed and the ventral cirrus has a pointed
tip as long as the setigerous lobe. The setae (Text-fig. 3/) have oval striated shaft-
heads and serrated tapering blades.

These South African specimens agree very well with Ehlers' description and figures
and have the same three characteristically dusky segments following those that
bear the tentacular cirri. But Ehlers' account of the tentacular segments is confused
and in the text he makes no mention of the fact that the first two segments are
fused to the prostomium. Bergstrom (1914, p. 149), who examined Kinberg's type
from Valparaiso found it in a poor condition and does not discuss the point, but his
text-figure 47 is an excellent illustration of the characteristic parapodium.

Monro's specimens were examined in the British Museum and, as he stated, they
have setae in the second and third tentacular segments. The coloration is also
different and they obviously belong to a different species. Incidentally there is no
aciculum in the dorsal cirrophore.

Family PILARGIDAE
Ancistrosyllis parva sp. n.

(Text-fig. 3 g-k)

RECORDS. FAL 390 (2) ; SCD 200 (2).

DESCRIPTION. The holotype is the largest specimen from station SCD 200
dredged at 34 10' S./23 32' E. in 97 metres. It measures 10 mm. by 0-7 mm. for
60 segments. The body is rather flattened, tapered posteriorly, and pale in alcohol.
There is no narrowed neck region.

The prostomium (Test-fig. 3 g) is broader than long without any posterior in-
dentation. A pair of palps with broad flattened palpophores and small palpostyles
arise from the anterior margin with a notch between them. There are three tapered
antennae each equal to half the width of the peristome but there are no eyes. The
proboscis is everted and muscular with 12 marginal papillae ; the 8 dorsal ones are
stout and conical but the 4 ventral ones are smaller. The peristome bears 2 pairs
of tentacular cirri which are the same length as the antennae. The first setiger has a
dorsal cirrus which is even longer than the tentacular cirri, a normal setigerous lobe
and a ventral cirrus. The second setiger has a rather short dorsal cirrus and a
normal setigerous lobe but no ventral cirrus. Subsequent parapodia (Text-fig.
3 h) are all essentially similar. The dorsal cirri are smoothly tapered and there is a
vague joint between the cirrus and its short stout cirrophore. The setigerous lobe
is bluntly truncate and the ventral cirrus is similar to the dorsal one but smaller.

There are no notosetae on the first three setigers. From the fourth onwards a
smooth, stout recurved hook arises from the dorsal cirrophore which is thus homo-
logus to a vestigal notopodium. The neurosetae are all simple capillaries. The
short outer ones (Text-fig. 3 k) have long spinules along one margin but the much
longer setae in the centre of the bundle (Text-fig. 3 _;') have smooth blades.

A . parva is allied to A . tentaculata Treadwell 1941 from Long Island Sound,
U.S.A. Both are small species with long antennae and dorsal cirri. Both lack a

396 J. H. DAY

ventral cirrus on setiger 2 and have notopodial hooks from setiger 4 onwards. But
Treadwell figures A. tentaculata as having the posterior margin of the prostomium
deeply excavate. In A. parva the posterior margin of the prostomium is smoothly
curved.

Two species of Ancistrosyllis have previously been recorded from southern Africa.
Ehlers (1908) recorded A. robusta from Angola. This is a much larger species
with notopodial hooks appearing on setiger 69. Ehlers does not mention whether
there is a ventral cirrus on setiger 2 or not. Ancistrosyllis cf. constricta described by
me (Day, 1957, p. 71) is almost certainly a juvenile. It lacks a ventral cirrus on
setiger 2 but the notopodial hooks start on setiger 8 and from setiger 3 onwards there
are 1-2 fine notopodial capillaries in all feet. Holotype : B.M.(N.H.) Reg. No.
1963 .1.22.

Loandalia capensis sp. n.

(Text-fig. 4 a-d)

RECORDS. SCD 220 (2).

DESCRIPTION. The holotype which was dredged at 34 02' S./23 28-4' E. in
49 metres consists of one specimen broken into three pieces with a total of 56 seg-
ments measuring 15 mm. The pygidium is missing. There is also a headless
fragment of a smaller specimen. The body is rounded in section and about 0-8
mm. broad. The segments are about twice as broad as long though the first few
segments are shorter, about 4 times as broad as long. The general appearance is
that of a Maldanid for the head appendages are not obvious, the parapodia are very
small and there are deep constrictions between the segments.

The head and first three setigers (Text-fig. 4 a] are covered with conical papillae
but the rest of the surface is smooth.

The prostomium is small and not clearly separated from the peristomium. It
bears a pair of palps with large palpophores and small palpostyles but there are no
antennae nor eyes. The proboscis is partly everted and is soft and unarmed. The
peristome is clearly separated from setiger i. Like the other anterior segments it is
broader than long and careful examination shows that it bears two pairs of tentacular
cirri. These are not much larger than the scattered papillae but quite definite.

The parapodia are inconspicuous throughout. There is a low papilliform notopodial
lobe supported by an aciculum but no dorsal cirrus. The neuropodia are a little
larger and broadly conical with 2 acicula and a minute ventral cirrus just below the
apex. None of the parapodia has any sign of gills.

The first 7 setigers have no notosetae, only an internal aciculum. From setiger
8 onwards each notopodium (Text-fig. 4 b) bears a single stout recurved hook (Text-
fig. 4 c) which is granular near the apex. There are no capillaries. The neurosetae
(Text-fig. 4 d) are 3-4 capillaries per parapodium. Each has a narrow hairlike blade
which is either smooth or very faintly serrated and tapers to a slender tip.

The absence of antennae and the presence of very small tentacular cirri places this
species in an intermediate position between Loandalia and Pilargis. The poorly
developed parapodia suggest that it is closer to Loandalia.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 397

L. capensis is the fourth recorded species of this rare genus. The other three lack
tentacular cirri and L. abermns Monro, 1936 which was recorded from Angola also
has gills on the posterior parapodia. Holotype : B.M.(N.H.), Reg. No. 1963.1.25.

Family HESIONIDAE
Oxydromus capensis sp. n.

(Text-fig. 4 e-j)

RECORDS. SCD 200 (i) ; FAL 390 (i + i juv.) ; WCD 112 (i), 118 (i).

DESCRIPTION. The holotype was selected from station WCD 112 dredged off
Saldanha Bay in 141 metres on dark green mud. It is 5 mm. long for 20 segments
but the posterior half is missing and the complete worm was probably double this
length. The body is uniformly pale in alcohol.

The prostomium (Text-fig. 4 h) is roughly rectangular and twice as broad as
long. The three antennae all arise from the anterior margin and the median is
only half the length of the laterals. The biarticulate palps are much stouter than
the lateral antennae but not longer. The palpostyle is ovoid and slightly shorter
than the palpophore. The prostomium bears two pairs of eyes and the anterior
pair are reniform and larger than the posterior pair. The proboscis is muscular
and lacks jaws but has 10 conical, widely separated papillae around its margin.

There are 8 pairs of faintly articulated tentacular cirri borne on 4 partially fused
segments. The first pair of tentacular cirri are shorter than the rest which are
considerably longer than the body is broad.

The first 5-7 segments are uniramous and the rest biramous. The dorsal cirri
are mounted on short cirrophores and are smooth and tapered, not annulated.
Apart from the first which is equal to three-quarters of the body width, the dorsal
cirri are about equal to half the body width. The first few cirrophores have inter-
nal acicula but no setae ; from setiger 5 onwards a notopodial papilla appears on
the ventral side of the cirrophore and by setiger 7 it obviously bears setae. Whether
small notosetae are present on setigers 5 and 6 is doubtful. On later parapodia
(Text-fig. 4 j) the notopodium is small but quite conspicuous and bears about 10
setae. The neuropodia are always well developed. Each is roughly cylindrical
with numerous setae, and bears a triangular presetal lobe, a shorter rounded
postsetal lip and a tapered ventral cirrus.

The notosetae include capillaries with smooth flattened blades (Text-fig. 4 /)
and 3-4 forked setae (Text-fig. 4 e) . The neurosetae (Text-fig. 4 g) are all compound
and falcigerous with bidentate blades of varying length. The margin of the blade
is minutely serrated and the bidentate apex is formed by a strong terminal tooth and
a long slender secondary tooth.

The only record of Oxydromus from southern Africa is Oxydromus sp. Augener,
1918, p. 225 from Angola. This is incompletely known and possibly has jaws, in
which case it belongs to a different genus. O. capensis is close to 0. arenicolus La
Greca, 1946 from the Mediterranean and to the subspecies 0. a. glabrus Hartman,
1961 from California. All three have 10 widely spaced marginal papillae on the
proboscis while other species have numerous marginal papillae. However, 0.

ZOOL. IO, 7 22

398 J. H. DAY

capensis differs from 0. arenicolus in that the notosetae do not appear before setiger
5 and include both forked setae and capillaries. Forked setae were also reported
by Augener for the doubtful Oxydromus sp. from Angola. 0. arenicolus is also said
to have annulated dorsal cirri whereas those of 0. capensis appear to be smooth.
Holotype : B.M.(N.H.), Reg. No. 1963.1.28.

FIG. 4. Loandalia capensis : (a) anterior end ; (b) I2th parapodium ; (c) notopodial hook ;
(d) neuroseta. Oxydromus capensis : (e) notopodial forked seta ; (/) notopodial capillary ;
(g) neuroseta ; (h) anterior end ; (;') anterior view of parapodium. Syllis benguellana :
(k) anterior end ; (1) parapodium ; (m) seta.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 399

Family SYLLIDAE
Syllis benguellana sp. n.

(Text-fig. 4 k-m)

RECORDS. WCD 91 (ca. 100).

DIAGNOSIS. A threadlike species with slender dorsal cirri having about 15 long
joints. Setae unidentate.

DESCRIPTION. Numerous specimens were obtained by a van Veen grab at station
WCD 91 at 32 05-5' S./i8 17-3' E. in 27 metres on white sand. Among them were
numerous mature males and ovigerous females. Ten specimens were selected as the
types and one of these, a mature male measuring 9 mm. by 0-4 mm. for 90 segmments
is designated the holotype. The anterior part of the body is colourless but from
segment 60 it is swollen and white with sperm.

The proboscis is extruded and shows a small dorsal tooth. The pharynx when
retracted would continue back to setiger 9 and the proventriculus which has 35
rows of points is the same length as the pharynx. The prostomium (Text-fig. 4 k)
is broader than long with 2 pairs of small eyes. The antennae are slender and
clearly annulated, the median which is posterior in origin having 14 long joints and
the laterals having 12. The palps are broad and separate at the base. There is no
occipital flap.

The tentacular cirri and dorsal cirri (Text-fig. 4 /) are essentially similar, all slender
with 12-14 long joints and equal to two-thirds of the segmental breadth. Posterior
cirri are shorter and have only 10-12 joints but they are never stout and fusiform
as in 5. armillaris. The ventral cirri are rather long and project beyond the seti-
gerous lobes.

There are two knobbed acicula per parapodium and about 6 compound setae.
Each of these (Text-fig. 4 m) has a straight unidentate blade of normal length. A
simple needlelike superior seta appears in the last few feet.

This species is obviously close to 5. armillaris and agrees with it in having uniden-
tate setae. However, the dorsal cirri are quite different, for in contrast to the short
fusiform dorsal cirri of 5. armillaris the cirri of 5. benguellana are filiform, with 10-14
elongated joints. Holotype : B.M.(N.H.), Reg. No. 1963.1.50.

Eusyllis assimilis Marenzeller, 1875

Eusyllis assimilis Marenzeller, 1875 : 158 ; Fauvel, 1923 : 294, fig. 112 a-g.
Eusyllis monilocornis : Fauvel 1923* : 6.

RECORDS. WCD 41 (i), 87 (i).

NOTES. E. assimilis was recorded from southern Angola by Fauvel (19230:)
under the name of E. monilicornis. Fauvel's statement that the bidentate setae
have both long and blades shows that his record refers to E. assimilis and not E.
blomstrandi.

This is the first record from South Africa.

400 J. H. DAY

Anguillosyllis gen. n.

Prostomium with 3 minute antennae and a pair of large pointed palps fused for
half their length. A single pair of minute tentacular cirri. Pharynx straight,
unarmed. Dorsal cirri long but not annulated. Ventral cirri present. Setigerous
lobe of the parapodium with a contractile dorsal projection.

Setae compound and falcigerous.

TYPE SPECIES. Anguillosyllis capensis Day, 1963.

Anguillosyllis capensis sp. n.

(Text-fig. 5 a-ct)

RECORDS. SCD 275 (2) ; WCD 106 (i).

DESCRIPTION. The type material consists of two incomplete specimens from
station SCD 275 dredged at 34 51' S./23 41' E. in 183 metres on fine sand. One
is extended and measures 3-5 mm. for 12 segments and the other is contracted and
measures 2-5 mm. for 13 segments. The description given below is based on the
extended specimen which is designated the holotype.

The head end (Text-fig. 5 b] is pointed due to the tapered palps which project
well forward and are fused for over half their length. The prostomium is oval
and broader than long. There are no eyes but three short club-shaped antennae
are present with the median further back than the two laterals and difficult to see.
The peristome bears a single pair of lateral tentacular cirri which are rather smaller
than the antennae.

The pharynx when dissected proved to be short, broad and straight and reached
setiger 3. Careful dissection failed to reveal the dorsal tooth which was expected
and the whole pharynx is smooth. The pharyngeal sheath appears to have 6
soft papillae. The proventriculus is barrel-shaped and extends through setigers
4, 5 and 6 and has about 30 rows of points.

The segments and parapodia are similar throughout. Each segment is more than
twice as broad as long and each parapodium (Text-fig. 5 c] has a long dorsal cirrus,
a characteristic setigerous lobe and a slender ventral cirrus. The dorsal cirrus is
very long and slender, without any sign of annulation and is usually twisted and
coiled. The setigerous lobe is well developed with a dorsal hood over the setae which
tapers to a cirriform projection. When retracted (Text-fig. 5 d) the dorsal hood is
no longer obvious and the end of the setigerous lobe then appears to have three lips ;
one superior, one anterior and one posterior. The ventral cirrus is slender and
arises from the distal end of the setigerous lobe just below the setae.

The setae (Text-fig. 5 a) are very numerous and all similar. Each is compound
and falcigerous with a rather long, minutely serrated blade and a blunt unidentate
tip.

The fusion of the palps and the presence of a single pair of tentacular cirri place
this species in the sub-family Exogoninae ; further the well developed dorsal cirri
suggest that it is closer to Sphaerosyllis than Exogone. But it is obviously distinct

THE POLYCHAETE FAUNA OF SOUTH AFRICA 401

from both these genera. The lack of a dorsal tooth on the pharynx is unusual and
the curious dorsal hood over the parapodium is unique. Holotype : B.M.(N.H.)
Reg. No. 1963.1.29.

Exogone normalis sp. n.

(Text-fig. 5 e-h]

RECORDS. SCD 130 (3), 212 (2), 215 (i) ; FAL 419 (3).

DESCRIPTION. The type material consists of three specimens from station SCD
130 obtained by a van Veen grab at 34 48' S./22 06' E. in 100 metres on a sandy
bottom on 3rd June, 1960. The largest specimen measuring 5-5 mm. by 0-3 mm.
for 50 segments is designated the holotype. The paratypes are slightly smaller.
All are creamy brown in alcohol.

The palps (Text-fig. 5 g) are large, about three times as long as the prostomium
and broader basally. They are completely fused dorsally but broadly grooved
ventrally. The prostomium is three times as broad as long with three pairs of
large eyes which are coalescent on each side in the holotype but there are only two
pairs of small eyes in the two paratypes. The three antennae arise close together on
the posterior margin of the prostomium. The median is longer than the prosto-
mium and reaches the base of the palps but the laterals are very small, only one-
third the length of the median. The pharynx has a single dorsal tooth and extends
back to setiger 5. The proventriculus which is the same length extends on to
setiger 10 and has about 23 rows of points.

The tentacular segment is about the same length as the prostomium and quite
distinct from it. There is a nuchal organ at the junction between the prostomium
and peristomium and the latter bears a single pair of very small tentacular cirri
which are only half the size of the first pair of dorsal cirri.

The body is cylindrical and quite smooth. The parapodia (Text-fig. 5 h) are
bluntly conical lateral projections. The dorsal cirri are small ovoid papillae well
above the parapodia. There is no dorsal cirrus on setiger 2. The ventral cirri
are twice the length of the dorsal cirri and project beyond the ends of the parapodia.

Anterior segments bear 12 setae but the number is reduced to about 5 posteriorly.
The first 9 segments bear compound setae only, then a superior simple seta appears
and finally, in the last few segments an inferior simple seta as well. All the compound
setae (Text-fig. 5 /) are similar with moderate shaft-heads and straight unidentate
blades of normal length. It is emphasised that there is no specialised superior
compound seta. The superior simple seta (Text-fig. 5 e) is a straight blunt needle
which gradually increases in thickness until, in posterior segments it is twice the
diameter of the compound setae. The inferior seta is the same diameter as the
compound seta.

The pygidium bears a pair of anal cirri. Holotype : B.M.(N.H.), Reg. No.
1963.1.30.

Three species of Exogone have been recorded from South Africa and there is a
very doubtful record of the Antarctic species E. heterosetosa from Madagascar.
E. clavator Ehlers and E. verugera Claparede both have dorsal cirri on setiger 2 and

402

J. H. DAY

e f

FIG. 5. Anguillosyllis capensis : (a] seta; (fc) anterior end ; (c) expanded parapodium ;
(d) contracted parapodium with dorsal cirrus omitted. Exogone normalis : (e) superior
simple seta ; (/) compound seta ; (g) anterior end ; (h) parapodium. Exogonoides anten-
nata : (j) simple seta ; (k) compound seta from 2nd foot ; (/) dorsal view of anterior end ;
(m) ventral view of head ; (n) parapodium.

both have a superior compound seta with a long swordlike blade. E. heterosetosa
Mclntosh and E. gemmifera Pagenstecher both lack a dorsal cirrus on setiger 2 and
in this respect they agree with E. normalis but their setae differ. Their normal

THE POLYCHAETE FAUNA OF SOUTH AFRICA 403

compound setae are bidentate and both have a specialised superior compound seta.
In E. heterosetosa it has a very swollen shaft-head and a broad triangular blade
and in E. gemmifera it has a normal shaft-head and a long swordlike blade.

Exogonoides gen. n.

Prostomium with three ovoid frontal antennae. Palps small, ventral, partially
fused. Pharynx straight without obvious teeth. Proventriculus poorly developed.
A single pair of small tentacular cirri. Dorsal cirri ovoid. Ventral cirri absent
(fused to setigerous lobe) . Setae compound anteriorly but simple posteriorly.

TYPE SPECIES. E. antennata Day, 1963.

Exogonoides antennata sp. n.

(Text-fig. 5 j-n)

RECORDS. SCD 246 (2).

DESCRIPTION. The type material was dredged at station SCD 246 at 34 02' S./
23 28' E. in 49 metres on sand, mud and rock. One of the two specimens has a
damaged head. The other, measuring 35 mm. by 0-5 mm. for over 200 segments is
designated the holotype. It is creamy brown in alcohol and lacks colour markings.
The general appearance is that of a very long Exogone, hence the name.

The prostomium (Text-fig. 5 /) is bluntly conical with three stout ovoid antennae
projecting forward from the anterior end, the median being superior to the two
laterals. There are two pairs of eyes of which the anterior pair is larger and wider
apart than the posterior pair. The palps (Text-fig. 5 m) are small and not visible
dorsally. They arise from the antero-ventral margin of the prostomium as two
cushion-like lobes in contact with one another. While they are obviously not fused
completely it is impossible to say that they are entirely separate and it is presumed
that they are partially fused. The pharynx and proventriculus when dissected
proved to be slender and attached by mesenteries throughout. Even when the
pharynx was removed and cleared, no teeth or trepan was seen and it is thought to
be unarmed. The pharynx which extends back to setiger 6 is straight anteriorly
but somewhat sinuous where it meets the proventriculus which is small, dark and
unusually thin walled. About 15 rows of lumps corresponding to the " points "
of a normal gizzard were seen.

The tentacular segment is as long as the body segments and bears a single pair
of small ovoid tentacular cirri similar to those found in the genus Exogone. There
is no sign of a second pair.

The body is very long and slender, anterior segments being about 8 times as
broad as long but posterior ones are longer, only twice as broad as long. The body
surface is quite smooth.

The parapodia (Text-fig. 5 n) are similar throughout. The dorsal cirrus is ovoid
in outline, somewhat compressed and about half the size of the setigerous lobe.
The setigerous lobe itself is a broad lateral projection with the ventral cirrus com-
pletely fused to the postero-ventral margin so that ventral cirri may be said to be

4 o 4 J. H. DAY

absent as in the Autolytinae. The setae are usually three in number and most of
them are simple spines with bluntly triangular tips. However, in the first few feet
the setae (Text-fig. 5 k) bear short, triangular bidentate blades so that the later
simple setae (Text-fig. 5 j) are obviously homologous with the shafts of the anterior
compound forms.

This species cannot be assigned to any known genus. The single tentacular cirrus
and short ovoid antennae and cirri are reminiscent of Exogone but the latter has
completely fused palps projecting forwards, the antennae are dorsal not terminal,
and there is a well marked dorsal tooth and distinct ventral cirri. In these respects
the present species is related to the Autolytinae but the latter has two pairs of ten-
tacular cirri. It is suggested that the subfamily Autolytinae be enlarged to include
forms with one pair of tentacular cirri and the new genus Exogonoides be included
in it. Holotype : B.M.(N.H.), Reg. No. 1963.1.34.

Family NEREID AE
Micronereides gen. n.

Small Nereidae generally similar to Micronereis but possessing two antennae.
Proboscis without paragnaths but with a pair of toothed jaws. Prostomium with
a pair of antennae and a pair of biarticulate palps. Two pairs of tentacular cirri.
No apodous segment behind the peristome. First two segments uniramous, sub-
sequent ones biramous. Setae are all homogomph spinigers.

TYPE. Micronereides capensis Day.

Micronereides capensis sp. n.

(Text-fig. 6 a-e)

RECORDS. SCD 233 (i).

The holotype is a single specimen measuring 6 mm. for 34 segments. It was
dredged at Station SCD 233 position 36 28-5' S./2i n' E. on the Agulhas Bank in
183 metres on a bottom of fine sand. There is no colour pattern. The pro-
stomium (Text-fig. 6 b) is broad, almost square with the eyes either indistinct or
absent. A pair of widely separated antennae is present. The palps have stout
palpophores completely fused to the prostomium and slender palpostyles longer
than the antennae. The proboscis lacks paragnaths but has a pair of large jaws
with 7 teeth. There are two pairs of short tentacular cirri arranged 2 : 2 showing
that they have been derived from two fused segments. There is no anterior apodous
segment. The first two setigerous segments are uniramous, the notopodial lobes
and notosetae being absent though the dorsal cirrus remains. Subsequent segments
have biramous parapodia which are similar in structure throughout. The pygidium
is unknown.

Normal parapodia (Text-fig. 6 c] are deeply cleft between the notopodia and
neuropodia. The dorsal cirrus is rather short and the superior lobe of the notopo-
dium is absent. The inferior lobe of the notopodium however, is elongated and

THE POLYCHAETE FAUNA OF SOUTH AFRICA

405

FIG. 6. Micronereides capensis : (a) seta. ; (b) anterior end. ; (c) parapodium. Nereis
(Neanthes) agulhana : (d) falcigerous neuroseta ; (e) anterior end ; (/) dorsal view of
proboscis ; (g) ventral view of proboscis ; (h) loth foot ; (j) posterior foot.

zooi. 10, 7

406 J. H. DAY

exceeds the dorsal cirrus. There is no intermediate (setigerous) lobe. The neuro-
podium consists of a stout, conical setigerous lobe, a very small inferior lobe and a
small ventral cirrus.

The setae (Text-fig. 6 a) are all homogomph spinigers though the inferior series
in the neuropodia have rather short blades and the shaft-heads might be considered
hemigomph.

REMARKS. The new genus Micronereides is obviously close to Micronereis
Claparede, 1863 in which Hartman (1959) recognised three species, M. halei Hart-
man, 1954 from South Australia, M. nanaimoensis Berkeley and Berkeley, 1953
from Western Canada and M. variegata Claparede, 1863 from France.

I have not seen a description of M. nanaimoensis but the other two species both
lack antennae and it would seem unwise to enlarge the generic description of the
well established genus Micronereis to include forms with two antennae. Holotype :
B.M.(N.H.), Reg. No. 1963.1.60.

Nereis (Neanthes) agulhana sp. n.

(Text-fig. 6 d-j)

Nereis (Neanthes) cf. kerguelensis : Day, 1960 : 321.

RECORDS. SCD 156 (3), 182 (4), 207 (6 juvs.), 237 (i juv.), 246 (i).

The holotype is a specimen measuring 20 mm. with 65 segments selected from the
material dredged at station SCD 156 at 34 03' S./25 59' E. in 84 metres on a rocky
bottom. The colour in alcohol is pale with touches of brown on the prostomium,
a well marked and characteristic bar across the dorsum of setiger 2 and a pair of small
spots on each of the succeeding anterior segments (see Text-fig. 6 e) .

The antennae and tentacular cirri are rather long and slender and the prostomium
itself is longer than broad.

The proboscis (Text-figs. 4/ and g) has group I = o ; II = 8-9 in a double row ;
III = a group of 5-6 ; IV = a wedge of 8-12 ; V = o ; VI = 2-4 close-set points
in a transverse line ; VII and VIII = a single row of 3-5.

Anterior feet (Text-fig. 6 h) have pointed lobes and slightly longer dorsal cirri.
The notopodium has three lobes, the middle lobe bearing the notosetae being well
marked but obviously smaller than the other two. The setigerous lobe of the
neuropodium has no projecting presetal or postsetal lips.

In posterior feet (Text-fig. 6 f) the dorsal cirrus is rather smaller and the notopo-
dium has lost its third or setigerous lobe. The two notopodial lobes that remain
are stout, markedly pointed and divergent. The structure of the neuropodium is
the same as in the anterior feet.

The notosetae are homogomph spinigers throughout, there being no notopodial
falcigers. The neurosetae include the normal series of homogomph and hetero-
gomph spinigers and heterogomph falcigers. The latter (Text-fig. 6 d) have rather
straight blades with a tendon attached to the blunt tip.

This species is close to N. kerguelensis Mclntosh, 1885 which has now been reported
from many parts of the world including the Mediterranean. However an examina-
tion of Mclntosh's type in the British Museum (Reg. No. 1885 : 12 : i : 170) and

THE POLYCHAETE FAUNA OF SOUTH AFRICA 407

other specimens identified as N. kerguelensis from sub-antarctic localities by Monro,
1930 shows that the South African material is distinct.

In the type of N. kergulensis the proboscis has group I with 2 points in line, V = o,
VI a single small point and VII and VIII a row of 5 points. The prostomium is
broad, almost oval, and the colour of the body is uniformly pale. In anterior feet
the third (setigerous) lobe of the notopodium is small and the setigerous lobe of
the neuropodium has a prominent, almost digitiform postsetal lip. This is not
shown in Mclntosh's pi. 35, fig. n but it is quite distinct in the type, and is referred
to by Monro, 1930. Holotype : B.M.(N.H.), Reg. No. 1963.1.61.

In brief the important differences between N. agulhana which is apparently com-
mon in deep water dredgings around South Africa and the sub-antarctic N. kergue-
lensis concern the distribution of paragnaths, the presence of a brown bar across
setiger 2 which is present in every specimen of N. agulhana, the structure of the
setigerous lobe of the neuropodium and according to Ramsay (1914) and Monro
(1936) the lack of heterogomph spinigers in the neuropodia of N. kerguelensis.

N. papillosa Day 1963 has a similar arrangement of paragnaths but is immediately
separated by the presence of papillae on the parapodial lobes of middle feet.

Family SPHAERODORIDAE
Sphaerodorum benguellarum sp. n.

(Text-fig. 7 a-c)

RECORDS. WCD no (i).

DESCRIPTION. The type locality is 32 08' S./iy 39' E. in 172 metres. The
single specimen (Text-fig. 7 6) is ovoid, 2-2 mm. long by 0-8 mm. broad with 24
segments. The head is retracted and its appendages are not visible, but a pair of
dark eyes was seen on clearing with glycerine. There is a large muscular gizzard.

The whole body is covered with papillae. Each segment has an anterior row
of 8 large papillae across the dorsum and a posterior, more irregular row of about
20 small papillae. The papillae above and below the parapodia are not larger than
the rest. The parapodia themselves (Text-fig. 7 c) are wrinkled and obviously
retractile. Each is conical in shape with a large presetal papilla, about 12 setae and
two smaller postsetal papillae.

The setae (Text-fig. 7 a) are all compound, each with a swollen shaft-head and a
long tapered falcigerous blade. The tip is minute but probably unidentate. Holo-
type : B.M.(N.H.), Reg. No. 1963.1.75.

Sphaerodorum capense sp. n.

(Text-fig. 7 d-f)

RECORDS. CP 452 i specimen found among Gunnarea tubes on the shore at
Mouille Point, Cape Town on ist October, 1955.

DESCRIPTION. The holotype (Text-fig. 7 d) is ovoid in shape 2-5 mm. long, 0-8
mm. wide and has 16 setigers. The colour is yellowish, with a pair of black eyes

4 o8 J. H. DAY

and the dark contents of the gut showing through the skin. The head is retracted
and its appendages are indistinguishable from the numerous spherical papillae which
cover the whole surface.

Each body segment has two transverse rows of papillae across the dorsum. The
anterior row consists of about 18 larger papillae and the posterior irregular row has
many more smaller ones about half the size of those in the anterior row. The papillae
above the parapodia are not obviously enlarged as in S. gracile. The parapodia
(Text-fig. 7 /) are small, obviously retractile and covered with papillae with a larger
one apically. There are about 10 long setae per parapodium. Each seta (Text-fig.
7 e) is compound with a broad oval shaft-head and an indistinctly separated blade
ending in a small, hooked unidentate tip. 5. capense may be distinguished from other
species of the genus by the characteristic setae and the distribution of the papillae.
Holotype : B.M.(N.H.), Reg. No. 1963.1.74.

Family GLYCERIDAE
Goniadella gracilis (Verrill, 1873)

Eone gracilis Verrill, 1873 : 596.

Goniadella gracilis : Hartman, 1950 : 42, pi. 5, figs. 4-8.

RECORDS. FAL 413 (3) ; SCD 212 (i), 218 (6), 237 (2), 245 (i), 275 (3), 288 (i).

NOTES. This small threadlike species which averages 10-15 mm. in length seems
to be common in grab samples around the Cape. The genus Goniadella is close to
Goniada but may be distinguished by the possession of both falcigerous and spini-
gerous neurosetae instead of spinigers alone. G. gracilis (Verrill) is the only species
known and the long digitiform presetal lip on the neuropodium is characteristic.
The South African specimens differ from Hartman's description only in a few minor
points. The proboscis has about 30 chevrons at its base, instead of 25, no eyes were
found on the terminal ring of the prostomium, only on the basal one and there are
fewer notosetae in posterior segments.

Family EUNICIDAE

Subfamily EUNICINAE
Marphysa purcellana Willey, 1904
Marphysa purcellana Willey, 1904 : 263, pi. 13, fig. 17 ; Day, 1953 : 435.

RECORDS. Seven specimens obtained by a dredger prospecting for diamonds
north of the Orange River mouth in about 5-10 fathoms.
NOTES. This is a new record for South West Africa.

Subfamily ONUPHIDINAE
Diopatra dubia Day, 1960

Diopatra dubia Day, 1960 : 348, fig. 10 g-k.

RECORDS. SCD 288 (i)

THE POLYCHAETE FAUNA OF SOUTH AFRICA

409

FIG. 7. Sphaerodorum capense : (a) seta ; (b) entire worm ; (c) parapodium. Sphaerodorum
benguellarum : (d) entire worm ; (c) seta ; (/) parapodium. Lumbrineris meteorana : (g)
anterior end ; (h) maxillae (_;') compound hook ; (k) simple hook ; (/) anterior view of
6th foot ; (ra) anterior view of 5oth foot ; (n) anterior view of far posterior foot.

NOTES. Although this species seems to be widely distributed on the Atlantic
coast of South Africa and is fairly common in False Bay, this record extends its
range to the Indian Ocean off East London.

4 io J. H. DAY

Onuphis quinquedens Day, 1951

Onuphis quinquedens Day, 1951 : 40, fig. 6 a-h.

RECORDS. Muizenberg beach (locally common).

NOTES. This species has not been seen since its original discovery in Zululand.
The new record extends its known range to False Bay. It is apparently common in a
small area at the sheltered western end of Muizenberg bathing beach where it burrows
in sand near low tide. The new specimens are much larger than the holotype and
reach a length of 350 mm. by 8 mm. The notch between the tentacular cirri is
distinctive.

Subfamily LUMBRINERINAE
Lumbrineris heteropoda difficilis nom. nov.

Lumbrineris heteropoda, difficilis nom. nov. pro Lumbrineris heteropoda var. atlantica (non Lum-
briconereis atlantica Kinberg, 1865 : 568 ; Fauvel 1934 : 2 7) Day, 1960 : 360, fig. 12 c-d.
Lumbriconereis heteropoda : Monro, 1930 : 137 ; Monro, 1936 : 154.

RECORDS. WCD 90 (2 + 2 juvs.), 94 (2 + 3 juvs.), 103 (12), 121 (2).

NOTES. This subspecies differs from L. heteropoda in the extreme length of the
posterior feet and the possession of dark acicula. Its relationship to L. h. heteropoda
has been discussed by me (Day, 1957, 1960) under the name L. heteropoda var.
atlantica. Hartman (1959) showed that the name atlantica has been used both by
Kinberg, 1865 and Fauvel, 1934 so that these South African specimens must be
renamed.

Lumbrineris meteorana (Augener, 1931)
(Text-fig. 7 g-w)

Lumbriconereis meteorana Augener, 1931 : 300, fig. 8 a-e (partim).
Lumbrineris cf. meteorana : Day 1960, p. 358.

RECORDS. WCD 60 (4), 61 (5), 63 (28), 64 (n), 66 (8), 67 (i), 97 (3), 109 (13),
112 (3 + 10 juvs.), 115 (i + 3 juvs.) ; FAL 390 (43) ; SCD 185 (i), 187 (15),
193 (4), 197 (13), 200 (14), 214 (3 juvs.).

DESCRIPTION. The abundant material now available has confirmed the suspicion
expressed by me (Day, 1960) that the two fragments described by Augener (1931)
do not belong to the same species. I suggest that the name L. meteorana should be
restricted to the posterior fragment with its characteristically long filamentous
parapodial lobes. A summarised description of a complete specimen from station
WCD 60 is given below.

Body slender, about 20 mm. long by 0-8 mm. broad with 120 segments. Yellowish
white in alcohol. Prostomium (Text-fig. 7 g) bluntly conical. Mandibles weak,
and poorly calcified with shafts in contact throughout their length. Maxillary
formula : I = i + i ; II = 3 + 3 ; III = ?i + ?i ; IV = i + i. The maxillary
supports (Text-fig. 7 h) are broadly triangular, the main fangs (Mx. I) are weakly

THE POLYCHAETE FAUNA OF SOUTH AFRICA 411

chitinised and blunt, each of the main dental plates (Mx. II) has three blunt bilobed
teeth, MX. Ill are cutting plates with a sharp edge but no definite tooth while MX.
IV are large and rounded, each with pale centre and a dark edge on which there is an
indistinct tooth.

, Anterior feet (Text-fig. 7 /) have rather flattened lobes, the low presetal one being
more developed superiorly while the larger postsetal one is compressed and bluntly
rounded. In the middle feet (Text-fig. 7 m) the two lobes are both small and sub-
equal. This condition persists over most of the body but near the posterior end
the two parapodial lobes (Text-fig. 7 n) elongate and become filamentous ; as always
the presetal one is slightly more superior than the postsetal.

The acicula are pale in all feet. Anterior setae include about 4 winged capillaries
and 2 compound hooks. The hooks are compound (Text-fig. 7 j) for the first 10-15
feet but then become simple with short hoods and unusually long teeth (see Text-fig.
7 k). One or two such hooks persist to the end of the body.

The capillaries decrease in number after the first few feet and disappear before
the middle of the body.

This species differs from L. bifilaris Ehlers in the structure of the maxillae and in
the possession of jointed hooks anteriorly. In both characters it is close to L.
albidentata but the posterior feet are quite different and the mandibular shafts are
in contact throughout, not widely divergent as in L. albidentata.

Lumbrineris aberrans sp. n.

(Text-fig. 8 a-f)

RECORDS. SCD 212 (i).

DESCRIPTION. The type locality is 33 58-8' S./25 42-2' E. in 26 metres. The
single specimen is 6 mm. long with 38 setigers but is incomplete posteriorly. The
anterior end however is quite characteristic.

The prostomium (Text-fig. 8 a] is enormously elongated, conical and about as
long as the combined length of the first two apodous segments plus the first
four setigers. The mandibles (Text-fig. 8 c) are small and lightly chitinised. The
maxillary supports (Text-fig. 8 6) are very long and triangular and the maxillary
formula is MX. I = i + i ; II = 3 + 3 ; III ?i + ?i (short plates with one
indistinct tooth) ; IV = o -f- o (oval plates without teeth).

The first two apodous segments are short and the first few setigerous segments that
follow have such small parapodial lobes that the setae appear to arise from the sides
of the segments. The first well formed parapodium is on setiger 7 and from there
on the parapodial lobes increase in size though they are never large. Within the
38 setigers investigated all the parapodia (Text-fig. 8 e) are essentially similar in
shape. Each has a low rounded presetal lobe and a small digitiform postsetal lobe
always shorter than the setae. There are 4-6 setae per parapodium. Capillaries
appear in the first setiger and simple hooks in the fourth. The first few capillaries
have very short broad blades but they become longer and more tapered posteriorly
(see Text-fig. 8 d) . The hooks (Text-fig. 8/) have short broad hoods and are distinctive
in having only two stout teeth instead of the usual series of one large and several
small ones. The acicula are pale.

4 I2 J. H. DAY

This species is obviously related to L. acuta Verrill, 1875 from Rhode Island,
U.S.A. and L. mucronata Ehlers, 1908 from the mouth of the Congo. The former
lacks hooks but the latter is closer for it also has bidentate hooks. However, in
L. mucronata the prostomium is relatively shorter, the maxillary supports narrower
and maxillary formula is different. Ehlers was only able to see the maxillae by
transparency in a cleared preparation but he figures MX. II with 6 teeth and MX. IV
with about 8 small ones. Both of these features are unusual and dissection may
show that Ehlers' interpretation was inaccurate. If so, L. aberrans may prove to be
a synonym of L. mucronata. Holotype : B.M.(N.H.), Reg. No. 1963.1.86.

Subfamily ARABELLINAE
Drilonereis falcata Moore, 1911
Drilonereis falcata Moore, 1911 : 298 ; Day, 1960 : 364, fig. 13 a-e.

RECORDS. One specimen obtained by a dredger prospecting for diamonds north
of the Orange River mouth in 5-10 fathoms.
NOTES. This is a new record for South West Africa.

Subfamily DORVILLEINAE
Protodorvillea egena (Ehlers, 1913)

Stauronereis egena Ehlers, 1913 : 501, pi. 35, figs. 1-6.

Non Stauronereis egena : Augener, 1918 : 377, pi. 5, figs. 102-103, fig- 45-

Dorvillea egena Day, 1960 : 371.

Protodorvillea egena Pettibone, 1961 : 180.

RECORDS. FAL 284 (i).

NOTES. The discovery of a specimen from the type locality (False Bay) allows me
to confirm Ehlers' description apart from minor points. The present specimen is
5-0 mm. long and thus slightly larger than the 4-5 mm. holotype. Only the second
large pair of eyes is visible, not the anterior minute pair. As stated by Ehlers the
short antennae consist of a single club-shaped joint, there is no dorsal cirrus on
setiger i but small ovoid dorsal cirri without cirrophores or internal acicula are present
on all subsequent feet. The presetal lobe projects well beyond the postsetal one.
The superior simple setae include 1-2 forked setae with subequal prongs and smooth
shafts and 1-2 capillaries which appear smooth under low power but are actually
faintly serrate at the base of the blade. The inferior compound setae number 3-4
per bundle. The blades vary in length, some being as long as that figured by Ehlers
(pi. 35, fig. 4) but others much shorter. They are all falcigerous and under high
power it may be seen that the tip is minutely bidentate.

Augener's specimens from South West Africa were considerably larger (9-16 mm.)
with multi-articulate antennae, the dorsal cirri are reported to be mounted on cirro-
phores and the forked setae are shown (Text-fig. 45 a] as having unequal prongs and
serrations on one side of the shaft-head.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

413

FIG. 8. Lumbrineris aberrans : (a) anterior end ; (b) maxillae; (c) mandibles ; (d) capillary
seta from 2oth foot ; (e) anterior parapodium ; (/) hook. Protodorvillea biarticulata : (g)
anterior end ; (h) superior capillary seta ; (;') forked seta ; (k) compound seta ; (/)
posterior view of parapodium. Scoloplella capensis : (m, m') thoracic neuroseta and details
of crenulations ; (ri) dorso-lateral view of head and first twelve setigers ; (p) 6th foot ;
(q) 24th foot.

414 J- H. DAY

Protodorvillea biarticulata sp. n.

(Text-fig. 8 g-l]

RECORDS. SCD 140 (2), 213 (3).

DESCRIPTION. The holotype is a complete specimen measuring 4-0 mm. by
0-3 mm. for 100 segments. It was obtained by grab from station SCD 140 at 34
35' S./2i 56' E. in 78 metres on a bottom of broken shell. The specimen is colour-
less in alcohol.

The two-ringed prostomium (Text-fig. 8 g) is smoothly ovoid anteriorly and has a
single pair of posterior eyes. The palps are long and coiled with irregularly annulated
or wrinkled palpophores and rather long oval palpostyles. They reach back to
setiger 10. The antennae are a little shorter than the length of the prostomium
and are biarticulate with two long joints. The mandibles and maxillae were not
clearly seen but appear to be normal. There are two apodous segments behind the
prostomium.

The body is somewhat flattened and consists of 100 short segments each about 8
times broader than long. The parapodia (Text-fig. 8 /) are relatively large. They
are the same width as the segments and extend outwards for a distance equal to
half the body width. Each one, including the first is provided with a small ovoid
dorsal cirrus without a cirrophore or internal aciculum. The setigerous lobe has a
blunt projecting presetal lobe longer than the dorsal cirrus, a very short postsetal
lobe or lip and a ventral cirrus similar to the dorsal one.

There are 5-6 setae. Superiorly there is a forked seta and a simple capillary and
inferiorly there are 3-4 compound setae. The forked seta (Text-fig. 8 j) has a
smooth shaft-head and subequal prongs which are flanged but not serrated on their
inner margins. The capillary seta (Text-fig. 8 h) has a narrow blade serrated
basally, while the compound setae (Text-fig. 8 k] are falcigerous with shaft-heads
which are minutely serrated on one side and blades which vary in length and end in
minutely bidentate tips.

The pygidium bears two pairs of long cylindrical anal cirri.

Pettibone (1961) has recently revised the Dorvilleidae in which she recognises
five genera. Her new genus Protodorvillea resembles Ophryotrocha in having unira-
mous parapodia and lacking dorsal cirrophores but may be distinguished from it
by the characters of the jaws, and by the possession of longer jointed palps and forked
setae. According to her species key the present species is most closely related to
P. kefersteini (Mclntosh) and P. gracilis (Hartman) and may be distinguished from
P. egena (Ehlers) by the presence of dorsal cirri on the first setiger and (one might add)
the possession of two-jointed antennae. P. gracilis (Hartman, 1938) from California
has single-jointed antennae and the pre- and postsetal lobes of the parapodia are
very short and subequal. P. kefersteini (Mclntosh, 1869) from Europe as re-
described by Fauvel, 1923 is a larger species up to 15 mm. long with possibly 4
indistinct annulations on the antennae and broad, subequal pre- and postsetal lobes
to the parapodia. Holotype : B.M.(N.H.), Reg. No. 1963.1.87.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 415

Ophryotrocha puerilis Clap. & Mecz., 1869

Ophryotrocha puerilis Claparede & Mecznikow, 1869 : 184 ; Fauvel, 1923 : 450, fig. 180 a-h ;
Hartman, 1944 : 191, pi. 15, figs. 325-330.

RECORDS. Specimens numerous in aquaria in Cape Town in June and August,
1962.

NOTES. The specimens are quite typical. Dissection of the maxillae showed that
there are 7 pairs of small dentate elements attached to a pair of chitinous strands
leading back to the main fangs. The 3 proximal pairs are dark and have about 5
teeth of decreasing size and the 4 distal pairs are pale and have about 12 minute
subequal teeth. These are used in the same way as a radula to scrape off attached
diatoms which fill the gut. Specimens were found breeding in August. This is the
first record of this cosmopolitan species from South Africa.

Family ORBINIIDAE

Subfamily PROTOARCIINAE

Scoloplella gen. n.

Small Oribiniidae belonging to the sub-family Protoariciinae with two achaetous
segments following the prostomium. Thoracic region with few segments. Thoracic
setae are crenulate capillaries in both rami. Abdominal region with both notopodial
and neuropodial lobes bearing capillary seta. No specialised setae. Branchiae
restricted to the abdominal segments.

TYPE SPECIES. Scoloplella capensis Day.

Scoloplella capensis sp. n.

(Text-fig. 8 m-q)

RECORDS. WCD 79 (2), 116 (3) ; SCD 237 (2).

DESCRIPTION. The type material comes from station WCD 116 at 33 06-5' S./
17 32-9' E. in 183 metres on a bottom of dark green mud. There are two incomplete
specimens. The holotype measures 15 mm. by 0-3 mm. for 48 segments while the
paratype measures 8 mm. for 25 segments. All are rounded in section and quite
pale in alcohol.

The prostomium (Text-fig. 8 n) is bluntly conical and lacks eyes. The proboscis
is retracted so that it is not possible to say whether its distal end is lobed or smooth.

The body is roughly cylindrical in section without any obvious flattening of the
anterior region. The first 9 segments are about three times as broad as long but
later ones rapidly increase in length ; the 20th setiger is about as long as broad, and
the 3oth is 4 times longer than broad.

The first two segments behind the mouth lack parapodia or setae. They are
well marked and there is no possibility that this is really a single biannulate segment.

Setigers 1-7 (segments 3-9) bear two bundles of crenulate capillaries arising
directly from the sides of the body (see Text-fig. 8 p). Minute postsetal papillae

4 i6 J. H. DAY

are present behind the notosetae and neurosetae from setiger 4 to 7 and on one of
the paratypes the notopodial papilla may be traced forwards to setiger i. Thus
the thorax consists of 2 apodous segments and 7 setigers.

The change from thorax to abdomen is marked by the appearance of a bilobed
neuropodial projection on setiger 8 (segment 10), the movement of the neuropodium
from a lateral to a dorso -lateral position and the gradual lengthening of the segments.
The abdominal parapodia (Text-fig. 8 q) are very simple. The notopodium consists
of a small digitiform postsetal lobe and the neuropodium a stout bilobed projection
from which the setae arise. The first pair of gills were only found on setiger 20 of
the holotype and on setiger 24 on one of the paratypes. A conical papilla represent-
ing the intermediate cirrus appeared two segments later. There was no sign of a
ventral cirrus. Thus the whole of the thoracic region and the anterior abdominal
segments lack gills.

All the setae are simple crenulate capillaries and no specialised setae were found.
The thoracic setae (Text-fig. 8 m, m'} both in the notopodium and the neuropodium
are more numerous than those of the abdomen and their blades are rather stouter
and the crenulations better marked. The abdominal notosetae are rather shorter
than those on the thorax but the neurosetae are twice as long. There are only 2-3
of them and they are very smooth. The crenulations if present at all must be very
poorly marked.

When first examined it was thought that these specimens belonged to the genus
Haploscoloplos, but the presence of two anterior achaetous segments shows that they
belonged to the new subfamily Protoariciinae erected by Hartman, 1957. This
contains three genera, Orbiniella, Protoaricia and Proscoloplos. Orbiniella lacks
gills and parapodial projections ; Protoaricia has uncini and subuluncini in some
thoracic neuropodia ; Proscoloplos has swan-shaped hooks in the abdominal neuro-
podia. Thus these South African specimens must be referred to a new genus for
which the name Scoloplella is proposed. Holotype : B.M.(N.H.), Reg. No. 1963.1.91.

Haploscoloplos cf . fragilis (Verrill, 1873)

Anthostoma fragile Verrill, 1873 : 598.

Hoploscoloplos fragilis : Hartman, 1957 : 2 7* pi- 2 5- fi g s - I ~3-

RECORDS. SCD 223 (2 + 1 juv.).

NOTES. The three specimens are incomplete and rather damaged. The thorax
consists of 15-16 setigerous segments sharply marked off from the abdomen by the
change in the neuropodia. The setae are crenulate capillaries in both rami through-
out the body, there being no hooks in thoracic neuropodia. In the thorax all the
notopodia have well developed postsetal lobes which are somewhat flattened.
Anterior neuropodia have a similar postsetal lobe (or foot papilla) behind the middle
of the setal fascicle but on the last two thoracic segments there are 3 postsetal
lobes. These all continue into the abdomen, the superior one becoming a neuro-
podial lobe and the two inferior ones becoming ventral cirri. This condition persists
for 4 or 5 segments and then one papilla disappears leaving only a single small
cirrus below the bilobed neuropodium. The abdominal notopodium has a slender

THE POLYCHAETE FAUNA OF SOUTH AFRICA 417

tapered postsetal lobe. There is also an interramal cirrus which appears on the last
thoracic segment and gradually decreases in size on later abdominal segments.
Branchiae appear on setigers 14-15, i.e. the penultimate thoracic segment. They
are well developed and larger than the abdominal notopodia.

The possession of an interramal cirrus and ventral cirri immediately distinguish
these specimens from H. kerguelensis. They agree in the main with Hartman's
description of H. fragilis but differ in the number of thoracic setigers (setiger 15-16
vs. 17-23) and the origin of the gills (setiger 14-15 vs. 17-21 or 30). More material
is required to determine how variable these characters are.

Family SPIONIDAE
Polydora maculata sp. n.

(Text-fig. 9 a-d]

RECORDS. SCD 284 (7).

DESCRIPTION. The type material was dredged at station SCD 284 at 33 01' S./
27 55' E. in 7 metres. The specimens were extracted from burrows in an old Bullia
laevissima shell inhabited by a hermit crab. A complete specimen chosen as the
holotype measures 20 mm. by i-o mm. for 130 segments. The body is flesh pink
apart from the palps which are barred with black.

The prostomium (Text-fig. 9 a] is faintly bilobed anteriorly and continues back
as a median ridge reaching setiger 2. There are 4 eyes. Between the bases of the
palps the prostomial ridge bears two occipital tentacles, one behind the other. The
peristome forms a broad support for the prostomium and bears a pair of stout
barred palps, which if folded back, would reach setiger 10.

The first setiger is small and bears cirriform notopodial and neuropodial lobes.
Neurosetae are present but no notosetae. The notopodial lobes on setigers 2 to
4 are well developed and larger than those on setiger 6 or later segments. The
neuropodial lobes increase in size on the first three setigers, decrease on setiger 4
and from setiger 6 onwards become insignificant and more ventrally situated.
Setiger 5 as usual is swollen but bears no parapodial projections. Branchiae (Text-
fig. 9 b) start on setiger 7 and continue to the posterior end. They are broad strap-
like lobes quite separate from the notopodial lobes and just meet their fellows on
the mid-dorsal line. The pygidium is small and saucer-shaped.

As mentioned above, there are no notosetae on the first foot. Subsequent feet
have winged capillaries in the notopodia and there are no specialised posterior
notosetae. The anterior neurosetae are also winged capillaries. Bidentate hooded
hooks (Text-fig. 9 d) appear in the neuropodium of setiger 9 and as they barely
project above the surface they are difficult to see. At first there are about 8-10
in a row but the number decreases to 4 near the posterior end. The enlarged hooks
of setiger 5 are arranged in a curved row centred dorsally. There are about 6 of
them alternating with abraded winged capillaries similar to those in other feet.
Each hook (Text-fig. 9 c] has a stout shaft which curves and flattens at the tip.

4 i8 J. H. DAY

Worn hooks are plain and blunt but a young hook has a narrow marginal flange.
There are certainly no accessory teeth. Neither is there any separate tuft of
capillaries on setiger 5.

The two striking features of this species are the possession of two occipital tentacles
and the late appearance of the neuropodial hooks on setiger 9. Apart from these
peculiarities there is some resemblance to P. antennata where the hooks appear on
setiger 8 and there is one occipital tentacle. However the gills of P. antennata do
not persist to the posterior end and the special hooks of setiger 5 are in a double,
horseshoe-shaped row. There are also some resemblances to P. kempi Southern
but again P. kempi has fewer gills and fang-like setae among the enlarged hooks of
setiger 5. Holotype : B.M.(N.H.), Reg. No. 1963.1.94.

Prionospio malmgreni Claparede, 1870
Prionospio Malmgreni Claparede, 1870 : 73 ; Fauvel, 1927 : 61, fig. 21 a-e.

RECORDS. WCD 82 (i), 97 (7) ; SCD 214 (2 + 1 juv.).

NOTES. These specimens agree in detail with Fauvel's description. There are
2 pairs of eyes of which the posterior, dorsal pair are the larger. There are 4 pairs
of gills on setigers 2 to 5 of which the 2nd and 3rd are short and smooth while the
ist and 4th are both longer and pennate. There is a well marked membraneous
ridge across the dorsum of setiger 7 and there are no genital pockets between middle
parapodia. Neuropodial hooks start on setiger n and notopodial hooks on setiger

36.

There has been considerable doubt about the presence of this species at the Cape
but as the above characters show, the present record is quite definite.

Prionospio steenstrupi Malmgren, 1867

Prionospio Steenstrupi Malmgren, 1867 : 202 ; Fauvel 1927, p. 60, fig. 21 f-i.
P. malmgreni var. dubia Day, 1961 : 489, fig. 3 j-n.

RECORDS. WCD 73 (i), 82 (i), 86 (i), 97 (i), 109 (35 + 8 juvs.), 112 (20 + 5 juvs.),
116 (2 juvs.), 118 (68 + 4 juvs.), 122 (19) ; FAL 390 (i) ; SCD 154 (i), 185 (i),
187 (3), 200 (3), 223 (4), 228 (i -f- 3 juvs.), 232 (6), 236 (2), 275 (4).

NOTES. The additional specimens show that the eyes which are indistinct in
juveniles are invisible in adults. The fourth pair of gills are pennate and in South
African specimens always shorter than the first pair. These features as well as the
later appearance of neuropodial hooks and the lack of a membraneous ridge across
setiger 7 distinguish this species from P. malmgreni.

Prionospio sp.

RECORDS. SCD 223 (2 juvs.).

NOTES. The larger of the two specimens measured 4 mm. and is obviously a
juvenile. However it certainly does not belong to any species recorded from South
Africa.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

419

e f g

FIG. 9. Polydora maculata: (a) anterior end ; (b) nth foot ; (c) enlarged hook from 5th
foot ; (d) neuropodial hook from nth foot. Aricidea curviseta : (e) notopodial capillary
from posterior foot ; (/) lateral view of posterior neuroseta ; (g) plan view of part of
posterior neuroseta ; (h) anterior end ; (j) 2oth foot and gill ; (k) posterior foot.
Cirrophorus branchiatiis : (1) notopodial spine ; (m) anterior end ; (ri) i6th foot ;
(0) looth foot.

420 J. H. DAY

The prostomium is oval or shovel-shaped anteriorly and produced back as a ridge
which reaches setiger 2. There are 2 pairs of eyes. The peristome lacks winglike
expansions and the first setiger is small and without gills. Very long slender
branchiae are present on setigers 2-10. Each gill is a smooth, cylindrical filament
about 4 times as long as the segmental breadth though the last few are shorter.
The postsetal lobes of the notopodia are never conspicuous ; in the branchial region
they are shorter than the notosetae and further back they are minute papillae no
larger than the neuropodial lobes. Dorsal crests were not seen. Neuropodial hooks
are certainly present on setiger 15 but possibly appear earlier.

These juveniles do not belong either to P. cirri/era or to P. cirrobranchiata the
only species with approximately the same number of smooth gills. They might be
juveniles of P. polybranchia Fauvel, 1929 which occurs in the tropical Indian Ocean
and has over 40 pairs of gills in the adult.

Family PARAONIDAE

REMARKS ON THE FAMILY. Until recently only three species had been reported
from South Africa, namely Cirrophoms branchiatus Ehlers, Paraonis gracilis (Tauber)
and Paraonis (Paraonides) lyra capensis Day. Kirkegaard (1959) reported Aricidea
fauveli Hartman and Day, 1961 added two new species namely Aricidea capensis
and Aricidea longobranchiata. More recent collections include most of these and
several new records. These are Aricidea (Aedicira) belgicae (Fauvel), Aricidea
curviseta sp. n., Aricidea jeffreysii (Mclntosh), Aricidea suecica simplex subsp. n.,
Paraonis gracilis oculata Hartman and Paraonis (Paraonides} lyra lyra Southern.

The identification of this material necessitated a review of generic characters.
Useful reviews of the family are to be found in Cerruti (1909) and Hartman (1957).

The shape of the head is reasonably constant but there is some doubt about the
buccal segment. According to Hartman " the first or second visible segment has
biramous parapodia ". Careful examination of the head shows that in Aricidea the
mouth opens on the ventral surface of the head and there are faint grooves indicating
that the head is formed of a dorsal prostomium with the achaetous peristomium
reduced and fused to its ventral surface. A complete and well preserved specimen
of Cirrophorus branchiatus Ehlers shows that the peristomium is not always so
reduced. In this species a dorsal fillet of the peristome is visible between the pro-
stomium and first setiger and the ventral part of the peristome is separated by a
groove from the prostomium. In Paraonis the mouth is further back and the ventral
surface of the first setiger forms the posterior lip. In brief, the first or buccal seg-
ment is reduced, always achaetous and fused to the ventral surface of the prosto-
mium ; the second segment is normally developed and bears the first pair of
parapodia.

Hartman (1957) recognised two genera based on the presence or absence of a dorsal
antenna and several subgenera based on the distribution of specialised setae. These
characters are clear-cut and easy to observe and as the whole family includes some
fifty species it is suggested that the subgenera be given full generic rank. The
following key covers the species recorded from South Africa and one or two others
marked with an asterisk which may be found in this area at a later date.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 421

KEY TO GENERA AND SPECIES OF PARAONIDAE FROM SOUTH AFRICA

1. Prostomium with a median dorsal antenna ....... 2

Prostomium without an antenna ...... 10

2. Specialised setae appear among the capillaries of posterior neuropodia (genus Ari-

cidea) ...... ...... 3

Specialised setae appear among the capillaries of posterior notopodia (genus Cirro-

phorus) ......... Cirrophorus branchiatus

No specialised setae among the capillaries either in the notopodia or neuropodia

(genus Aedicira) ^ . . Aedicira belgicae

3. Specialised posterior neurosetae numerous with a stout shaft abruptly tapered to a

slender blade ........... 4

Specialised posterior neurosetae as a few (5 or less) sigmoid hooks often with a hood 5

4. Specialised neurosetae with an incomplete joint at the junction of shaft and slender

tip ........... Aricidea fragilis*

Specialised neurosetae with a kink but no joint at the junction of shaft and blade

Aricidea curviseta

5. Specialised neurosetae are sigmoid acicular hooks sometimes with a terminal filament

but no hood ............ 6

Specialised neurosetae are hooks with a short or tapering hood over the apex . . 7

6. Acicular hooks with a terminal filament in middle segments but quite plain posteriorly

Eyes present ........ Aricidea suecica suecica*

Acicular hooks without a terminal filament even in middle segments. No eyes

Aricidea suecica simplex

7. Specialised neurosetae with a rounded spioniform hood covering the blunt apex

Aricidea Jeffrey sii
Specialised neurosetae with a delicate pointed hood or arista near the apex . . 8

8. Pointed hood or arista on the convex side of the apex. Posterior branchiae longer and

stouter than middle ones Aricidea fauveli

Pointed hood or arista on the concave side of the apex ...... 9

9. Last few branchiae with swollen bases and greatly elongated tips

Aricidea longobranchiata
Last few branchiae decrease in size ...... Aricidea capensis

10. Specialised setae among the capillaries of posterior notopodia (genus Paraonides) . n
Specialised setae among the capillaries of posterior neuropodia (genus Paraonis} . 13

11. Modified notosetae are short, stout, broad -winged capillaries . Paraonides neapolitana*
Modified notsetae are forked setae . . . . . . . . .12

12. Postsetal lobe of anterior notopodia one-third the length of the branchia. Prostomium

as broad as long ......... Paraonidea lyra lyra

Postsetal lobe of anterior notopodia are minute, inconspicuous papillae. Prostomium

longer than broad Paraonides lyra capensis

13. Specialised neurosetae have a blunt, curved apex with a ventral guard. Branchiae

foliaceous Paraonis fulgens*

Specialised neurosetae are unidentate acicular hooks without hoods. Branchiae cirri-
form .............. 14

14. Prostomium with eyes Paraonis gracilis oculata

Prostomium without eyes Paraonis gracilis gracilis

422 J. H. DAY

Aricidea curviseta sp. n.

(Text-fig. 9 e-k)

SCD 197 (i), 237 (3)-

DESCRIPTION. The type material comes from station SCD 237. The holotype is
16 mm. long by 0-7 mm. across the branchial region but the posterior region is more
slender and rounded. There are 83 segments but the tail end is missing.

The prostomium (Text-fig. 9 h) overlies the small achaetous peristomium and is
fused to it though lateral grooves remain. The whole forms a bluntly triangular
head about as broad as long. There are no eyes and the median antenna does not reach
the tip of the prostomium. The first few setigers are narrowed but in the branchial
region each segment is about four times as broad as long. Branchiae appear on
segment 4 and persist to segment 40. In other specimens as many as 40 pairs of
gills have been observed. Each gill (Text-fig. 9;) is a stout tapered projection which
just meets its fellow in the mid-dorsal line. The last few decrease gradually in
size and there is no posterior groups of elongate branchiae. The postsetal lobe of
the notopodium is evident from the first segment and is at first rather short and
conical. It soon becomes more slender and in the branchial region it is a little
less than a third the length of the branchia. There is a small postsetal neuropodial
papilla behind the slight lateral swelling on which the seta arise in the branchial
region.

In the posterior region the segments are longer, hardly twice as broad as long.
The postsetal lobe of the notopodium (Text-fig. 9 k] is a slender filament no longer
than that in the branchial region and only half the length of the notosetae. There
is no neuropodial projection.

There are two types of setae. In the branchial region both the notosetae and the
neurosetae are tapered capillaries without obvious blades or marked curvature.
In the posterior notopodia the capillaries (Text-fig. 9 e) are more slender and the
number of setae is reduced. Posterior neurosetae (Text-fig. 9 / and g) however are
quite different. All neurosetae are similar throughout the posterior region and
each consists of a stout shaft which suddenly narrows to a long slender tip. When
seen in surface view (Text-fig. 9 g) the sudden narrowing is the only obvious feature
but when seen in lateral view (Text-fig. 9 /) it is evident that there is an abrupt
curve or kink at the origin of the slender tip. There is no hood or projecting tooth
at the point of curvature but the kink forms a slight step and the tip may break off
leaving only the stout shaft.

Possibly this species is most closely related to A . uschakovi Zachs, 1925 from Arctic
seas or possibly to A. fragilis Webster, 1879 from Sweden. In the former each
posterior neuroseta has a fine hairlike arista at the end of the stout sigmoid shaft,
which might be a further development of the abrupt decrease in thickness which
occurs in A . curviseta.

In A. fragilis as described by Hartman (1957) there is an incomplete joint or
pseudoarticulation at the junction of the stout shaft and the slender blade. There
is no sign of such an articulation in A. curviseta. Holotype : B.M.(N.H-), Reg. No.
1963.1.106.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 423

Aricidea Jeffrey sii (Mclntosh, 1879) sensu Cerruti, 1909

Scolecolepis Jeffrey sii Mclntosh, 1879 : 506, pi. 65, figs. 13, 14.

Aricidea Jeffrey sii : Cerruti, 1909 : 409, pi. 18, figs. 1-6, 9-18, 22-26 ; Fauvel, 1927 : 75, fig. 25
a-e.

RECORDS. FAL 397 (i).

NOTES. The single specimen is incomplete, measuring 9 mm. by 0-4 mm. for
70 segments. The prostomium is bluntly rounded anteriorly and lacks eye-spots.
The median antenna is cylindrical and just projects beyond the anterior margin of
the prostomium. Branchiae start on setiger 4 and extend over 18 segments. Each
is bluntly cylindrical and its length equals half the width of the segment. The
postsetal lobes of anterior segments are tapered and half as long as the gills ; in
posterior segments they remain the same length but become very slender. Most of
the setae are capillaries which are curved anteriorly but fine and straight posteriorly.
The specialised neuropodial setae are 4-5 hooks which appear in the postbranchial
region. The tip curves to a blunt apex which has a rounded spioniform hood.

This is a new record for South Africa but the above description agrees with the
account given by Fauvel with three exceptions. The South African form lacks
eye-spots, has a longer antennae and considerably shorter gills than shown by
Fauvel's fig. 25 a.

Hartman (1957) reported that Mclntosh's description of the holotype from Green-
land is incomplete and that the description given by Cerruti (1909) and Fauvel (1927)
may refer to a different species.

Cirrophorus branchiatus Ehlers, 1908
(Text-fig. 9 l-o)

Cirrophorus branchiatus Ehlers, 1908 : 124, pi. 17, figs. 5-9.
Aricidea (Cirrophorus} branchiata : Hartman, 1957 : 323.

RECORDS. WCD 82 (i).

NOTES. Ehlers' holotype from the Agulhas bank was incomplete and the present
complete and well preserved specimen allows me to amend the original description
of this rare species.

The complete worm measures 25 mm. by 0-4 mm. for 120 segments. The body is
somewhat flattened with short anterior segments about 3-4 times as broad as long
and longer posterior segments less than twice as broad as long. The intersegmental
constrictions are very marked posteriorly so that this region appears almost moni-
liform.

The prostomium (Text-fig. 9 m) is a blunt depressed cone without eyes but has
well-marked nuchal slits. The median antenna is short and stout, less than half
the length of the prostomium. Ehlers, p. 17, fig. 6, shows the first segment as bearing
setae but in fact there is a short peristomial segment between the prostomium and
the first setiger. It is just visible dorsally but extends forward below the prosto-
mium ventrally.

424 J. H. DAY

In the anterior region the postsetal lobes of the notopodia (Text-fig. 9 n) are well
developed but in the middle of the body they are reduced until they are no longer
than the notopodial spines and then in the posterior region (Text-fig. 9 o) they become
very long and slender. At the very end of the body there are 4 pairs of cirriform
projections. One pair certainly arises from the pygidium but the others may be
the postsetal lobes of developing segments. The neuropodia are mere lateral swel-
lings throughout the whole length of the body.

Branchiae appear on setiger 5 and continue for the next 23 segments. Each is a
cylindro-conical organ arching forward over the back and its length is equal to
2/3 the segmental breadth. The last 2-3 pairs are shorter.

Anterior setae up to segment 12 are all fine capillaries without a trace of a blade.
From setiger 13 however, 1-2 heavy spines appear in the notopodia and the noto-
podial capillaries are reduced in number from over 20 in anterior segments to about
6 in posterior ones. Each spine (Text-fig. 9 I) is straight and pointed with a very
fine curved filament arising some distance below the pointed apex. These filaments
were not noted by Ehlers and are easily mistaken for developing capillaries but the
examination of several parapodia shows that they occur on all spines that are
unbroken. No specialised setae appear in the neuropodia but the capillaries in-
crease in length until they are longer than the segmental breadth and twice as long
as the notosetae.

Aedicira belgicae (Fauvel, 1936)

Paraonis Belgicae Fauvel, 1936 : 29.

Aricidea belgicae : Monro, 1939 : 127, fig. 16 a-b.

Aricidea (Aedicira) belgicae : Hartman, 1957 : 327.

RECORDS. WCD 109 (3).

NOTES. Aedicira was erected by Hartman (1957) as a subgenus of Aricidea
which lacked specialised neurosetae in posterior parapodia. The type species is
Aricidea (Aedicira) paciftca Hartman, 1944. As explained earlier, Aedicira is here
recognised as a full genus characterised by the possession of a median antenna and
the lack of specialised setae either in the notopodia or neuropodia.

The three specimens are all incomplete. The longest is 18 mm. by 0-9 mm. for 84
segments. The prostomium which is broadly rounded anteriorly agrees very well
with Monro 's figure. The peristome is not visible dorsally but may be distinguished
ventrally where it is largely fused to the prostomium.

The anterior region is rather flattened with segments 6 times as broad as long ;
the posterior region however is more rounded with segments only twice as broad as
long. Branchiae appear on setiger 4 and extend over 24 segments. An average gill
is a stout tapered organ whose length is equal to 2/3 the breadth of the segment.
The last few gills however are markedly swollen basally and then taper abruptly
to short filamentous tips. The postsetal lobe of the notopodium is well developed
and tapered and on branchial segments it is half the length of the gill. Posteriorly
it becomes very slender but remains the same length.

The capillary setae of both rami in the anterior region are numerous and curved.
In the postbranchial region the number of setae decrease particularly in the noto-

THE POLYCHAETE FAUNA OF SOUTH AFRICA 425

podium where only a few fine capillaries remain. The neurosetae are also fine and
about as long as the segment is broad. There are no modified setae in either ramus.
Aedicira belgicae is an Antarctic species and has not been recorded from South
Africa before. The genus is characterised by the lack of specialised setae and as the
specimen is broken at segment 84 some doubt as to the identification must remain.

Paraonis gracilis gracilis (Tauber, 1879)

Aonides gracilis Tauber, 1879: 115.

Paraonis gracilis : Hartman, 1957 : 330, pi. 44, figs. 4-5.

Aonides gracilis : Ehlers, 1913, p. 512.

RECORDS. SCD 223 (7).

NOTES. The only other record of this species from South Africa is Ehlers' doubt-
ful record from False Bay (" Simonstown "). In Ehlers' brief description there is
no mention of eyes, a character which separates P. gracilis gracilis from P. gracilis
oculata Hartman, 1957, a subspecies which has recently been recorded by me (Day,
1963) from 1 200 metres west of Cape Town.

The present specimens agree in detail with the stem form. They lack eyes and
there are about 12 pairs of branchiae starting on setiger 7-9. The postsetal noto-
podial lobes are small throughout and the specialised neurosetae of the posterior
region are 2-3 stout sigmoid hooks with blunt unidentate tips and no trace of a
hood or arista.

This species seems to be world wide in distribution.

Paraonides lyra lyra (Southern, 1914)

Paraonis (Paraonides) lyra Southern, 1914 : 94, pis. 9-10, fig. 22 a-c ; Fauvel, 1927 : 72, fig. 24
a-f.

RECORDS. WCD 112 (common).

NOTES. Paraonides lyra capensis (Day, 1955) was reported by me as Paraonis
lyra var. capensis from the Knysna estuary. When the present specimens were
compared with paratypes of P. I. capensis the following differences were observed.

In P. I. capensis the prostomium is obviously longer than broad, the postsetal
lobes are always minute or invisible and the setae are shorter than the width of
the body. In P. I. lyra the prostomium is about as broad as long, the postsetal
lobes of the notopodia are well marked in the branchial region and in the posterior
region the neuropodial capillaries are much longer than the segmental width.
Southern has suggested that the latter character may be a sexual feature.

? Paraonides sp.

RECORDS. WCD 92 (2).

NOTES. Two minute threadlike worms 4 mm. long by o-i mm. but broken after
the 46th setiger. At first sight there appear to be no projections of any sort.

426 J. H. DAY

Examination under high power shows an ovoid prostomium without an antenna
but with a pair of minute eyes, a reduced achaetous peristome quite separate from
the prostomium and behind this numerous biramous segments. The notopodia
bear minute postsetal papillae which are easily overlooked and there are no gills.
There are no neuropoolial projections. There are slender capillaries in both rami
throughout the body and in addition there are 1-2 short forked setae in the notopodia
from setiger 2. These are similar to the forked setae of Paraonides lyra with the
longer ramus bearing spinules on its inner margin.

The absence of gills is peculiar. It is possible however that they are juveniles
and that gills develop later.

Family GOSSURIDAE nov.

REMARKS. The family Cirratulidae to which the genus Cossura Webster &
Benedict, 1887 has hitherto been referred, has been a dumping ground for a wide
diversity of genera. Fauvel (1927) recognised 12 European genera divided between
two subfamilies. Hartman (1959) agreed with earlier workers in transferring
Streblospio to the Spionidae and recognised Ctenodrilidae Kennel, 1882 as distinct
from Cirratulidae Carus, 1862. Within the Cirratulidae she recognised 14 valid
genera. Among these are Acrocirrus, Ledon and Macrochaeta all of which have
prostomial projections and compound setae and Caulleriella, Chaetozone, Cirratulis-
pio, Cirratulus, Cirriformia, Cossura, Dodecaceria, Pentacirrus, Pseudocirratulus,
Tharyx and Timarete which lack prostomial projections and have no compound
setae.

The genus Cossura is now known through 5 species. It agrees with the typical
cirratulids in lacking prostomial projections and projecting parapodia and in posses-
sing capillary setae with flattened blades. However it differs from typical cirratulid
genera in three important features. It lacks grooved food-gathering peristomial
palps or the homologous grooved tentacles inserted on anterior segments. It lacks
paired cylindrical gills above the notosetae and it possesses 3 long anal cirri. The
only projection is a single very long median gill which arises from the second or
third setigerous segment. Some workers have referred to this as a " tentacle "
but it is quite different in structure from the grooved tentacles of the typical
cirratulids.

There is no more reason to include Cossura among the Cirratulidae than there is to
include Paraonis which does possess gills. It is suggested in fact that Cossura should
be placed in a new family, Cossuridae, denned as follows :

Small threadlike worms with numerous similar segments. Prostomium conical.
Pharynx eversible, soft and unarmed. No head appendages either on the prosto-
ium or peristomium. No parapodial lobes on any segment. The first 1-2 segments
behind the peristomium achaetous but subsequent segments with one or two bundles
of simple capillary setae which usually have flattened blades with spinulose margins.
A single very long branchial filament arises from the middle of the dorsum of setiger
2 or 3. Pygidium often with 3 long anal cirri.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 427

Cossura coasta Kitamori, 1960

Cossura coasta Kitamori, 1960 : 1082, fig. i a-f.

RECORDS. WCD 53 (i juv.), 79 (i), 106 (12) ; SCD 228 (3), 232 (i).

NOTES. The prostomium is a blunt cone a little longer than broad. There are
no eyes nor head appendages. Behind the prostomium there are two apodous rings
with the ventral mouth opening between them. Each of the rings is as long as one
of the setigerous segments. They are regarded here as the peristome and the second
segment but it is possible that the first ring is part of a biannulated prostomium or
that the two rings together form part of a biannulate peristome. The proboscis
which was everted on one specimen has a lobed margin.

Behind the second apodous ring there are 40 cylindrical setigerous segments in the
longest fragment which measures 7 mm ; none of the specimens was complete.

There are no parapodial projections and the setae arise directly from the sides of the
body. In the posterior segments the setae obviously arise in two fans, so that these
segments are clearly biramous. Further forward however the two bundles approach
one another very closely, until those of the first setiger form a single bundle. The
second and subsequent setigers seem to have two bundles but some doubt must
remain. In each ramus the setae are arranged in two rows ; an anterior row of
shorter setae and a posterior row of longer ones, almost twice the length of the anterior
row. All setae are capillaries with rather flattened blades having a spinulose or
hispid margin.

A single very long slender gill (the " tentacle " of other workers) arises from the
anterior margin of setiger 3 in the mid-dorsal line. In contracted specimens it is
not very clear whether this gill actually arises from setiger 3 or setiger 2.

A discussion of the species of Cossura is given by Reish (1958). The specific
differences are based on the number of apodous segments behind the prostomium,
the number of uniramous setigerous segments and the position of the gill. As shown
above these differences are somewhat slender evidence for specific identification
but the characters of the present specimens fit the description of C. coasta Kitamori
from Japan better than others. C. coasta however is very close to C. Candida
Hartman (1955) from California.

As far as I am aware this is the first record of Cossura from the Southern hemis-
sphere.

Family GIRRATULIDAE
Tharyx marioni (St. Joseph, 1894)

Heterocirrus Marioni Saint- Joseph, 1894 : 56.

Tharyx marioni : Fauvel, 1927 : 100, fig. 35 a-b ; Day, 1961 : 503 with synonymy.

RECORDS. One specimen obtained by a dredger prospecting for diamonds north
of the Orange River mouth in 5-10 fathoms.
NOTES. This is a new record for South West Africa.

428 J. H. DAY

Family OPHELIIDAE
Ophelia cf. roscoffensis Augener, 1910

Ophelia roscoffensis Augener, 1910 : 237.

RECORDS. SCD 247 (i + i juv.).

NOTES. The larger of the two specimens is 18 mm. long. It has the usual
fusiform shape with 8 prebranchial, 20-21 branchial and 3-4 posterior abranchiate
segments, there being one more gill on one side. Thus the total is 32 setigers.
There is no achaetous segment before the pygidium. Branchial fenestrations are
absent.

Crimped dorso-lateral ridges extend from setiger 29 to the pygidium which bears
an arc of 12 small cirri dorsally and two very stout conical cirri ventrally. The
neurosetae are half the length of the notosetae and the gills taper evenly.

0. roscoffensis was redescribed by Tebble (1952 : 565, fig. 5) from a large (51 mm.)
specimen obtained off Plymouth ; it has 8 abranchiate, 23 branchiate and i posterior
abranchiate setiger making a total of 32 setigers as in the present specimens. How-
ever it has 2-3 more branchial segments and correspondingly fewer posterior
abranchiate setigers. It may be that additional branchiae appear on larger indivi-
duals. It also agrees in the lack of branchial fenestrations and in the details of
the pygidium. Ophelia praetiosa (Kinberg) from Patagonia as described by Hartman
(1948 : 115) is also close with 8 abranchiate, 18 branchiate and 5 postbranchial
setigers making a total of 31 setigers. This species also lacks branchial fenestrations
but has 19 small anal cirri as well as the large ventral pair.

More material is required to determine the variability in numbers of branchiae.

Family SCALIBREGMIDAE
Asclerocheilus capensis sp. n.

(Text-fig. 10 a-/)

RECORDS. SCD 236 (i).

DESCRIPTION. The type locality is 34 51' S./23 41' E. in 183 metres. The
holotype is a single incomplete specimen 5 mm. long by i-o mm. broad with 16
setigers and probably represents half of the complete worm.

The prostomium (Text-fig. 10 a) is quite pale without any trace of eye-spots and
bears a pair of short stout lateral lobes. The peristome is achaetous, narrow dorsally
and incomplete ventrally so that the lower lip appears to be formed by setiger i.

The body is oval in section and somewhat depressed with the first 10 to 12 setigers
broader than those which follow. The first three setigers are not clearly annulated
but from the 6th onward all segments are quadriannulate and tessellated dorsally.
The ventral marks are not so regular or clearly defined. There are no parapodial
lobes or gills and the setae arise from slight swellings on the body wall. Anterior
notosetae are dorsal in origin but later ones from about the I2th segment are lateral
like the neurosetae.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 429

Setigers i and 2 bear curved acicular setae and capillaries in both rami. There are
15 or more curved acicular setae in the notopodium of setiger i and 4-6 long capil-
laries, but in the neuropodium there are only a few acicular setae, many more capil-
laries, and the distinction between the two types of setae is not so marked. In
setiger 2 (Text-fig. 10 b) the notopodium has only about 6 acicular setae and an
equal number of capillaries. All subsequent segments (Text-fig. 10 c) lack acicular
setae and bear long fine capillaries and a few short forked setae in both rami.

The capillaries (Text-fig. 10 /) are fine and tapered without any ornamentation.
The acicular setae (Text-fig. 10 e) are fairly stout and taper abruptly near the end to
fine tips. The forked setae (Text-fig. 10 d) have two unequal prongs and the inner
margins of both are spinulose.

Two species of the genus Asclerocheilus have been described, namely A. inter-
medium (St. J.) with acicular setae in both rami of setigers i, 2 and 3, and A. berin-
gianus Uschakov, 1955 from the Bering Sea with acicular setae confined to the noto-
podia of setigers i and 2. A. capensis appears to be intermediate between the two,
but further work may show that all three are merely growth forms of A. intermedium.
Holotype : B.M.(N.H.), Reg. No. 1963.1.132.

Parasclerocheilus capensis Day, 1961

Parasclerocheilus capensis Day, 1961 : 517, fig. 10 c-f.

RECORDS. Collected by a dredger prospecting for diamonds in 10-20 metres
just north of the Orange River mouth.
NOTES. This is a new record for South West Africa.

Family MALDANIDAE
Euclymene sp.

(Text-fig. log)

RECORDS. SCD 187 (fragments).

NOTES. Although the material contains both anterior and posterior ends, the
number of setigers is unknown. Nevertheless the characters suggest that the species
is new to South Africa and thus worth recording.

The prostomium is bluntly triangular but lacks ocelli. The cephalic plate is
oval with a well developed rim which slopes back smoothly to a median posterior
notch. There are no lateral notches. The nuchal grooves are long and straight,
fully 4/5 the length of the cephalic plate.

Setigers 1-5 are short but 6 and 7 are longer. Setiger 8 is shorter than 7 but subse-
quent ones become 4-5 times longer than broad. Glandular bands are well marked
from setiger 3 to 6, and there is a faint midventral streak from setiger i onwards.
The posterior end bears two achaetous preanal segments in front of the well marked
pygidial ring and funnel. The two preanals together are just shorter than the last
setiger and the first is twice as long as the second. The funnel has 18-24 c i rr i which
alternate in length ; the ventral one is much longer than the rest. The anus is
sunk in the funnel and a ventral valve is lacking.

430 J. H. DAY

Notosetae include the usual winged capillaries with a few fine feathered forms in
posterior notopodia. The neurosetae of setigers 1-3 are represented by 2-3 acicular
spines (Text-fig. 10 g) with sharply bent ends bearing 3 vestigal denticles above the
rostrum but no trace of a tendon below. Later neurosetae are numerous well
developed hooks with a vertical series of 6-7 teeth above the main fang and strong
tendons below.

The only species of Euclymene recorded from South Africa with a smooth (non-
crenulate) cephalic margin and 2-3 acicular neurosetae with denticles is E. oerstedii.
Kirkegaard's (1959) record from the Cape is based on a fragmentary specimen and
thus doubtful. The present material however has only two achaetous preanals, no
ocelli and lacks lateral notches in the cephalic margin.

Euclymene cf. quadrilobata (Sars, 1865)

(Text-fig. 10 h-j)
Clymene quadrilobata Sars, 1865 : 15.

RECORDS. SCD 223 (3 fragmentary specimens).

NOTES. The specimens are all broken. The longest fragment has 13 setigers
and measures 45 mm. by i-o mm. The fragments obviously belong to a species new
to South Africa but the number of setigerous segments is unknown. The main
characters may be summarised as follows :

Anterior end truncate with a rounded cephalic plate (Text-fig. 10 h) almost at right
angles to the body. Prostomium small and rounded. Eye-spots present. Nuchal
grooves straight, very short, less than half the length of the cephalic plate. Cephalic
rim very low and thick with a pair of lateral depressions and a single median posterior
notch. Number of setigerous segments unknown. Glandular bands well marked
on setigers 3 to 7. A faint mid-ventral streak from setiger I onwards. Posterior end
with 4 achaetous preanals each half as long as the last setiger. Pygidial ring and
funnel well marked, the latter bearing 24 unequal cirri. Anus sunken in the funnel.
No ventral valve. Notosetae are winged capillaries. Neurosetae of setigers 1-3
represented by a single stout spine (Text-fig. 10 j) with faint indications of denticles
above the rostrum. Later neurosetae are rows of numerous hooks each with 4-5
well developed teeth above the main fang and a few fine tendons below.

The main characters are generally similar to those described by Arwidsson (1906)
for Pseudoclymene quadrilobata (Sars) and there are also resemblances to Clymene
(Isocirrus] wolft Kirkegaard, 1959 from the Gold Coast.

Praxillella cf. affinis (Sars, 1872)
(Text-fig. 10 k-l]

Clymene affinis Sars, 1872 : 412.

Clymene (Praxillella) affinis : Fauvel, 1927 : 180, fig. 62 /-/.

RECORDS. WCD 112 (2).

NOTES. The largest of the two specimens was complete and measured 23 mm.

FIG. 10. Asclerocheilus capensis : (a) anterior end ; (b) 2nd foot ; (c) zath foot ; (d)
forked seta ; (e) notopodial acicular seta from 2nd foot ; (/) capillary seta. Euclymene
cf . oerstedii : (g) neuropodial spine from 2nd foot Euclymene cf . quadrilobata : (h) head ;
(j) neuropodial spine from 2nd foot. Praxillella cf. affinis : (k) neuropodial spine of 2nd
foot ; (/) neuropodial hook from 5th foot. Rhodine gracilior : (m) dorsal view of anterior
end ; (n) lateral view of head ; (o) lateral view of posterior end ; (p) neuropodial hook.

432 J. H. DAY

by 0-4 mm. It is uniformly pale brown in alcohol. There are 18 setigerous seg-
ments, 3 achaetous preanals followed by the pygidium which has a circle of 12
subequal cirri and a protruding anal cone with a large ventral valve.

The prostomium is bluntly triangular without eyespots. The cephalic plate is
oval and the marginal rim has a pair of faint lateral notches and a better marked
posterior notch. The rim is high in front of the lateral notches and distinctly lower
behind them. The nuchal grooves are long and straight, fully 4/5 the length of the
cephalic plate. Setigers 1-8 are 2 to 3 times longer than broad, setigers 9-15 about
4 times as long as broad, but later ones gradually shorten until the 3 preanals are
together just equal to the length of the last two setigers. Setiger 1-3 have slight
anterior glandular bands, setigers 4-8 are markedly glandular and from setiger
8 onwards there is a narrow mid-ventral glandular streak.

Setiger 1-3 have 2-3 hooks in each neuropodium but each of the later segments has
a row of about 10. The hooks of the first three setigers (Text-fig. 10 k) have 3
well marked teeth above the rostrum and there is even a trace of a tendon below.
The hooks of later setigers (Text-fig. 10 /) are much better developed with a vertical
crest of about 5 denticles above the rostrum and well marked tendons below. The
notosetae include about 4 stout capillaries with narrow wings and 3 fine capillaries
without visible wings.

There are several differences between the South African form described above and
Fauvel's account of P. affinis. Eye-spots are absent, there is no enlarged anal cirrus,
the finer notosetae are not spinulose and above all the neurosetae of setigers 1-3
have much better developed denticles and even possess faint tendons, so that the
anterior neurosetae approach the condition found in the genus Axiothella. However,
these South African specimens are small and these may be juvenile characters.
Some doubt as to the identification must remain until larger specimens have been
found.

As shown earlier (Day, 1955), Praxilla capensis Mclntosh, 1885 is indeterminate
since the posterior end is missing. In any case. the neurosetae of setiger 1-3 are
smooth, straight and acicular, one per parapodium.

Lumbriclymene cylindricauda Sars, 1872

Lumbriclymene cylindricauda Sars, 1872 : 413 ; Arwidsson, 1906: 40, pi. i, figs. 15-24; pi. 7,
figs. 219-221.

RECORDS. SCD 237 (i juv.), 251 (i).

NOTES. The adult specimen is broken into two pieces which together measure
60 mm. by i mm. while the juvenile was complete with 19 setigers and measured
35 mm. There are 3 well developed achaetous preanals and then 3 indefinite ones
preceding the bluntly conical pygidium.

These specimens agree with Arwidsson's description with one or two minor excep-
tions. The colour markings appear to differ. There is a dorsal brown bar on the
buccal segment and similar bars in front of and behind each parapodium of the first
three setigers ; there is also a well marked midventral glandular streak stretching

THE POLYCHAETE FAUNA OF SOUTH AFRICA 433

from setiger n to setiger 14, and the neuropodial hooks on setiger 4 have 4 teeth
above the main fang, not 2-3. Later hooks also have 4, but there are minor lateral
teeth as well.

The juvenile was encased in a brittle tube composed of coarse sand and foramini-
feran shells.

The only other species of the genus recorded from South Africa is L. minor which
differs in having only 3 achaetous preanals and a dorsal anus.

Rhodine gracilior Tauber, 1879
(Text-fig. 10 m-p)

Rhodine loveni gracilior Tauber, 1879 : 123.

Rhodine gracilior : Arwidsson, 1906 : 74, pi. 2, figs. 53-58 ; pi. 7, figs. 237-241 ; pi. 8, figs. 242-

243 ; Wesenberg-Lund, 1949 : 344 ; Kirkegaard, 1959 : 65.
Rhodine cf. gracilior : Day, 1961 : 514.

RECORDS. WCD 86 (3), 112 (2), 118 (2), 121 (i).

NOTES. The additional material includes an anterior fragment with a head and
13 setigers (Text-fig. 10 m), a posterior fragment with 7 setigers, 10 achaetous pre-
anals and the pygidium and finally a juvenile with 21 setigers but without a head
or posterior end. It is estimated that an adult was about 50 mm. long by i mm.
wide with a minimum of 23 setigers.

The striking features include the very long head and first setiger (Text-fig. 10 n],
the dark red 4th setiger, the smooth-edged posterior collars on posterior segments,
(Text-fig. 10 o) the numerous achaetous preanals, the bluntly conical pygidium and
the backwardly bent and expanded heads of the hooks (Text-fig. 10 p] .

The tube is fragile and membraneous.

Family PECTINARIIDAE
Pectinaria capensis (Pallas, 1766)

Nereis cylindraria capensis Pallas, 1778 : 118.

Pectinaria capensis : Mclntosh, 1904 : 76, pi. 7, figs. 35-36 ; Day 1955, p. 432.

RECORDS. In 10 metres on sand, 20 miles north of Orange River mouth.

NOTES. This species has not been recorded from South West Africa before.
It was obtained by a suction dredger prospecting for diamonds. The specimens are
large (up to 120 mm. long) and unusual in that the tube is composed of sand grains
arranged in regular transverse rows. All Cape specimens examined have tubes
composed of short lengths of sponge spicules arranged like bonded bricks. Careful
comparison of the Cape and South West African forms showed no differences in
the animal but the description given by me (Day, 1955) needs amendment.

There may be n or as many as 15 paleae on each side, there is one, not two achae-
tous prescaphal segments and a slender anal cirrus is present on the ligule though
often broken off. It may be added that the scaphe is oval in outline, and concave
dorsally with 2 pairs of cirriform marginal papillae near the base and one pair distally
near the origin of the ligule.

434 J.H.DAY

Pectinaria (Lagis) koreni cirrata subsp. n.

(Text-fig, ii a)

RECORDS. NAD 27 (2) dredged on I3th July, 1959 off the Natal coast at
29 53-5' S./3i 06-5' E. in 71 metres on dark brown sand.

DESCRIPTION. The holotype is n mm. long by 2 mm. across the first setiger.
The tube is composed of coarse sand grains.

Operculum (Text-fig. 11 a) with 10 -f 10 paleae. Margin of the operculum smooth
and unusually high. Its ventral ends are fused to the bases of the first pair of
tentacular cirri. The veil which bears about 18 slender marginal cirri is also fused
to the bases of the tentacular cirri on each side so that the tentacular cirri mark the
junction of the veil and opercular margin, i.e. the veil is completely fused to the
operculum. First pair of tentacular cirri slender and longer than the paleae ;
second pair a little shorter. The semicircular ventral flange between the second
pair of tentacular cirri is incised to form 15-20 slender cirri. Two pairs of lamellate
gills. Ventrum of branchiferous segments and setiger 1-2 with three glandular
ventral pads.

Behind the branchiferous segments there are 3 segments with notosetae only, 12
segments with notosetae and uncini, and then 2 achaetous prescaphal segments
followed by the scaphe.

The scaphe is oval and the margin is incised to form a median basal lappet and
5 pairs of marginal lappets. Scaphal hooks were not seen but were probably re-
tracted. The anal ligule is tongue-shaped with a smooth margin and a small anal
cirrus.

Notosetae include a series of broad-bladed capillaries and a smaller series of
capillaries with bent and finely dentate tips. Uncini have 2 rows of 5-6 large teeth
and a close set group of numerous small teeth above the basal gouge.

The main characters are similar to those of P. (Lagis) koreni but the slender cirri
on the ventral flange between the second pair of tentacular cirri are quite different
from the 8-10 blunt lobes found on the same flange of P. koreni. Holotype :
B.M.(N.H.), Reg. No. 1963.1.138.

Family AMPHARETIDAE
Isolda pulchella Muller, 1858

Isolda pulchella F. Muller, 1858 : 219 ; Augener, 1918 : 517, pi. 7, fig. 229, text fig. 88.
Isolda warnebroensis Augener, 1914 : 82, pi. i, figs. 14-15, text-figs. 13 a-c.
Isolda sibogai Caullery, 1944 : 102, fig. 83.

MATERIAL EXAMINED. British Museum (Natural History) Reg. No. 76 : 10 : 4 : 13
from Brasil /. pulchella (Miiller det.) and Reg. No. 1933 : 3 : 18 : 67-69 from Burma,
/. warnebroensis (Monro det.).

S. A. RECORD. SCD 187 (2+1 juv.).

A comparison of the specimen of I. pulchella from Brasil, of /. warnebroensis from
Burma and the large fresh specimens from South Africa shows that they are all

THE POLYCHAETE FAUNA OF SOUTH AFRICA 435

identical. Since there is no modern description of /. pulchella and some doubt as
to its relations with other species, a short description is given below and the various
species discussed.

Body up to 45 mm. long with 60 segments. Colour brown, buccal cirri speckled
when fresh. Prostomium snoutlike, rounded anteriorly and pinched in at the sides.
No eyes nor glandular ridges. Buccal cirri short, stout, without papillae but grooved
along one side. Two swollen lateral folds along the sides of the head extend back
to segment VI and join a transverse dorsal fold which has a smooth margin. Two
groups of 4 gills on segment III, each group united basally and arranged 2 in front
and 2 behind. The two outer gills of each group are smooth and tapered ; the two
inner ones are bipennate for more than half their length with well developed lateral
pinnules. Segments III, IV and V with fine, pointed neurosetae embedded in the
lateral fold. No neurosetae on segment VI. No notosetae on segment III which
bears the gills. A pair of stout dorsal hooks on segment IV behind the gills. Seg-
ments V and VI with small notopodial capillaries. Segment VII and the next 12
segments bear smooth-winged notopodial capillaries and neuropodial uncini.
Altogether a total of 17 thoracic setigers. Abdomen with 32 or more segments.
Abdominal uncini borne on square pinnules each of which has a small superior papilla.
Uncini with a single series of 5 teeth above the small rostral point and a recurved
basal prow.

Augener (1914) was doubtful as to whether /. warnebroensis from South West
Australia was identical with /. pulchella, but the only difference he noted was the
presence of eye-spots. Specimens from Burma in the British Museum (Natural
History) identified as /. warnebroensis are identical with /. pulchella. /. sibogae
Caullery, 1944 from Indonesia is even closer for it lacks eye-spots.

/. whydaensis Augener, 1918 from tropical West Africa is a different species. A
specimen in the British Museum (Reg. No. 1953 : 3 : i : 876) from Accra identified
by Tebble 1955 has only 16 thoracic setigers of which 12 bear uncini. Moreover
the branchiae are long and slender, the inner two pairs bearing numerous minute
lateral pinnules which are quite different from the long paired pinnules of /.
pulchella.

Samythella affinis sp. n.

(Text-fig, ii b-e)

RECORDS. SCD 237 (7) ; 275 (8).

DESCRIPTION. The type material (B.M.(N.H.), Reg. No. 1963.1.142-144) was
obtained from station SCD 237 at 34 51' S./23 41' E. in 183 metres on a bottom
of fine sand. Only one specimen has a complete set of gills. Complete specimens
measure 10 mm. by 0-8 mm. across the anterior thorax.

The diagnostic characters may be summarised as follows : Prostomium (Text-fig.
II e) depressed and bluntly triangular. No eyes nor glandular ridges. Buccal
tentacles smooth. Buccal segment and segment II well defined. Segment III
(the paleal segment) with a well developed branchial ridge bearing 3 pairs of smooth
gills in a single transverse row. No median gap nor nephridial papillae on the

436

J. H. DAY

FIG. 1 1 . Pectinaria koreni cirrata : (a) ventral view of head ; Samythella affinis : (b) pos-
terior abdominal pinnule ; (c and d) lateral and face view of thoracic hook ; (e) anterior
end. Pista unibranchia : (/) dorso-lateral view of anterior end ; (g and h) face-on and
lateral views of hook from first row. A mphitrite pauciseta : (j,j') notoseta and details
of tip ; (k) lateral view of anterior end ; (/ and m) face-on and lateral views of thoracic
hook.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 437

branchial ridge. Individual gills tapered and project far in front of the prostomium.
Segment III with about 10 paleae on each side of the branchial ridge. Paleae
longer than normal capillaries. Segments IV and V with a small notopodia bearing
a few small capillaries, segment VI with normal capillaries, segment VII and the
next ii segments with both normal capillaries and uncini. No specialised posterior
notosetae. Altogether a total of 15 thoracic setigers behind the paleae. Thirteen
thoracic segments with glandular ventral pads.

Eleven abdominal segments with uncini. Anterior abdominal segments with
short uncigerous tori, later ones (Text-fig. 11 b) with uncigerous pinnules which are
expanded distally. No dorsal cirri or rudimentary notopodia.

Thoracic notosetae are smooth-winged capillaries. Thoracic uncini (Text-fig,
ii c and d] with a single series of 5 teeth and a smaller attachment process above the
recurved basal prow.

The generic position of this species is uncertain since there has been no uniformity
in the definition of the genera. Seven genera of the subfamily Ampharetinae lack
glandular ridges on the prostomium and have 3 pairs of smooth gills ; they are
Amythas, Aryandes, Eusamytha, Glyphanostomum, Neosabellides, Samytha and
Samythella. Only Aryandes Kinberg has paleae and this is a doubtful genus of
which the type species has been lost. Of the rest only Samythella and possibly
Eusamytha agree in having smooth tentacles and a single row of teeth on the thoracic
uncini. It is suggested that the characters of the genus Samythella be amended to
include species with paleae. The proposed definition is as follows :

Samythella Verrill, 1873 (characters amended).

Prostomium a tongue-shaped lobe without glandular ridges. Tentacles smooth.
A transverse row of 6 smooth gills on the branchial ridge formed of right and left
groups of 3 each. No nephridial papilla on the branchial ridge. Segment III
(the branchiferous segment) with notosetae either absent or enlarged to form
paleae. Segments IV, V and VI with notopodial capillaries only. Segment VII
and the next n segments with both notopodial capillaries and uncini. No specialised
posterior notosetae. Uncini with a single row of teeth. Abdomen of ii or more
segments sometimes bearing vestigial notopodia above the uncigerous pinnules.

TYPE SPECIES. S. elongata Verrill, 1873.

Family TEREBELLIDAE
Artacama proboscidea Malmgren, 1865

Artacama proboscidea Malmgren, 1865 : 394 ; Hessle, 1917 : 194, pi. 2, fig. 13.

RECORDS. WCD 109 (i).

NOTES. The single specimen is broken posteriorly. The main characters may
be summarised as follows. Tentacular lobe horse-shoe-shaped with a dorsal notch.
No eye-spots. Tentacles numerous and either broken or very short. A swollen
proboscis covered with coarse papillae below the mouth opening. Segments 2 and
3 flanged but do not form lateral lobes. Three pairs of filiform gills on segments

438 J. H. DAY

2, 3 and 4. Each gill is a tuft of filaments arising from a basal stump. A large
nephridial papilla on segment 3 and small ones on segments 6, 7, 8 and 9. Seventeen
bundles of notosetae starting on segment 4. Uncini start on segment 5 and are
arranged in double rows on the posterior thorax. Eleven glandular ventral pads.
Abdomen incomplete but has 20 segments with large lamellar uncigerous pinnules.
The base of each pinnule is constricted and the uncini are limited to part of the ventral
margin. Notosetae are winged capillaries with smooth tips and the uncini are
avicular with a short base, a stout main fang and a crest of about 20 denticles in
4 irregular arcs.

This is the first record of this genus from South Africa.

Pista unibranchia sp. n.

(Text-fig, ii f-h)

RECORDS. FAL 375 (i), 413 (12).

DIAGNOSIS. A single gill with a stalk and a terminal pom-pom of short filaments
First row of uncini without basal shafts.

DESCRIPTION. The holotype is a complete specimen measuring 10 mm. by
0-5 mm. dredged from Station FAL 413 in False Bay at 34 12-5' S./i8 37' E. in 48
metres on sand and shell. It is colourless in alcohol and the fragile tube is en-
crusted with coarse sand grains.

The tentacular lobe (Text-fig, n / ) is short and collarlike with numerous long
grooved tentacles. Eye-spots, if present, must be few and minute. The upper lip
is well developed. Glandular lateral lobes grow forward from segments 2 and 3
to form a stout sheath which supports the tentacles. The lateral lobes are unusual
for they are not limited to the sides of the body but extend dorsally and ventrally
practically encircling the anterior end. Small lobes also occur on segment 4. A
single median gill is present on setiger 2. There is no sign of any other gills in all
the thirteen specimens examined. The gill consists of a stout trunk bearing a
terminal tuft of short filaments arranged in 6-7 whorls. The trunk and the tuft
of filaments are about the same length.

Notosetae start on segment 4 and there are 17 bundles in all. Uncini start on
segment 5 and are arranged in alternating rows on the posterior thorax. Fifteen
segments with poorly defined ventral pads.

The abdomen is gently tapered and consists of about 26 long segments. The
uncigerous tori are short throughout and only towards the end of the abdomen do
they project as square pinnules. The pygidium is missing.

The notosetae are all broad-winged capillaries with smooth tips. The first bundle
on segment 4 is rather smaller than the rest. The uncini (Text-fig, n g and h)
are arranged in rather short rows and none of them, even those of the first row,
have basal prolongations or shafts. The base is rounded, the main fang is very stout
and above this is a cap of denticles arranged in 4-5 alternating close-set rows. In
profile there appear to be 4-5 teeth but a face-on view shows that the dental formula
is roughly MF : 4 : 5 : 7 : 10.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 439

At first sight this species resembles a juvenile Pista cristata which has lost all but
one of its four pom-pom shaped gills. A closer examination shows no sign of scars
where the missing gills should be. The lack of shafts on the first row of uncini
suggests P. macrolobata Hessle but this species has dichotomously branched gills.
Holotype : B.M.(N.H.), Reg. No. 1963.1.153.

Amphitrite pauciseta sp. n.

(Text-fig, ii j-m)

RECORDS. WCD 83 (3), 85 (common).

DIAGNOSIS. This species may be recognised by the presence of pigment granules
at the bases of the two pairs of gills and the possession of only 13 thoracic setigers.

DESCRIPTION. The type material consists of 9 specimens selected from station
WCD 85 dredged at 33 06' S./I7 45' E. on mud at a depth of 160 metres.

The holotype is 18 mm. long by i-o mm. across the thorax and is pale brown in
alcohol. The tube is very fragile with adherent sand grains of various sizes. The
tentacular lobe (Text-fig, n k) is collar-shaped with numerous long grooved tentacles
and a few eye-spots. The lower lip is well developed. Small lateral lobes are
present on segments 2, 3 and 4 but are not obvious on some of the paratypes. Two
pairs of gills on segments 2 and 3. The trunk of each gill is short and the branches
are digitiform so that the first impression is that each gill consists of a tuft of stout
filaments. At the base of each gill there is a group of reddish subdermal spots
probably associated with the vascular system. These spots were seen on every
specimen examined and are very characteristic. The holotype also has a few similar
spots next to the notopodium of segment 4.

Notosetae start on segment 4 and are present on 13 segments. Uncini start on
segment 5 (setiger 2) and are arranged in double rows on posterior thoracic segments.
A nephridial papilla is present on segment 3 just lateral to the base of the gill. Ten
glandular ventral pads are present from segment 3 to segment 14.

The abdomen is tapered and consists of 42 segments. The uncigerous tori
gradually lengthen to become long pinnules on the last few segments. The anus
has a pleated margin.

The notosetae (Text-fig. 11 j) have narrow blades and end in fine, minutely ser-
rated tips (Text-fig, ii j'). The uncini (Text-fig, n / and m) have short bases and
when seen in profile appear to have 5 indistinct teeth above the main fang ; in full
face however it may be seen that there are numerous irregular arcs of denticles
approximating to the formula MF : ca 5 : ca 7 : ca 9 : ca 12.

As far as I am aware this is the only species of Amphitrite with less than 17 thoracic
setigers. Holotype : B.M.(N.H.), Reg. No. 1963.1.145.

Family SABELLIDAE
Fabricia filamentosa sp. n.

(Text-fig. 12 a-g)

RECORDS. WCD 83 (2), 109 (10), 122 (2) ; SCD 223 (4).

DESCRIPTION. The holotype is a complete specimen from Station WCD 109

440 J. H. DAY

collected by a grab from 32 08' S./iy 39' E. in 170 metres on a bottom of dark
green mud. There are 9 paratypes.

The whole worm (Text-fig. 12 a) is encased in a fragile tube of flocculent debris.
It is very slender measuring 4 mm. in length by 0-2 mm. broad and is colourless in
alcohol. Each brancial lobe (Text-fig. 12 b) bears 3 radioles, each with 5 pairs of
very long slender pinnules which extend like a tuft of filaments twice as far as the
radiole itself. There are no palps. The whole branchial crown is about a quarter
the length of the body. The collar is well developed and continuous ventrally.
It is notched back dorsolaterally on each side to form a tongue-shaped median dorsal
lappet but the rest of the collar is continuous and has a smooth margin. The median
lappet is soldered onto a pair of dorsal lips inside the collar. A pair of conspicuous
eyes is present inside the collar at the base of the branchiae. No otocysts were seen.

The slender body consists of 8 thoracic and 3 abdominal segments which differ in
length. The first 3 setigers are about twice as long as broad, the 4th and 5th are
longer and the 6th, 7th and 8th are about five times as long as broad. The three
abdominal segments decrease in length until the third is broader than long. All
abdominal segments have swollen tori while the pygidium is tapered and bears a
pair of terminal eye spots.

Setiger i bears 4-5 narrow-bladed capillaries. The notosetae of setigers 2-8 are
4 capillaries (Text-fig. 12 c) with blades which are broad at the base but taper to
slender tips. The neurosetae are rows of about 4 long-shafted hooks (Text-fig.
12 d and e) each with a transverse arc of about 5-7 teeth above the rostrum. In the
abdomen each neuropodium bears 3 slender-bladed capillaries and the notopodium
has a row of about 15 long-handled uncini (Text-figs. 12/andg), each with 6-7 rows
of teeth with 3-4 long teeth per row. Holotype : B.M. (N.H.), Reg. No. 1963 . i . 169.

This species is most closely related to F. capensis (Monro) . It has a similar collar
and elongated segments in the posterior thorax. The important distinctions con-
cern the setae and the branchiae for F. capensis has abdominal uncini with very
numerous teeth (18 rows of 8 each), thoracic uncini with 2 teeth above the rostrum
and then an arc of 12-14 smaller denticles. Moreover F. capensis does not have such
extremely elongated branchial filaments. It is a much larger worm living gregari-
ously on the shore in sandy tubes.

Fabriciola cf. mossambica (Day, 1957)

Fabricia mossambica Day, 1957 : 115, fig. 8 e-o.

RECORDS. WCD 83 (2), 108 (4) ; SCD 233 (4).

NOTES. These specimens are referred to the tropical East African form with
hesitation, for the abdominal uncini have a greater number of teeth. They agree in the
possession of filamentous palps, a triangular lobe below the mouth in place of a collar,
similar thoracic setae and abdominal capillaries but the uncini have many more
teeth. A juvenile had 8 rows of 4 teeth and an adult had 14 rows of 8 teeth. In F.
mossambica there are 7 rows with 2-3 teeth per row. It is possible that this character
is more variable than had been realised.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

441

FIG. 12. Fabricia filamentosa : (a) entire work ; (b) head ; (c) thoracic notoseta ; (d and e)
lateral and face-on views of thoracic hook ; (/ and g) lateral and face-on views of
abdominal hook. Spirorbis adeonella : (h) tube ; (_/) plan view of operculum ; (k and /)
edge-on and lateral views of thoracic uncinus ; (m) side view of operculum to show talon ;
() fine capillary from first bundle ; (0) specialised collar seta ; (p) smooth-winged capil-
lary from second bundle.

442 J. H. DAY

According to Banse (1956) this species with its filamentous palps should be referred
to the genus Fabriciola.

Family SERPULIDAE
Hydroides heteroceros (Grube, 1868)

Serpula (Eupomatus) heterocercos Grube, 1868 : 639.

Hydroides heteroceros : Pixell, 1913 : 75, pi. 8, fig. 2 a-c ; Fauvel, 1953 : 459, fig. 2416.

RECORDS. " William Scoresby " station WS 1022, at 25 26' S./46 07' E. in
58 metres on sand and shell.

NOTES. This is a new record for the area which is just south of Madagascar.
H. heteroceros is widely distributed in the tropical Indian Ocean and the present
specimens agree with PixelTs description.

Spirorbis (Paralaeospira) adeonella sp. n.

(Text-fig. 12 h-p)
Spirorbis (Laeospira) sp. Day, 1961 : 557.

RECORDS. SCD 154 (i) ; MB 7 (1 living).

DESCRIPTION. The additional material allows me to amplify and amend the
preliminary account in 1961 of Spirorbis sp. The holotype was dredged in Mossel.
Bay growing on the calcareous Polyzoan Adeonella sp. at a depth of 19 metres
The shell (Text-fig. 12 h) is 1-5 mm. in diameter and is thick and vitreous with well
marked annular ridges. It is coiled clockwise when viewed from above (sinistral).
The operculum (Text-fig. 12 _;') is oval in outline with a concave surface. The talon
(Text-fig. 12 m) is a smoothly swollen knob almost at right angles to the opercular
plate which is strengthened by a median ventral ridge. The branchiae consist of
7 unequal radicles with slender tips. Incubation is in the tube. The collar is
well developed and the two halves overlap dorsally but are not united. There
are 4 thoracic segments but the 4th is asymmetrical. Thus setiger i bears the collar
setae, setigers 2 and 3 bear both notosetae and uncini but setiger 4 has only a short
row of uncini on the concave side of the body. The abdomen is short and consists
of about 13 segments.

The collar setae includes a few fine capillaries (Text-fig. 12 n) and several larger
setae (Text-fig. 12 o) with very finely serrated blades and a separate dentate lobe at
the base. Setigers 2 and 3 also have a few fine capillaries in addition to the numerous
smooth-winged capillaries (Text-fig. 12 p). There are no " setae of Apomatus ".
The thoracic uncini (Text-fig. 12 h and /) have 2-3 rows of teeth with about 18 teeth
per row. The gouge is expanded and truncate.

Two other species of the subgenus Paralaeospira have been recorded from South
Africa, namely 5. (P.) patagonicus and 5. (P.) capensis. S. (P.) adeonella differs
from both in having an annulated shell and in lacking setae of Apomatus. The
talon is also different. Holotype : B.M.(N.H.), Reg. No. 1963.1.179.

THE POLYCHAETE FAUNA OF SOUTH AFRICA 443

REFERENCES

ARWIDSSON, I. 1906. Studien tiber die skandinavischen und arktischen Maldaniden nebst

Zusammenstellung der iibrigen bisher bekannten Arten dieser Familie. Zoo/. Jb. Suppl. 9 :

1-308.
AUGENER, H. 1910. Bemerkungen iiber einige Polychaeten von Roscoff, iiber zwei neue

Polynoiden des Berliner Museums und iiber die Brutpflege von Hipponoe gaudichaudi

Aud.-M. Edw. Zoo/. Anz., 36, 232-249.

1914. Polychaeta Sedentaria. In Michaelsen, W. & Hartmeyer, R., eds. Die Fauna

Sudwest Austr aliens., 5 : 1-170, Jena.

1918. Polychaeta. In Michaelsen, W., ed. Beitrage zur Kenntnis der Meeresfauna Westa-

frikas., 2 : 67-625. Hamburg.

1931. Die bodensassigen Polychaeten nebst einer Hirudinee der Meteor-Fahrt. Mitt.

zool. St. Inst. Hamb., 44 : 297-313.
BANSE, K. 1956. Beitrage zur Kenntnis der Gattungen Fabricia, Manayunkia und Fabriciola

(Sabellidae, Polychaeta). Zoo/. Jb. (Syst.), 84 (4-5) : 415-438.
BERGSTROM, E. 1914. Zur Systematik des Polychaeten Familie der Phyllodociden. Zoo/. Bidr.

Uppsala, 3 : 37-224.

1916. Die Polynoiden des schwedischen Siidpolar-Expedition 1901-1903. Zoo/. Bidr.

Uppsala, 4 : 249-304.

CAULLERY, M. 1944. Polychetes Sedentaires de 1'Expedition du Siboga. Siboga-Exped.

Leiden, 24 (2) : 1-204.
CERRUTI, A. 1909. Contributo aH'Anatomia, biologic e sistematica delle Paraonidae (Levin-

senidae) con particolare riguardo alle specie del golfo di Napoli. Mitt. Zool. Stat. Neapel.,

19 : 459-512.
CLAPARISDE, E. 1870. Les Annelides Ch6topodes du Golfe de Naples. Mem. Soc. Phys.

Geneve 20 (2) 365-542.

& MECZNIKOW, E. 1869. Beitrage zur Kenntniss der Entwicklungsgeschichte der

Chaetopoden. Zeitschr. wiss. Zool. Leipzig, 19, 163-205.

DAY, J. H. 1951. The polychaet fauna of South Africa. Part i : The intertidal and estuarine
Polychaeta of Natal and Mosambique. Ann. Natal Mus., 12 : 1-67.

1953- The polychaet fauna of South Africa. Part 2 : Errant species from Cape shores

and estuaries. Ann. Natal Mus., 12 : 397-441.

1955- The Polychaeta of South Africa. Part 3 : Sedentary species from Cape shores

and estuaries. /. Linn. Soc. (Zool.}, 42 : 407-452.

1960. The polychaet fauna of South Africa. Part 5 : Errant species dredged off Cape

coasts. Ann. S. Afr. Mus., 45 (3) : 261-373.

1961. The polychaet fauna of South Africa. Part 6 : Sedentary species dredged off Cape

coasts with a few new records from the shore. /. Linn. Soc. (Zool.}, 44 : 463-560.
DAY, J. H. 1963. The polychaete fauna of South Africa. Part 7 : Species from depths between

1,000 and 3,300 metres west of Cape Town. Ann. S. Afr. Mus. 46 (14) : 353-371.
EHLERS, E. 1901. Die Polychaeten des magellanischen und chilenischen Strandes. Festschr.

konig. Gesell. Wiss. Gottingen (Math. -Phys.}, 232 pp. Berlin.

1908. Die bodensassigen Anneliden aus der Sammlungen der deutschen Tief see-Ex-
pedition. Wiss. Ergebn. " Valdivia ", 16 : 1-67.

1913. Die Polychaten-Sammlungen der deutschen Siidpolar-Expedition 1901-1903.

Dtsch. Siidpol.-Exped., 13 : 397-598.

FAUVEL, P. 1923. Polychetes errantes. Faune Fr., 5 : 1-488.

19230. Sur quelques polychetes de 1' Angola Portugaise. Handl. Goteborg Kon.

Vetensk., 20 : 1-13.

1927. Polychetes sedentaires. Faune Fr., 16 : 1-494.

1934. Annelides polychetes de Rovigno d'Istria. Thallasia, Jena., 1 (7) : 1-78.

1936. Polychetes. Res. voy. " Belgica " 1897-99. Zool., 1-44.

GRUBE, A. E. 1857. Annulata Oerstediana. Part 2.

1868. Ueber neue Anneliden. Jb. Schles. Gesells. Breslau, 45, 50-52.

444 J. H. DAY

HARTMAN, O. 1944. Polychaetous annelids. Part 5. Eunicea. Allan Hancock Pacif. Exped.,
10 : 1-238.

1948. The marine annelids erected by Kinberg with notes on some other types in the

Swedish State Museum. Ark. Zool., 42A : 1-237.

- 195- Goniadidae, Glyceridae and Nephtyidae. Allan Hancock Pacif. Exped., 15 : 1-181.

1953. Non-pelagic Polychaeta of the Swedish Antarctic Expedition, 1901-1903. Further

Zool. Res. Swed. Antarct. Exped., 4 (2) : 1-183.

- 1957- Orbiniidae, Aspitobranchidae, Paraonidae and Longosomatidae. Allan Hancock
Pacif. Exped., 15 (3) : 211-392.

- 1959- Catalogue of the polychaetous annelids of the world. Part I, pp. 1-353 ; Part 2,
pp ; 354-628. Allan Hancock Publ. occ. paper, 23.

HESSLE, C. 191?- Zur Kenntnis der terebellomorphen Polychaeten. Zool. Bidr. Uppsala,
5 : 39-258.

KINBERG, J. G. H. 1865. Annulata nova. Fork. Oefv. Vet. Akad. Stockholm, 21 : 559-574.

KIRKEGAARD, J. B. 1959. The Polychaeta of West Africa. Sci. Res. Danish Exped. Trop. W.
Afr. 1945-46, " Atlantide " Rep., 5 : 7-117.

KITAMORI, R. 1960. Two new species of Cirratulidae and Nephthydidae (Annelida Poly-
chaeta). Bull. Jap. Soc. Sci. Fish. ,26 (n) : 1082-1085.

MclNTOSH, W. C. 1876. On British Annelida. Trans. Zool. Soc. Lond., 9 : 371-394.

- 1879. On the Annelida obtained during the cruise of H.M.S. " Valorous " to the Davis
Strait in 1875. Trans. Linn. Soc. Lond. (n.s.) 1 : 499-511.

1904. Marine annelids (Polychaeta) of South Africa. Mar. Invest. S. Afr., 3 : 17-92.

- 1900. A monograph of the British Annelids. Polychaeta. Amphinomidae to Sigalionidae.
1(2): 215-442. Ray Soc., London.

MALMGREN, A. J. 1865. Nordiska Hafs-Annulater. Fork. Oefv. K. Vetensk. Akad. Stockholm,

22 : 355-410-
1867. Annulata Polychaeta Spetsbergiae, Groenlandiae, Islandiae et Scandinaviae

hacterus cognita. Fork. Oefv. K. Vetensk. Akad. Stockholm, 24 : 127-235.
MARENZELLER, E. VON. 1875. Zur Kenntniss der adriatischen Anneliden. Zweiter Beitrag.

(Polynoiden, Hesioneen, Syllideen.) Sitzber. Akad. Wiss. Wien., 72 (i) : 129-171.

- 1893. Zoologische Ergebenisse II. Polychaten des Grundes, gesammelt 1890, 1891 und
1892. Deutschr. Akad. Wiss. Wien., 60 : 25-48.

MONRO, C. C. A. 1930. Polychaete worms. " Discovery " Rep., 2 : 1-222.

1936. Polychaete worms. II. " Discovery " Rep., 12 : 59-198.

1939- Polychaeta. Rep. Antarctic Res. Exped. 1929-1931. Ser. B. (Zool. Bot.), 4 (4) :

89-156.

MOORE, J. P. 1911. The polychaetous annelids dredged by the U.S.S. " Albatros " off the

coast of Southern California in 1904. Euphrosynidae to Goniadidae. Proc. Acad. Nat.

Sci. Philad., 63 : 234-318.
MULLER, F. 1858. Einiges iiber die Anneliden Fauna der Insel St. Catharina an der Braziliani-

schen Kiiste. Arkiv. Naturg. Berlin, 24 (i) : 211-220.
PALLAS, P. S. 1766. Miscellanea Zoologica. 1-7, Hague.
PETTIBONE, M. H. 1961. New species of polychaete worms from the Atlantic ocean with a

revision of the Dorvilleidae. Proc. Biol. Soc. Wash., 74 : 167-186.
PIXELL, H. L. M. 1913. Polychatea of the Indian Ocean together with some species from the

Cape Verde Islands. The Serpulidae with a classification of the genera Hydroides and Eupo-

matus. Trans. Linn. Soc., 16 : 69-92.
RAMSAY, L. N. G. 1914. Polychaeta of the family Nereidae collected by the Scottish National

Antarctic Expedition (1902-1904). Trans. Roy. Soc. Edinb., 50 : 41-48.
REISH, D. J. 1958. Description of a new species of Cossura (Annelida Polychaeta) from the

Mississippi delta. /. Wash. Acad. Sci., 48 : 53-55.

THE POLYCHAETE FAUNA OF SOUTH AFRICA

445

SAINT-JOSEPH, Baron A. DE. 1888. Les annelides polychetes des cdtes de Dinard. Part 2.
Ann. Sci. not. Paris, 5 (7) : 141-338.

1894. Les Annelides polychetes des cdtes de Dinard. Pt. 3. Ann. Sci. Nat. Paris

(ser. 7) 17 : 1-395-

SARS, M. 1865. Forsatte Bidrag til Kunskaben om Norges Anneliden. Fork. Vidensk.
Selsk. Chrisiiania, 1865, 5-20.

1872. Diagnoser af nye Anneliden fra Christiania-fjorden. Fork. Vidensk. Selsk.

Christiania 1872, 406-417.

SCHMARDA, L. K. 1861. Neue wirbelose Thiere. 1 : 1-164, Leipzig.

SOUTHERN, R. 1914. Archiannelida and Polychaeta (Clare Island Survey). Proc. Roy. Irish

Acad. Dublin, 31 (47) : 1-160.

TAUBER, P. 1879. Annulata Danica. 1-144. Copenhagen.
TEBBLE, N. 1952. On three new species of the genus Ophelia (Polychaeta) from British and

adjacent waters. Ann. Mag. nat. Hist., (12) 5 : 553-571.

1955- The polychaete fauna of the Gold Coast. Bull. Brit. Mus. (nat. Hist.) Zool.,

3 : 59-148.

VERRILL, A. E. 1879. Report upon the invertebrate animals of Vineyard Sound and the

adjacent waters, with an account of the physical characters of the region. Rep. U.S. Com.

Fish. Wash. 1871-1872, 295-778.
WESENBERG-LUND, E. 1949. Polychaetes of the Iranian Gulf. Danish Sci. Invest. Iran, 4 :

247-400.
WILLEY, A. 1902. Polychaeta. In : Report on the collections of natural history made in

the Antarctic regions during the voyage of the " Southern Cross ". London.

1904. Littoral Polychaeta from the Cape of Good Hope. Trans. Linn. Soc. (Zool.),

9 (6) : 255-268.

ZOOL. IO, 7

THE FRESHWATER GASTROPOD
MOLLUSCA OF ANGOLA

C. A. WRIGHT

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 10 No. 8

LONDON: 1963

C. A. WRIGHT

British Museum (Natural History)

Pp. 447-528 ; Plates 1-16 ; 126 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 10 No. 8

LONDON: 1963

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
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.

This paper is Vol. 10, No. 8 of the Zoological series.

Trustees of the British Museum 1963

Issued June, 1963 Price Thirty-three Shillings

THE

FRESHWATER GASTROPOD MOLLUSCA

OF ANGOLA

By C. A. WRIGHT

British Museum (Natural History)

THIS account of the freshwater gastropod fauna of Angola is based largely on
material collected by the author during the months of October and November
1957. Due to the enormous size of the territory and the short duration of the visit
the collection cannot be considered to be completely representative of the Angolan
snail fauna, but it includes the majority of species recorded from the area as well as
several others not previously known to occur there. Wherever possible the type
material of the species concerned has been examined and photographs of these
specimens are included. Anatomical descriptions of the pulmonate species are
given and an attempt has been made to indicate the range of morphological variation
encountered in many of them, particularly those members of the Planorbidae
which are known or potential intermediate hosts of the human blood-flukes.
Synonymies are on a purely regional basis since it is felt that there is not yet sufficient
information from the whole continent to justify more sweeping conclusions.

One of the primary objects of the expedition was to obtain topotype material
of species described from Angola, but it became apparent that, due to inadequate
definition of original localities, and to recent changes brought about by development
of the country, this was not possible. The main effort was therefore concentrated
on obtaining large samples of specimens from four regions of the country, the coastal
plain in the neighbourhood of Luanda, the escarpment zone around Salazar, the
northern part of the plateau around Malange and the southern area of the plateau
in the Ganguelas district.

It is a great pleasure to acknowledge the assistance which I received from members
of the Angolan Health Department, in particular the Director, Dr. Barrata Feio,
also Dr. Pereira Nunes and all those who were so generous with their hospitality
and time. Many other people, both private individuals and local administrators,
contributed to the success of the field-collecting work and to all of them I am
indebted as well as to Mr. D. Claugher who undertook the tedious work of making
a great number of radula preparations.

The history of the malacological exploration of Angola covers just over a century.
The first reports of shells of freshwater snails are those of Dunker (1845 and 1853)
based on material collected by Dr. George Tarns on a voyage to Lower Guinea in

C. A. WRIGHT

MOSSAMEDES

the years 1841 and 1842. These collections were made only in the coastal region
in the neighbourhood of Luanda and Benguela. In 1866 Morelet published a pre-
liminary account of the molluscs collected by Dr. Friederich Welwitsch during his
historic exploration of Angola between the years 1853-61 and in 1868 Morelet
followed his earlier paper with a more comprehensive report of the malacological
results of this expedition. Welwitsch's contribution to the exploration of Africa
was quite outstanding ; a brief summary of the extent of his journeys is an essential
part of this account, for it was the areas he visited that were the primary objectives
of the present expedition. He arrived in Angola in 1854 and spent the first year
exploring the coastal plain from Quizembo (north of Ambriz) to the mouth of the
River Cuanza, a distance of about 120 miles. He then followed the course of the
Rio Bengo inland for about 125 miles to Sange in the Golungo-Alto district and

FRESHWATER GASTROPOD MOLLUSCA OF ANGOLA 451

spent two years based on this centre, until ill health forced him to move up to the
plateau. He travelled through Ambaca to Pungo Andongo where he set up temporary
headquarters and spent eight months exploring this part of the plateau region.
His journeys took him to the Cuije and Lombe rivers in the east, also to the islands
of Calemba in the Rio Cuanza between the cataracts of Condo and the Quissonde
rapids, a distance of about 250 miles from the coast. He visited salt marshes near
Quitage and the forests on the north bank of the Cuanza, then he travelled towards
the Rio Luxillo to the west of Pungo Andongo before beginning his return journey
to Sange. On the way back to Sange he tried to follow the Rio Lucala up into the
Duque de Braganza district but his porters deserted and he was forced to abandon
this part of his journey. On his return to Luanda he made a brief trip to the
Libongo area near the mouth of the Rio Lifune, then, in June 1859, ne travelled
by sea to Benguela, stayed for a short while in the neighbourhood of the town,
and went on to Mossamedes where he stayed for about two months before exploring
to the south to Pinda, Cap Negro and the Bahia dos Tigres. He returned to
Mossamedes and followed the Rio Mayombo inland to Bumbo and, crossing the
Serra de Xella on to the plateau of Huila, he stayed for about seven months until
a tribal rebellion forced him to return to the coast and thence back to Luanda and
eventually to Lisbon where he arrived in January 1861.

The collection of land and freshwater molluscs which Welwitsch accumulated
during the course of these journeys was the most comprehensive ever made in
Angola and was one of the major contributions to African malacology in the
nineteenth century.

In 1873 Sowerby described as a new species Physa benguellensis and gave the
type locality as Benguela, Egypt. At first sight this appears to be merely another
of Sowerby's regrettable lapses of geographical accuracy but the existence of a
coastal village called Egito near Benguela provides a satisfactory explanation.
The next contribution to the literature on the Angolan mollusc fauna was that of
von Martens (1882) with descriptions of a number of terrestrial species and a new
melaniid, and some years later Nobre (1905) described the collections made by
Francisco Newton in 1903.

For nearly forty years nothing further was written concerning the freshwater
molluscs of Angola, but in a more general work on the bulinids of Africa Bequaert
and Clench (1931) included photographs of some of the paratype specimens of species
described earlier by Morelet. During this period several most important works on
the fauna of neighbouring territories were published and amongst these were
Pilsbry and Bequaert's (1927) exhaustive monograph on the freshwater molluscs of
the Belgian Congo and Connolly's (1930) paper on the non-marine Mollusca of
South-west Africa. Both of these works helped to bring the existing knowledge
of the Angolan fauna into perspective with that of the rest of Africa.

Interest in African freshwater snails received a considerable stimulus from
Leiper's (1915) demonstration of the part played by certain members of the Planor-
bidae as intermediate hosts for human blood flukes of the genus Schistosoma, but it
was not until 1942 that this influence became active in Angola with Bruno de

452 C. A. WRIGHT

Mesquita's account of schistosomiasis in the territory and his publication of a list,
compiled by Germain, of possible snail hosts for the parasites. Mesquita published
another paper in 1953 and Sarmento wrote on a number of foci of schistosomiasis
in the south in 1944, but neither of these contributed much to the knowledge of
the molluscs. In 1951 the Endemic Diseases Mission started a survey of the fresh-
water snails of Angola and this has resulted in several recent papers such as Azevedo
and Medeiros (1956) on the anatomy of an Angolan Biomphalaria, Janz and Carvalho's
(1956) general account of schistosomiasis in Angola with lists of mollusc species
collected and, most recently, Carvalho and Janz (1960) on the Angolan species of
Bulinus.

The province of Cabinda, although administratively a part of Angola, is separated
from the rest of the territory by the Congo river and a narrow strip of the Congo
Republic. It was not visited during this expedition and little is known of its
mollusc fauna apart from some information in Pilsbry and Bequaert (loc. cit.) and
Dartevelle's 1934 work on the aquatic molluscs of the principal river, the Shilloango.

Biomphalaria salinarum (Morelet)

Planorbis salinarum Morelet, 1868: p. 85, pi. 5, fig. 4.

MATERIAL. Quissol (2), Quastimbala stream, near Malange. 13 specimens
23rd October, 1957.

HABITAT. On aquatic vegetation in a very gently flowing part of a partially
dammed stream.

SHELL (PI. i, figs. 7, 8). All of the specimens collected were immature, the largest
had about 3^ whorls. The mean dimensions of the sample and the range (in
millimetres) is maximum diameter 7-35 (3-5-9-8), umbilicus 2-35 (-88-3-32) and
height 3-28 (2-45-4-37). The upper side of the whorls is slightly flattened and the
lower side gently rounded without any well marked angulation. The first whorl
on the upper side is deeply sunken and the umbilicus is shallow. The shells are
white beneath the encrusting algae and detritus and a fine spiral microsculpture
intersects the growth lines to give a granular appearance to the first 2^3 whorls
on the under side and, although this is present on the upper side it is not so well
marked.

ANATOMY. The mantle varies in colour from creamy- white to pale grey with
discrete black spots of fairly regular outline. The intermediate ridge between the
kidney and rectum is well developed. The male copulatory organ is well developed
in the specimens examined and the penis-sheath is approximately equal in length
(occasionally a little longer) to the preputium and the penis is roughly as long as
the sheath (Text-figs. 1-3). The prostatic diverticula are short and close-packed,
most are divided but few have tertiary branches. The seminal vesicle consists of
an upper part of complicated digitate diverticula from the hermaphrodite duct and

FRESHWATER GASTROPOD MOLLUSCA OF ANGOLA 453

a lower portion of more regular rounded protuberances. The female reproductive
system is not fully developed in this material ; the receptaculum is clavate and
roughly equal in length to its duct which is in turn a little longer than the ,vagina
(Text-figs. 2, 3).

./.
RADULA. There are about 23 teeth in each half-row of which 6-8 are tricuspid

laterals. Subdivision of the endocone may occur in the fourth or fifth lateral.
The endo- and mesocones of the first laterals are about equal in length, they are broad
and partly fused at their bases and the tip of the mesocone is bluntly arrow-head
shaped (Text-fig. 17). The ectocones of the marginals are irregularly sub-divided
into several cusps of unequal size.

Biomphalaria salinarum was described by Morelet from material collected by
Welwitsch from a stream near Dungo near the Rio Cuije. There is no village of
this name in the area today and it was not possible to find the exact type locality
but the Quastimbala stream is a tributory of the Rio Cuije and the material des-
cribed here is therefore from the original district. The juvenile specimens collected
do not resemble the type series (PI. i, figs. 1-6) closely because of the differential
changes in shell form which occur with growth. However, if the ratios of shell
diameter to umbilical diameter and shell diameter to height are plotted and the
graphs extrapolated, the measurements for the type specimens fall close to the
appropriate lines. The dimensions in millimetres of the type series are : maximum
diameter 16-3, 14-0 and 13-8; umbilicus 6-65, 5-6 and 5-4; height 4-8, 4-7 and 4-2.

The affinities of B. salinarum appear to be closer to the alexandrina species group
defined by Mandahl-Barth (i957a) than to any other. The intersection of , the graphs
of the ratios mentioned above occurs at a shell diameter of about 13 mm. which is
close to that shown by Mandahl-Barth for B. alexandrina and the number of whorls
in the 14-0-16-3 mm. shell diameter group supports this. The broadly arrow-head
shaped mesocones of the first lateral radula teeth and the subdivided ectocones of
the marginals are also characteristic of the alexandrina group. The proportions
of the male copulatory organ do n