SESS ee
Baten aianit aoe TS

ssesaeataerss

7 he€2TOO TOEO gO

nig,

1OHM/18W

if

oo Patagonia, BAGsiRbo.

a Re B. HATCHER, in Cuarce

EDITED BY

WILLIAM B. SCOTT

- BLAIR PROFESSOR OF GEOLOGY AND PALZONTOLOGY, PRINCETON UNIVERSITY

VOLUME III

ZOOLOGY aK
Pane H,—Bapracnrans anp-Rerrires :

L—SEEJNEGER

U. S. NATIONAL MUSEUM

OF THE Agcuitavas eat ae THEORY

BY

C. EIGENMANN

INDIANA UNIVERSITY

Ee

jg

(Pp. 225-374. Pls. XXX-XXXVII)

PRINCETON, N. J.
THE UNIVERSITY
oy eae

= 1909

oF SPRESS DR ss».
HE NEW ERA PRINTING COMPANY
"| LANCASTER, PA

PARG iit.

THE FRESH-WATER FISHES OF PATAGONIA AND
AN EXAMINATION OF THE ARCHIPLATA-
AKCHHELENIS THEORY.

BY
CARE oH. EIGENMANN,

Inpiana UNIVvERsITY.
PROFESSOR W. B. Scott,
PRINCETON, NEw JERSEY.

Dear Sir: 1 enclose the MS. of my report on the fishes collected by
the late J. B. Hatcher in Patagonia. Instead of confining my report to
the few specimens collected by Mr. Hatcher I have utilized, as far as pos-
sible, the knowledge gained from previous collections, and have dealt
monographically with the fresh-water fishes of the area south of the line
joining the mouth of the Rio Negro and Santiago, Chili. Ichthyolog-
ically this area constitutes a faunal unit sharply defined from temperate
and tropical America to the north of it.

Since 1887 I have been busied more or less with the fresh-water fishes
of South America, and the present opportunity seemed to me the best to
use our knowledge concerning them to test the claims of the Archiplata-
Archhelenis theory. This theory must stand or fall by the evidence of
the fresh-water fishes. The task proved much more onerous than ex-
‘pected, but I feel amply repaid by the definite ideas gained by this
review.

It seems quite certain: (@) that tropical America obtained the elements
of its fauna in common with Africa before the Tertiaries, from some insig-
nificant common ground inhabited by Cichlids, Characins, and Catfishes
(Nematognaths) and perhaps types of wide distribution which remain only
as relicts; (6) that tropical America has not been accessible and received
few, if any, immigrants from other land areas since that time; but (c) that

225

° > ‘we .
Ss a@s eS BIW we ted

aN,

. :

226 PATAGONIAN EXPEDITIONS: ZOOLOGY.

the three types mentioned have undergone unparalleled adaptive radiation
in the growing and metamorphosing continent, the Cichlids developing
about 150 known species, the Characins 500, besides differentiating the
Gymnotids, and the Catfishes, 500 or more species, distributed in several
autochthonous families.

The grounds for all of these conclusions together with their corollaries
you will find set forth in detail, all of which I hope will meet with your
approval.

This report is divided therefore into a section dealing with The Fresh-
water Fishes of Patagonia, a section dealing with the Archiplata-Archhelenis
theory and a section listing all of the fresh-water fishes of South and Middle
America.

I must at this place acknowledge the most courteous assistance received
from Mr. Richard Rathbun and Mr. Barton A. Bean, of the United States
National Museum; Dr. G. A. Boulenger and Mr. C. T. Regan, of the
British Museum ; Dr. Franz Steindachner, of the K. K. Museum of Vienna ;
Mr. S. Henshaw, of the Museum of Comparative Zoédlogy, and Dr. R.
Gestro, of the Museo Civico di Storia Naturale, Genoa, Italy.

Respectfully submitted,
CarRL H. EIGENMANN,
Professor of Zoology, Indiana University, and Director of
the Indiana University Biological Station.

IVE ORIN Siva bein Noid es lke
Pa tACcON Te

INTRODUCTION.

The interest in the Patagonian fresh-water fish fauna is entirely out of
proportion to its diversity and centers largely in its origin. Only about
twenty-nine species of fishes are known to live or enter the fresh waters
south of the line joining Valparaiso and Bahia Blanca. These few species
fall, according to their origin, into four distinct groups.

1. Immigrants from the sea are: (@) in the process of acclimatization,
species of Mentdia and Atherinopsis, or (6) may be looked upon as long
established, species of Percichthys and FPercilia. Members of a are
found in all the rivers; members of 4 are found in the north chiefly, but
reach the Santa Cruz river.

2. Immigrants from the fresh waters on the north: a very small overflow
from the extremely rich fauna to the north and still retaining their generic
affinity with northern forms. Here belong the species of the genera
Chetrodon and Astyanax, which are very widely distributed in tropical
South America and are not known to extend much south of the Rio
Negro. Here also belongs Hafcheria, a southern modification of the
widely distributed Pygedium.

3. Autochthons, or of doubtful origin. Here belongs the highly inter-
esting Dzplomysfe, which is found on the northern border of Patagonia,
but is not a derivative of the tropical American fauna. It is a relict of
‘the original catfishes, in which the maxillary is still functional as a tooth
bearer. Here belongs also Vematogenys, a catfish related to Pygzdium.
Like the members of group 2, these species are confined to the northern
fringe of the Patagonian area.

4. There remain unaccounted for the members of the 4Aochitonide, Ga-
laxtide and Petromyzonide, chiefly of southern Patagonia. The two
former Boulenger puts in his Haplomi, an order of odds and ends of
fishes, containing such relicts as the Percopside, Amblyopside of North

227

228 PATAGONIAN EXPEDITIONS: ZOOLOGY.

America, Kuertzde of Africa, Daliiide of Alaska and Siberia, and the
dominant Peciiide of the tropical fresh waters, the Asoczde of the north
temperate zone, and the Scopeiide of the deep sea. By Gill, on the other
hand, they are placed in the order of Malacopterygii with the Clupeoids
and Salmonoids.

Of the AAplochitonide there are two genera, 4fplochiton with an unde-
termined number of species in the Patagonian region, and Pvofotroctes
with three species, one in Queensland, one in South Australia and one in
New Zealand.

Of the Galaxude there are two genera, /Veochanna (apoda) from New
Zealand, where it frequently burrows in damp clay away from water, and
Galaxtas, With about 30 species from New Zealand, New South Wales,
South Australia, Tasmania, Cape of Good Hope, southern South America
and the Falkland Islands. The Petromyzonide are found in all temperate .
fresh waters and seas, in both the northern and southern hemispheres.

The distribution of the two former families is of interest in connection
with the theory of a former antarctic continent connecting the land masses
in which they are found. In favor of a former land connection it may be
argued, and with justice, that while these species descend to the sea, the
probability that any pair of individuals should migrate from Cape Horn
to New Zealand or vice versa is highly improbable. (This objection
loses some weight if they spawn in the sea, as is reported.) There are
no intermediate places that might be colonized and serve as new centers
of distribution. It may further be urged that these species could readily
have been distributed to their present homes by migration from stream to
stream along a continuous coast line or on a land wave moving from one
place to another. An obvious objection comes from the paucity of the
forms with this peculiar distribution. If there was a continental mass
connecting South America with New Zealand and Australia fit to be
inhabited by fishes, there must have been an abundant and diverse fish
fauna which has disappeared. If the antarctic continent depended entirely
for its existence on the evidence from the distribution of the fresh water
fishes, its existence would be very highly theoretical and precarious.
Concerning the distribution of Ga/axzas, Boulenger says :

“The fact that certain species of Ga/axzas live both in fresh water and
in the ocean suffices to explain the curious distribution of this small family
without the necessity of appealing to the existence in the past of an

2
EIGENMANN: FRESH WATER FISHES. 229

antarctic continent. It is probable that the Galaxiidaee were formerly
distributed as marine forms about the entire globe south of the tropic of
Capricorn and that certain species on adapting themselves entirely to a
life in freshwater have become localized in points as widely isolated from
each other as those in which we know them now.” See also the quota-
tion under Ga/axias, p. 271.

However, the evidence from other sources of a former land connection
has become conclusive, and I am of the opinion that during the submerg-
ence of large parts of Patagonia during the late Pliocene the formerly
abundant freshwater fauna became exterminated with the exception of
those forms that were indifferently freshwater or marine. —

The Petromyzonidz offer still another difficulty. There is no place on
the American continents between the Mexican plateau and Central Chili
that harbors any species of the family. The northern and southern species
belong to distinct genera. At least two of the South American genera

are peculiar, while two others are found in Australia and New Zealand.
* A comparison of the Atlantic and the Pacific regions of Patagonia shows
the two slopes to be largely inhabited by the same forms, a fact easily
explained by the character of the species (anadromous or indifferently
marine or aquatic) together with the continuous way by ocean and series
of streams emptying at intervals between the Rio Negro and Santiago.

A comparison of the species inhabiting the southern part of the region
from the Rio Santa Cruz south and the northern part from the Rio Santa
Cruz to the north shows us that the southern part of Patagonia is inhabited
by an extreme fauna limited to Galaxias, Aplochiton, Geotria and
E:xomegas.

To these are added on the north, on both sides, Percichthys, Diplomyste
and Pygidium, and on the west alone, Percilia, Nematogenys and Chetrodon.
On the east 4styanax and Gymnocharacinus appear before the La Plata
fauna proper is reached. They are outliers of the largest of the South
American families, the Chavacide.

BOUNDARY BETWEEN THE BRAZILIAN AND
PATAGONIAN REGIONS.

The definitive boundary between these regions has not been determined,
since little is known of the fauna of the Rio Negro and nothing from the
region about 500 miles wide extending north from a line connecting the

230 PATAGONIAN EXPEDITIONS: ZOOLOGY.

mouth of the Rio Negro with Santiago, Chili. No systematic collecting has
been done in the Rio Negro, but the few specimens recorded from this
region are, with the exception of 4styavax, Patagonian; on the other hand,
the few specimens recorded from just south of Buenos Aires are Brazilian.

The southern boundary of the Brazilian fauna has been extended on
the east by the La Plata~Paraguay river. The headwaters of this river,
the Parana and especially the Paraguay, are in contact with the southern
tributaries of the Amazon, and have opened a way for the southward
migration of equatorial forms like members of the Gywnotde, the Ser-
rasalmonine and the Characine, which have not succeeded in migrat-
ing so far south along the east coast, where there does not exist a
continuous water way. (See the chapter on southeastern Brazil.) The
distance which the La Plata-Paraguay has extended the tropical fauna
can be measured by comparing the fauna of Buenos Aires, which is com-
posed entirely of Amazonian types, with the fauna of the Rio Grande
at the boundary of the United States, whose mouth is 10° nearer the
equator and the Amazon than Buenos Aires, and only 6% per cent. of whose *
fauna is made up of equatorial types. This southward extension of the
equatorial fauna was also facilitated by the entire absence of competitors.
The entire fauna of the Parahyba near Rio de Janeiro is equatorial, while
only 15 +per cent. of the Rio Panuco fauna in the same latitude in Mexico
is equatorial, the rest being mostly intrusives from the north. Similar
irregularities in the distribution of species are introduced in other parts
of the globe, where a river has a north and south extent over different
climates or regions, as in the case of the Nile.

On the Pacific slope one genus of tropical lowland forms, Chezrodon,
reaches the southern limit of the Chilian plateau, and one mountain genus,
Pygidium, reaches as far south as Santiago, Chili, and it is continued still
further south as Hafcheria. The southern boundary of the northern fauna
is here, so far as records show, the desert of Atacama. The southern
boundary of the Brazilian fauna seems to be determined by climate and
the means of ready migration.

The factors determining the northward extent of the Patagonian fauna
are not so apparent. With the exception of Geofvia and Exomegas, no
Patagonian forms reach so far north as Buenos Aires on the Atlantic side ;
on the Pacific side the fauna in the latitude of Buenos Aires is still dis-
tinctly Patagonian.

EIGENMANN: FRESH WATER FISHES. 231

SYNOESIS OF THE -PRESH-WATER. DISHES ,OF
PATAGONIA.

Class l, MAKSIZFOBRANCHIL

= Marsipobranchi Bonaparte, Nuovi Annali delle Sci. Nat. t. 2, 1838 ;
Miiller, Abhandl. K. Akad. Wiss. Berlin, 1844.

< Myzontes Agassiz, Contrib. to Nat. Hist. of U. S., I, 187, 1857.

= Dermoptert Cope, Proc. Acad. Nat. Sci. Phila. 1868, 256.

= Cyclostoma Gegenbaur, Grundriss der vergl. Anat. 577, 1870.

= Cyclostomata Schmidt, Handb. der vergleich. Anat. 6, Aufl. 259, 1872.

= Marstpobranchiates Gill, Arrangement Fam. Fishes, IX, 25, 1872.
Skeleton cartilaginous ; skull imperfectly developed, not separate from

the vertebral column ; no jaws, no limbs, no shoulder girdle, no innomi-

nates ; no branchial arches, gills in sacks, 6 or more; nares single, median ;

mouth suctorial, subcircular; no bulbus arteriosus; alimentary canal

straight, simple, without coecal appendages or pancreas; an abdominal

pore ; development with a metamorphosis.

Order) ELEY PEROAR EI.

= Hyperoartia Miller, Abhandl. k. Akad. Wiss., Berlin, 1844.
Eel-shaped ; nasal duct a blind sack, not penetrating the palate ; gill-

opening close behind the head, communicating with a common branchial

passage which opens into the pharynx; intestines with a spiral valve.

Family I. PE TROMYZONIDAZ:.

= Lampredini Rafinesque, Indice d’Ittilog. Siciliana, 1810.
= Petromyzonide Bonaparte, Nuovi Annali delle Sci. Nat. II, 133, 1838 ;
Jordan & Evermann, Bull. U. S. Nat. Mus. 47, 8, 1896.
= Petromyzontide Girard, Expl. and Surv. for R. R. Route to Pacific
Ocean, X, Fishes, 376, 1858.
General distribution: Temperate regions in both the northern and south-
ern hemispheres, anadromous or landlocked.
Body eel-shaped, subcylindrical anteriorly, compressed behind; the
mouth nearly circular, suctorial, usually armed with horny teeth, or tooth-
like tubercles which are simple or multicuspid, resting on papillz ; those

232 PATAGONIAN EXPEDITIONS: ZOOLOGY.

immediately above and those immediately below the cesophagus more or
less specialized ; eyes developed in the adult; gill-openings 7, arranged
in a row along the sides of the ‘chest’; nostril on top of head just in
front of the eyes; lips present, usually fringed; dorsal fin more or less
deeply divided by a notch, the posterior part commonly continuous with the
anal around the tail ; intestines with a spiral valve ; eggs small.

Key To THE SoutH AMERICAN GENERA OF PETROMYZONID®.

a. Supraoral lamina median and undivided ; anus under the anterior part of the second dorsal ;
second dorsal fin distinct from the caudal.

6. Lingual teeth distinct and conic; supraoral lamina transversely extended, quadricuspid ;

infraoral crescent-like, sinuous or denticulate on the edge; discal teeth numerous and

in divergent series. 1. Geotria.
66. Lingual teeth three, standing on the same base, pointed and curved, the median smallest ;

discal teeth in concentric series, the outer containing the largest teeth, about 24 on each
side. 2. Exomegas.

aa. Two lateral supraoral laminz entirely separate, triangular, each with three conic teeth ; infra-
oral lamina crescent-shaped, with about nine acute conical cusps; discal teeth in somewhat
distant series, radiating from the center, those of a series more or less confluent ; lingual
teeth serrate, in two pairs; anus under the posterior part of the second dorsal ; posterior
dorsal continuous with the caudal. 3. Caragola.

i. GEOTRIA, Gray:
(Plate XXX.)

> Geotria Gray, Proc. Zool. Soc. London, 1851, 238, pl. IV, fig. 3, and pl.
Viid. List Spec; Fish, Brit. Mus. 1,142, pl..a, ig.3,.and pl. 2.1851
(australis).
> Velasia Gray, Proc. Zool. Soc. London, 1851, 239, pl. IV, fig. 4; id.
List Spec. Fish. Brit. Mus. I, 143, pl. 1, fig. 4, 1851. (chdlensis).
= Velasta Gimther, Cat. Fish. Brit. Mus. V, 508; 1870; Gill, Proc. U.S.
Nat. Mus. V, 524, 1882 (chzdenszs).
Macrophthalmia Plate, Sb. Ges. Naturf. Fr. Berlin, 1887, 137 (chzlensis).
Type: Geotria australts Gray.
Distribution: Temperate regions of the southern hemisphere.
Distinguished from Avomegas by having distinct lingual teeth.
Plate distinguishes between the three species of this genus as follows:
a. Labial teeth numerous and close together.
6. Outer lingual tooth very large, with 2 large and 1 median short point ona low base;

two small inconspicuous inner lingual teeth. Supraoral plate plainly denticulate ; a small
gular sack in the ripe 9. Australia, New Zealand and South America. 1, chilensis.”

EIGENMANN: FRESH WATER FISHES. 233

66. Outer lingual tooth with 3 uniform points; two larger, inner, lingual teeth each larger
than the outer; supraoral plate smooth or denticulate; no gular sack. New Zealand,
Tasmania. stenostomus.'

aa. Labial teeth wide apart ; outer lingual tooth large with two large points on a high base; 2
inconspicuous lingual teeth ; supraoral plate smooth or nearly smooth ; an enormous gular
sack even in immature specimens. Australia and South America. 2, australs,

1. GEOTRIA CHILENSIS (Gray).
' (Plate XXX, Figs. 1 and 1a.)

Velasia chile.sts Gray, Proc. Zool. Soc. London, 1851, 239, pl. iv, fig. 4
(Chili, in fresh water) ; id. List. Spec. Fish. Brit. Mus. 143, pl. 1, fig.
A, Eegu; id; Ann, & Mag. Nat," Bist. X1H,63) 18545 Philippi,
Wiegm. Arch. 1863, 207. pl. x, fig. a (Valdivia).

Geotria chilensis Giinther, Cat. Fish. Brit. Mus., XIII, 509, 1870; Hutton,
Trans. New Zealand Inst., V, 271, 1872 and XIII, 216, 1875; Eigen-
mann & Eigenmann, Proc. U. S. Nat. Mus., XIV, 24, 1891; Gill,
Mem. Nat. Acad. Sci. Washington, VI, 110, 1893; Berg, An. Mus.
Nac. Buenos Aires, IV, 122, 1895 (Rio de la Plata); Delfin, Cata-
logo de los Peces de Chile, 13, 1901. Plate, Zool. Jahrb. Suppl. V,
660, 1902 (Australia; New Zealand; Valdivia; Puerto, Montt.).

Macrophthalmia chilensts Plate, Sb. Ges. Naturf. Fr. Berlin, Oct. 19,
1897, 137. (Source of the Rio Maullin from the Lago Llanquihui.)

Flabitat: Streams of Chili and Argentina; New Zealand and Australia.

“Outer lobes of the maxillary dental lamina broad with a sharp convex
edge, inner narrow and pointed; mandibular lamina crescent-shaped, with
numerous obtuse points; suctorial teeth in numerous series, close
together; a series of larger, broad, scale-like teeth round the mandibulary
‘lamina; suctorial disk not dilated, circular; first and second dorsal fins
widely separate from each other; back greenish; sides and abdomen
silvery.’”’ Giinther.

No specimens of this species were obtained by Hatcher, nor has it been
actually reported from any of the streams of Patagonia. Its presence in
New Zealand, Chili and La Plata makes its occurrence in Patagonian waters
certain. The fact that it has. not been taken is easily accounted for by
its anadromic habit. So few specimens of the lampreys are known that
nothing is definitely settled concerning the generic or specific identity.
It is not improbable that Berg’s Geotria chilensts is in reality an xvomegas.

'Plate XXX, fig. 2.

234 PATAGONIAN EXPEDITIONS: ZOOLOGY.

2. GEOTRIA AUSTRALIS Gray.
(Plate XXX, Figs. 3 and 32.)

Geotria australis Gray, Proc. Zool. Soc. 1851, 238; id. List Spec. Fish
Brit. Mus. 142, pl. ii; 1851 (Inkarpinki River); Giinther, Cat. Fish.
Brit. Mus. VIII, 508, 1870; Plate, Zool. Jahrb. Suppl. V. 668, 1902
(Valdivia, Chili; King George’s sound, Australia).

Habitat: Australia and Chili.

No specimens of this species were obtained by Hatcher. The follow-
ing is a description based on the type specimen.

“Skin of the throat very lax, forming a large pouch; maxillary lamina
thin, crescent-shaped, with four sharp teeth, the middle pair of which are
only half as broad as the outer. Mandibulary lamina very low, slightly
sinuous; suctorial teeth in numerous series, rather distant from one
another, aniscuspid; only those nearest to the mouth somewhat larger,
the others small; only one transverse series of very small teeth between
the mandibular lamina and the posterior lip, which is beset with numer-
ous broad, leaf-like fringes, as in the remainder of the margins of the
disk; suctorial disk subtriangular with the lateral lobes very broad; first
and second dorsal fins rather widely separated from each other.”’ Giinther.

2. EXOMEGAS Gill.

E‘xomegas Gill, Proc. U. S. Nat. Mus., V, 524, 1882 (macrostomus).

Type. Petromyzon macrostomus Burmeister.

Giinther (Cat. Fish. Brit. Mus. VIII, 505, 1870) was the first to recog-
nize the aberrant nature of this genus. He says: “ The following species
would appear to be the type of a distinct genus, but it will be better to
leave its creation to an ichthyologist who will be able to characterize it
from an autopsy of examples.” The genus was named by Gill, Zc. It
is not improbable that it is synonymous with Geofva. Smitt says on this
point:

‘Cette structure de la ventouse et de la bouche, comparée aux descrip-
tions et aux figures des Geotvia et des Velasza chez les auteurs, ne nous
donne aucun signe de divergence de forme mais bien de développement.
C’est un plus-ou-moins du développement des cartilages et de la sub-
stance cornée du méme type. Si l’on place les espéces décrites dans

EIGENMANN: FRESH WATER FISHES. 235

l'ordre suivant: Velasza chilensis, Geotria australis, Exomegas macrosto-
mus, et enfin notre exemplaire, on voit une série continue de développe-
ment de ce plus-ou-moins. Les termes de cette série, s’ils se trouvent
constants a des différents locaux, peuvent naturellement étre regardés
comme des espéces distinctes ou des formes locales—c’est la une question
de convenance —mais certainement ils ne méritent pas d’étre des types
génériques.”’

Larve very similar to those of Petromyzon but with more than 80
preanal myomeres ; a triangular nasal opening; dentition simple, a lingual
tooth with 3 points, of which the median is very small; suctorial teeth
radially arranged, papillary; no median, supraoral or infraoral teeth;
mouth bordered by cutaneous lamina; dorsals remote.

Distribution of the Species : Gallegos to Rio de la Plata.

a. Back rounded ; a gular pouch. 3. macrostomus.
aa. Back with a flattened area bordered by two dermal folds, merged behind and continued in
front on the snout; no gular pouch. 4. macrostomus gallegensis.

3. ExomEGAS MACROSTOMUS (Burmeister).

Petromyzon macrostomus Burmeister, Anal. Mus. Buenos Aires, 1868;
Act.Soce:-Paleont: XXXVI; Gunther, Cat. Fish, But. Mus: VIII,
506, 1870.

Exomegas macrostomus Gill, Proc. U. S. Nat. Mus. V, 524, 1882; id. Sci-
ence X MIME 204 Jan.. 19; 1694: 1d. Proc. U. S.Nat. Mus; XVII,.190;
1894.

Geotria macrostomus Berg, Anales del Museo de La Plata I, 2, plate 1,
1893 (Island of Flores, near Montevideo).

This species, if distinct from the variety gal/egensis, is known from but
two specimens; one, 40 cm. long, described by Burmeister was found
alive in the Calle de Mexico in Buenos Aires in 1867, immediately after
a heavy rain; the second, 44 cm. long, was obtained in 1890 and was
recorded by Berg. If distinct from the Gallegos specimens, this species
is extralimital.

‘‘No teeth in the circumference of the mouth; outer row of the 10-11
concentric series largest, about 24 on each side; gular sack large.”” Berg.

236 PATAGONIAN EXPEDITIONS: ZOOLOGY.

4. EXOMEGAS MACROSTOMUS GALLEGENSIS (Smitt).
(Plate XXX, Figs. 4 and 4a.)

Geotria macrostoma gallegensis Smitt, Bihang Till K. Sv. Vet. Akad.
Handl. Band 26, afd. iv, No. 13, 23 (Rio Gallegos and its tributaries,
Rio Ruben and Rio Turbio).
This variety is known from a single adult individual and several larve.
It differs from szacrostomus chiefly in the flat dorsal region.

3. CARAGOLA Gray.

< Caragola Gray, Proc. Zool. Soc. London, 1851, 239, pl. iv, fig. 5; id.
List Spec. Fish Brit..Mus., 1) 143;‘plx, figsas; 1851 ;-Gill, Proeal
S. Nat. Mus., V, 525, 1882 (/apicida).

<Mordacia Gray, Proc. Zool. Soc. London, 1851, 239, pl. iv, fig. 6; id.
List Spec. Fish Brit. Mus., I, 143, pl. f, fig. 6, 1851.

= Mordacia Ginther, Cat. Fishes Brit. Mus., VIII, 8, 507, 1870; Plate,
Zool. Jahrb. Suppl., V, 1902, 654.

Type -Caragola lapicida Gray.

This genus is considered but a subgenus of Geofvia by Smitt, as a dis-
tinct genus by Giinther, and the type of a distinct family by Gill.

Second dorsal continuous with the caudal ; two pairs of serrated lingual
teeth; maxillary dentition consisting of two triangular groups, each with
three conical acute cusps.

Plate considers three species as valid which he distinguishes as follows :

a. Outer lingual tooth a broad transverse plate with numerous points; eyes small.
6. Wreath of inner labial teeth closed in the middle in front with three single pointed teeth.
5. mordax.
66. Wreath of inner labial teeth closed in the middle in front with one single pointed tooth.
6. lapicida.
aa. Outer lingual tooth triangular, with but two retrorse hooks. Eyes relatively large.
7. acutidens.

5. CARAGOLA MORDAX (Richardson).
(Plate XXX, Fig. 5.)

Petromyzon mordax Richardson, Voy. Erebus and Terror, Fish., pl. 38,
1845 (Tasmania).

Mordacia mordax Gray, List Spec. Fish. Brit. Mus., 144, pl. 1, fig. 6, 1851
(copied) ; Giinther, Cat. Brit. Mus., VIII, 507, 1870 (Valparaiso; Tas-

EIGENMANN: FRESH WATER FISHES. 237

mania); Steindachner, Zool. Jahrb. Suppl., IV, 334, 1898 (Tumbes) ;
Delfin, Catalogo de los Peces de Chile, 13, 1901 (Rio Tucapel and
Puerto Montt); Plate, Zool. Jahrb. Suppl., V, 654, 1902 (Murray
River). i

Habitat: Coastwise streams of Chili, south to at least Puerto Montt
(Tasmania).

‘‘Mandibular lamina crescent-shaped, with about nine acute conical
cusps, three of which are larger than the others; suctorial teeth in some-
what distant series, radiating from the center; the teeth of the series be-
tween the mandible and posterior lip as numerous as those of the other
series but rather more confluent; anterior labial teeth converging and con-
fluent behind, each tooth of the posterior pair like one half of an elon-
gated oval; suctorial disk elliptic, with a free lip behind.”” —_Giinther.

6. CARAGOLA LAPICIDA Gray.
(Plate XXX, Fig. 6.)

Caragola lapicida Gray, Proc. Zool. Soc., 1851, 239 (Valparaiso). List
Spee Fish Brit: Mus, 143, plz 1; fic. 5:

Mordacia lapicida Plate, 1. c., 654 (Tumbes, Chili).

Petromyzon anwandtert Philippi, Wiegm. Arch. 1863, 207, tab. X, fig. 6
(Valdivia).

Flabitat : Chili.
7. CARAGOLA ACUTIDENS (Philippi).
(Plate XXX, Fig. 7.)

Petromyzon acutidens Philippi, Ann. & Mag. Nat. Hist., XVI, 1865, 221."
Mordacta acutidens Plate, |. c. 657.
Habitat: Chili.

Class 11. PISCES.

<Pisces Artedi; Genera Piscium 1738.
<Pisces Linnzeus Sytema Nature ed. X, I, 239, 1758.
=Pisces Miiller, Abhandl. K. Akad. Wiss. Berlin, 1844.

' This is the translation of an article credited to ‘‘ Wiegm. Arch. 1864, 107.” It is not found
in the place cited.

238 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Superorder OSTARIOPHYSI.

Ostariophyst Sagemehl, Morph. Jahrbuch X, 1-120, plates I and II.

Anterior four vertebrz codssified, much modified ; frequently a chain of
four ossicula auditus forming the Weberian apparatus attached to the
anterior vertebree and connecting the air-bladder with the auditory appa-
ratus ; air-bladder connected with the alimentary canal by a duct. Parie-
tals separating the frontals from the supraoccipital or fused with the latter.
Shoulder girdle suspended from the skull by a long post temporal ; meso-
coracoid arch well developed ; opercle well developed. Ventrals abdomi-
nal, without spines.

The fishes belonging to the families making up the order or superorder
of Ostariophysi are the dominant forms in the fresh waters of the world.
They fall naturally into two groups or orders:

1. Nematognathi with no subopercle or symplectic bone; the supra-
occipital and parietals codssified ; the maxillary usually without teeth and
reduced to form the core of a variously developed barbel ; no scales.

II. Plectospondyhi, a less solid group; with subopercle and symplectic
bones and with a well formed maxillary and frequently with teeth.

Of these the Nematognathi are found both in the northern and south-
ern hemispheres and have also become inhabitants of the sea.

The Plectospondyli fall into three groups of familes, those related to
the Cyprinidae, confined chiefly to the northern hemisphere but overflow-
ing into northern Africa, those related to the Chavracide inhabiting Africa
and South America but overflowing into Central America and Mexico, and
those related to the Gymnotide and confined to tropical America.

Of the 1,200 or more species of fresh-water fishes recorded from tropical
America, about 1,000 belong to the Ostariophysi; in round numbers about
500 of these belong to the Nematognathi and another 500 to the Plecto-
spondyli. Of the African fresh-water fishes a much smaller per cent. are
Ostariophysi.

Of this vast array of American species, very few are found within the
boundaries of the region under consideration and most of these barely
cross its northern border.

EIGENMANN: FRESH WATER FISHES. 239

Order Il. NEMATOGNATHI.

Nematognatht Gill, Fishes of East Coast, 1870.

The following opinions expressed several years ago in a then obscure
and now extinct journal are still, in part, pertinent and are repeated.with a
few verbal changes.

Not many fossils of this order have been found, and those which have
been discovered do not help us materially in determining the interrela-
tions of the higher groups.

The date of the differentiation of the orders of the Ostariophysi is
uncertain. Silurinz, Bagarinze and Pimelodinz were separated as early
as the beginning of the Tertiary, and as these subfamilies are quite remote
from the most primitive living nematognathid, the Nematognathi must
extend into the Secondary period.

The Nematognathi reach their highest development in the neo-tropics,
where they constitute forty per cent. of the entire freshwater fish fauna.
Most of the families, sub-families and genera now inhabiting this region
have undoubtedly originated here. In most of the families the maxillary
bone is a mere vestige and serves only as a support for a highly special-
ized dermal appendage, the maxillary barbel. So greatly has this been
modified, even in its development, that Ryder seriously questioned whether
the basal bone of the maxillary barbel in the North American cat-fishes
is in reality the vestige of a maxillary. No doubt need, however, be
entertained on that score, since in Dzflomyste the basal bone of the short
primitive maxillary barbel is a functional dentiferous maxillary.

The comparative development of the maxillary is, then, an excellent
guide to the determination of the rank any family is to occupy in the sys-
tem.

Other structures valuable in this respect are the different barbels so
highly developed in some forms. Of the development of the barbels of
Lctalurus albidus, Ryder says :

‘The remarkably developed barbels of the embryos of the species make
their appearance very early, especially the maxillary pair; these appear on
the second day. . . . The barbels on the lower jaw do not appear till the
fourth day of the development is completed. . . . The last of all to be
developed is the nasal pair, . . . (which) does not appear until the seventh
day.”

240 PATAGONIAN EXPEDITIONS: ZOOLOGY.

In Diplomyste only short maxillary barbels are developed, and, as stated
above, the maxillary is dentiferous and forms the mouth border above. It
therefore has its barbels less specialized than the remaining families, while
its maxillary is more generalized and resembles that of other fishes. The
Diplomystide may therefore be considered the most primitive of the
Nematognathi. At present this family is restricted to the fresh waters of
central Chili and northern Patagonia.

The next important factor to be considered is the relative development
and specialization of the air-bladder and Weberian apparatus. In the
Bunocephalide and the Siluride, with the exception of the Agenezosine,
the air-bladder is large and lies below the codssified vertebrae. In the
remaining South American families the air-bladder has been split into
lateral halves, and, with the corresponding bones of the Weberian appa-
ratus, has been enclosed in a bony capsule.

In the Bunocephalide there are no indications of any approach to the
enclosed air-bladder, and indeed this family shows little resemblance to
the remaining families. It was early differentiated from the descendants
of the Diplomystide, but not before the maxillary had become quite ves-
tigial. Two subfamilies have become differentiated ; they differ chiefly in
the number of vertebre in the tail and the number of-anal rays. This
family inhabits the La Plata basin, the Amazon and the South American
rivers to the north of it.

The cosmopolitan family, S7/urid@, is composed of a large number of
subfamilies, seven of which are American, or have representatives in
America. They vary-considerably from the most generalized Tachisurine,
which most resemble the DiAlomystide, to the Doradinz and Ageneiosine,
in which last we observe an approach to the conditions obtaining in the
families with closed air-bladders.

As is intimated above, the Tachisurinz are the lowest of the living
Siluvide. The members of this subfamily are mostly marine, and from
it have, in all probability, been derived the remaining subfamilies. Some
members of the Tachisurinz have only the maxillary barbels developed,
while in others the mental barbels have been added.

From the Tachisurinz have unquestionably been derived the Callo-
physinz and Pimelodinz, which differ from each other in dentition. Both
these subfamilies are neotropical in their distribution, the Callophysinz
being found in the Amazon and northward, the Pimelodinz in the whole

EIGENMANN: FRESH WATER FISHES. 241

region from the La Plata to Mexico. One of the chief features modified
in passing from the Tachisurinz to the Pimelodinz is the space between
the anterior and posterior nares. In the Pimelodinz the barbels obtain
their greatest specialization. The nasal barbels are, however, not developed
before the Bagrinze are reached, to which all the North American species
belong.

From the Pimelodinz have also been derived the Doradinz, in which
dermal ossifications first make their appearance, although minute plates
are noticed along the anterior portion of the lateral line of Platystoma-
tichthys sturio (Pimelodine).

In the Doradinz and the succeeding subfamilies of Sz/urde the Weber-
ian apparatus varies greatly and the whole group seems to be in a state
of unstable equilibrium.

In the Ageneiosinze appears for the first time the great reduction and
splitting into lateral halves of the air-bladder. From the Ageneiosinz
were undoubtedly directly derived the Hypophthalmide.

From the primitive Ageneiosinz and Auchenipterine have evidently
been derived, as a lateral offshoot, the Pygzdde, of which the Cetopsinz
retain the most ancestral traits. In this family nasal barbels again appear.
Some members of Stegophelinz have lost the habit of free swimming and
live as commensals or parasites in the gill-cavity of large species of the
Seluride.

The three remaining families were, by Dr. Giinther, united in one of
his inferior groups, Hypostomatina. They have little external resem-
blance to the Sz/uride.

The Caliichthyide have two series of smooth plates covering the sides,
while the Lovicaritde have several series of rough plates. The teeth and
dentiferous bones of the latter have also been considerably modified. The
Argide are dwarfed forms inhabiting high mountain streams. They are
naked but otherwise similar to the mailed Loricariide.

Family I. D/PLOMYSTIDZ.

= Diplomystde Eigenmann, Zoe, Vol. I, 14, 1890.

Air-bladder well developed, simple or with transverse constrictions,
lying free in the abdominal cavity. Mouth terminal, teeth incisor-like and
villiform ; intestines short, arranged in longitudinal folds. Body naked.
Diaphragm membranaceous. Tip of scapular process touching basioccip-

242 PATAGONIAN EXPEDITIONS: ZOOLOGY.

ital. Dorsal short, confined to the abdominal portion of the vertebral
column. Opercle well developed and movable ; adipose fin present ; occip-
ital process not forming a bony bridge from the occipital to the dorsal
plate. Caudal vertebrae not compressed, the neural spines simple, spine-
like. Maxillary well developed, provided with teeth and forming the sides
of the jaw; nares approximated ; gill membranes somewhat joined, but free
from the isthmus. A single genus.

4. DIPEOMYSTE,

Diplomyste Dumeril, Ichthyol. Analyt., 487, 1856 (papz//osus).
Diplomystes Bleeker, Nederl. Tijdschr. Dierkunde, I, 92, 1863 (fafpellosus).
Diplomystax Giinther, Cat. Fish. Brit. Mus., V, 180, 1864 (fafzllosus).

The distinctive character of this genus is the development of the
maxillary bone. It forms the sides of the mouth and carries a very nar-
row band of teeth. No mental barbels; maxillary barbels thick, com-
pressed at the base. Occipital process not visible externally; no bony
orbit. Skull with two nearly parallel median ridges above. Fontanel
extending to the base of the occipital process, with a rather broad inter-
ruption behind the eyes. A single species.

Habitat: Chili, near Valparaiso and Santiago.

8. DIPLOMYSTE PAPILLOsSUS (Cuvier & Valenciennes).
(Plate XXXI, Figs. 1, 1a and 14.)

Molina “Hist. Nat. Chile, 199, No. 9.”

2? Selurus chilensis Linnzus, Syst. Nat. Ed. XIII, 1359, 1788 (Chill).

?? Pimelodus chilensis Lacépéde, Hist. Nat. Poiss., V, 114, 1803 (Fresh-
waters of Chili).

Arius papillosus Cuv. & Val. Hist. Nat. Poiss., XV, 118, pl. 431, 1840
(Valparaiso; Santiago); Gay, Hist. Chile, II, 305, pl. fig. 1, 1848
(Chili) Philippi, MB. Ak. Wiss. Berlin, 1866, 710 (Chili).

Diplomystax papillosus Giinther, Cat. Fish. Brit. Mus., V, 180, 1864
(Chili).

Diplomystes papillosus Eigenmann & Eigenmann, Proc. Cal. Acad. 2d
ser. vol. I, 1888, 149 (Rivers of Santiago).

Arius carcharias ‘‘Leybold, Anales de la Universidad de Chili, 1859
(Chile)”’; Philippi, MB. Akad. Wiss. Berlin, 1866, 711 (Chill).

EIGENMANN: FRESH WATER FISHES. 243

Arius villosus Philippi, 1. c., 712 (Santiago).
Arius squalus Philippi, |. c., 713 (Santiago).
Arius micropterus Philippi, |. c., 713 (Santiago).
Arius synodon Philippi, 1. c., 713 (Santiago).

Flabitat: Central portion of Chili.

The numerous species of Philippi seem to be based on individual differ-
ences. The seventh dorsal ray is always divided to near its base, and in
one of the specimens examined by me there are eight dorsal rays.

Slender, terete forward, compressed toward the tail; the width, below
the dorsal spine, equals the depth. Head short and blunt, somewhat
pointed in young, the profile strongly decurved in front; head entirely
covered with a layer of muscle and skin, the surface of the bones irregular,
rugose, the interorbital portion flattish, postorbital portion with two nearly
parallel median ridges, fontanel between the two median ridges, extend-
ing from the occipital process to in front of the eyes, with a broad bridge
behind the eyes. Occipital process a deep but short crest.

Eye small, circular, 3 in snout, little more than 1 in the interorbital, 3
in the interocular, 7, in the head.

Nostrils close together, with a membrane around the entire margin
forming an 8-shaped figure, the portion of the membrane separating the
nasal openings highest.

Maxillary barbel broad and flattened at its base, scarcely reaching the
base of the pectoral.

Snout sometimes long and pointed, usually rounded, projecting con-
siderably beyond the lower jaw. Lips thick and strongly papillose.
Maxillary bones reaching to below the eye, 22 in the head. The anter-
ior half of the maxillary consists of a slender peduncle bearing two series
of teeth; the posterior portion is flattened and about three times as wide
as the anterior portion. Premaxillaries with a crescent-shaped patch of
teeth, the outer four or five series compressed, incisor-like ; teeth of the
inner series very much smaller, not incisor-like. Vomer with two oval
patches of large conical teeth, the patches becoming united with age.
Teeth of the lower jaw compressed, incisor-like; the band deep in front
tapering rapidly to the rictus.

Gill-membranes usually separate to below the eye, entirely free from
the isthmus. Gill-rakers 5+ 8. Pseudobranchiz well developed, not
covered by a membrane.

244 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Skin covered with minute cirri or papilla, which are especially con-
spicuous over the humeral region and head, the tail sometimes smooth.

Distance of dorsal fin from tip of snout 23-2 in the length, the spine
stout and smooth, 13-13 in head. Distance of the adipose fin behind
the dorsal about equal to the length of the adipose fin or one-half longer,
5-6 in the length.

Caudal emarginate, with numerous, short, accessory basal rays.

Free margin of the anal straight or slightly convex, the fifth ray highest,
about 14 in the head, rapidly decreasing to the last ray, which is about
half the height of the fifth.

Pectoral spine very strong and with a leathery prolongation ; height of
the spinous portion 13-2 in the head; outer margin of the spine smooth,
inner margin with strong, recurved teeth along its entire length.

Bases of all the fins, lower part of head, suborbital regions and opercles
and a median line on the back purplish; rest of body and fins yellowish
and under the lens dark.

Head 4-44; depth 43-54; Br. 7; D. I, 7 or 8; A. 9-12.

No specimens were collected by Hatcher and the above description is
based on the material in the Museum of Comparative Zoélogy at Harvard
University.

Family III. PYG/DI/DAZ.
> Siluvoidet trichomycteriformes Bleeker, Neder]. Tijdschr. Dierk. I, 112,
1863.
> Siluvide opisthoptere Giinther, Cat. Fish. Brit. Mus. V, 4, 1864.
>Siluvide branchicole Giinther, |. c.
> Trichomycteride Gill, Arrangement of Families of Fishes, 19, 1872.
=Pygidide Eigenmann & Eigenmann, Am. Nat. July, 1888, and South
Amer. Remo cone: 316, 1890.

Distribution of the species: East and west slopes from Colombia to
La Plata and along the Andes to Patagonia.

Air-bladder vestigial, part on either side of the coalesced vertebree and
enclosed in a capsule formed by the lateral processes of the coalesced
vertebre only ; the external orifice of the air-bladder capsule as in //yp-
ophthalmide, its anterior side closed. Adipose fin none; dorsal usually
on the caudal, sometimes on the abdominal portion of the vertebral
column ; anal short. Derm naked. Dorsal and pectoral spines scarcely
developed.

EIGENMANN: FRESH WATER FISHES. ; 245

But three genera of this family are known to occur within the limits of
the area under consideration; they are /ematogenys, Hatcheria and
Pygidium. Their relationship will be seen from the following key to all
the genera of the family.

a. Dorsal entirely in front of the ventrals; vomer with teeth; head compressed; anterior nares
almost labial. Gill-membrane broadly united with the isthmus. Eye almost entirely con-
cealed under the skin. A single maxillary barbel ; two pairs of mental barbels. Opercles
unarmed. (Cetopsine.)
6. Teeth all conical or incisor-like, those on the vomer in a single series.

Hemuicetopsis Bleeker.
66. Teeth on the premaxillary villiform, in a band, those on the vomer and on the mandible

incisor-like, in a single series. ; Cetopsis Agassiz.
666. Teeth on the premaxillary and on the mandible villiform, in bands, those on the vomer
in a single series. Pseudocetopsis Bleeker.
6666. Teeth all villiform, in bands. Paracetopsis Eigenmann & Bean.

aa. Dorsal above or behind the ventrals ; no teeth on the vomer.
c. Gill-openings broad, the gill-membranes almost free or forming a free membrane across
the isthmus. (Pyrgidine.)
d. Anal short ; eyes superior; mouth terminal; gill-membranes free or almost free from
the isthmus.
e. Ventrals present.
J. A single maxillary barbel ; opercles and preopercle unarmed.

g. Nasal barbels present; dorsal placed over the ventrals ; one pair of

mental barbels. Nematogenys Girard.
gg. Nasal barbels obsolete ; dorsal placed behind the ventrals ; two pairs
of mental barbels. Pariolius Cope.

if Two maxillary barbels; opercle and preopercle with osseous prickles ;
nasal barbels present.
h. Caudal peduncle broad; caudal with numerous accessory rays.
Pygidium Meyen.
hh, Caudal peduncle slender; caudal with very few accessory rays;
dorsal emarginate. Hatcheria Eigenmann,
ee. Ventral fins none; otherwise like Pygidium. Evemophilus Humboldt.
dd. Anal long, partly in front of the dorsal; head greatly depressed ; eyes infringing
on the upper and lower surfaces of the head; mouth inferior; gill-membranes
broadly joined to each other, free from the isthmus; opercle and preopercle with
osseous prickles ; a series of fine labial teeth. Zridens Eigenmann & Eigenmann.
cc. Gill-membranes confluent with the skin of the isthmus, the gill-opening a narrow slit in
front of the pectoral; opercle and preopercle armed with spines; vent far behind the
middle of the body ; anal short. (Stegophiline.)
7. Upper lip with several series of numerous small, movable teeth ; each jaw with sev-
eral series of minute teeth; mouth inferior.
j. Gill-membranes broadly united with the isthmus.

246 PATAGONIAN EXPEDITIONS: ZOOLOGY.

k. Caudal widely forked, the upper lobe produced ina filament ; a single max-

illary barbel. Pseudostegophilus Eigenmann & Eigenmann.
kk. Caudal emarginate.

7, A single maxillary barbel.
Zl. Two maxillary barbels.
kkk, Caudal rounded.
m. A single maxillary barbel. Stegophilus Reinhardt.
mm. Two maxillary barbels. Miuroglanis Eigenmann & Eigenmann.
jj. “Rimae branchiales confluentes, membrana branchiostega cum isthmo gulare
haud connexa;” caudal rounded; maxillary barbel single; no accessory

caudal rays. Acanthopoma Litken.
zt. No labial teeth ; teeth in the jaws in a single series.

n. Teeth pointed, in a single series in the intermaxillaries only ; mouth sub-inferior ;

a single maxillary barbel. Vandellia Cuvier & Valenciennes.
nn. Teeth broad, incisor-like in both jaws; caudal forked ; two maxillary barbels.

Henonema Eigenmann.
Homodietus Kigenmann.

Paretodon Kner.

5. NEMATOGENYS Girard.

Nematogenys Girard, Proc. Acad. Nat. Sci. Phila, VII, 1854, 198
(cmervimes).

Type: Trichomycterus tnermts Guichenot.

Dorsal fin placed over the ventrals, without a spine. Fontanel ex-
tending to base of occipital process, interrupted above the posterior margin
of orbit. Opercle and preopercle unarmed. A single barbel on each
maxillary and one pair of mental barbels.

Habitat - Central Chili.

g. NEMATOGENYS INERMIS (Guichenot).

(Plate XXXI, Fig. 2, and Plate XXXII, Fig. 2.)

Trichomycterus tnermzts Guichenot, in Gay Hist. Chile Zool., II, 312, pl.
1m, ies 2) 1648 (Chil).
Nematogenys tnermis Girard, Proc. Acad. Nat. Sci. Phila. VII, 1854,
198; id. U. S. Naval & Astron. Exped. 240, pl. xxxii, 1855 (Rio de
Maypu near Santiago); Philippi, MB. Ak. Berl., 716, 1866 (Chili) ;
Giinther, Cat., V, 272, 1864 (copied); E. & E. Proc. Cal. Acad. Sci.,
2d Ser., II, 50, 1889 (Curico; Santiago); id. South Am. Nematog-
nathi, 322, 1890 (Curico; Santiago).
Nematogenys nigricans Philippi, 7 ¢., 716 (Chill).
Nematogenys pallidus Philippi, 2 ¢c., 716 (Chili).
Habitat: Fresh waters of Central Chili.

EIGENMANN: FRESH WATER FISHES. 247

N. nigricans and pallidus seem to be color varieties of WV. zxermis.

Tail compressed, head depressed, the caudal peduncle about as deep
as the body. Head entirely covered with soft skin, little longer than wide,
its depth 13-2 in its length. Eye small, superior; interocular width little
less than length of snout; orbit without a free margin. Mouth wide, ter-
minal, each jaw with a rather deep band of villiform teeth.

Gill-membranes narrowly joined to the isthmus.

Anterior nasal opening with a barbel which is about 13 times as long
as the eye; a series of pores extending from it backward below the eye;
a prominent pair of pores between the eyes.

Pectoral pore minute, above and behind the pectoral spine.

Origin of dorsal fin one-fifth nearer tip of snout than to base of middle
caudal rays in specimens .12 meter long, one-fifth nearer base of caudal
rays than to tip of snout in a specimen .26 meter long. ©

Margin of all the fins rounded. Caudal with numerous accessory rays,
the middle caudal rays 1-13 in the head. Origin of ventrals below or
slightly behind the first dorsal ray. Inner margin of pectoral spine ser-
rate, its lower surface spiny.

Light brown mottled with darker, a series of about 5 light areas along
the lateral line; fins speckled.

Head 43-34; depth 7-6; D. 10; A: 11; Br. 10=12.

This species does not belong to the tropical American fauna. Nor can
it technically be referred to the Patagonian fauna, since no specimens have
been recorded south of Central Chili. It occupies intermediate ground
and possibly touches the Patagonian region on its northwestern frontier.

This account is based on material in the Museum of Comparative Zodélogy.’

"PYGIDIUM Meyen.

Trichomycterus Valenciennes, Humboldt, Rec. d’Obs. Zool. et Anat. II, 348, 1833 (#zgricans) not
Thrichomycterus Humboldt.
Thrychomycterus Cuv. & Val. XVIII, 485, 1846 (misquoted).
Thrichomycterus Girard, Proc. Acad. Nat. Sci. Phila., VII, 1854, 198; Girard, U. S. Nav
Astron. Exped. II, 242, 1855 (misquoted).
Pygidium Meyen, Reise, I, 475, 1835 (fuscum).
Type: Pygidium fuscum Meyen.
Habitat: Andes from Central Chili to Colombia; Callao Bay ; Amazon to Cudajas; South-
eastern Brazil ; Central Argentine Republic.
The reasons for using the name Pygidivm instead of the older Trichomycterus were given
by Eigenmann & Eigenmann, S. A. Nematognathi, 325. Inasmuch as some naturalists still per-
sist in using the name Z7yichomycterus, the statement made by us may be repeated :

248 PATAGONIAN EXPEDITIONS: ZOOLOGY.

6. HATCHERIA Eigenmann gen. nov.
Type. Hatcheria patagoniensis sp. nov.
Dorsal long, emarginate ; caudal with very few accessory rays, which
are graduated ; anal below the posterior half of the dorsal; caudal
peduncle very slender ; caudal emarginate, the lobes rounded.

SPECIES OF HATCHERIA.

a. D. 21; origin of dorsal equidistant from occiput and tip of caudal ; anal inserted below the ninth
dorsal ray and terminating under the seventeenth ; ventrals extending to vent. 10. macrei.
aa. D. 15; origin of dorsal equidistant from tip of snout and tip of caudal; last ray of anal under
last ray of dorsal; distance of anal from base of caudal 3? in the length. 11. maculata.
aaa. D. 15; origin of dorsal equidistant from tip of caudal and posterior nares; gill openings
not extending forward to eye; last anal ray under last dorsal ray; distance of anal from
base of caudal about 44 in the length. 12. patagoniensts.
aaaa. D. 14; origin of dorsal equidistant from tip of caudal and occiput, its last ray over fourth
ray of anal; gill openings extending forward to below eye, the membrane free from the
isthmus ; distance of anal from base of caudal 5 in the length. 13. areolata.

10. HATCHERIA MACR41 (Girard).

(Plate XXXII, Figs 1, 1a and 1d.)

Thricomycterus macre? Girard, U. S. Naval and Astronomical Expedition
245, 1855 (Uspullata, elevation, 7000 feet).

The only claim for retention the name Zrichomycterus possesses is its distinctiveness from
Thrichomycterus Humb. = Eremophilus Humboldt. Twenty-two years before any species of
Trichomycterus Val. was known, the name Thrichomycterus was proposed by Humboldt as an
alternative for his Eremophilus, if future investigations should prove Evemophilus to be objection-
able. Upon proposing the name 7richomycterus Valenciennes states: ‘ Nous prenons pour le
nouveau genre le nom de 7richomycterus imaginé par M. de Humboldt,” so the names can hardly be
considered distinct, being either a misprint or a lapsus digiti. Valenciennes afterwards misquotes
himself, spelling the name Zhrychomycterus instead of Trichomycterus. Girard also misquotes
Valenciennes or quotes Humboldt correctly, using Thrichomycterus instead of Trichomycterus.

This genus has a wide distribution, extending along the Andes from Caracas to Chili on both
slopes. It is distinctly a mountain form, but it descends the Pacific slope to the coast at Callao
and the Amazon to Cudajas. It further occurs in the highlands of Guiana and Eastern Brazil.
There are in all about 35 species known. Some of these are found on both sides of the Andes,
but none have a wide distribution. On the contrary, the species seem to be local differentiations
of this widely distributed genus. Contrary to the general rule of distribution of South American
fresh-water fishes, that a genus with many species and a wide distribution has many representa-
tives in certain favorable localities, there is no place on record harboring more than one species
of this genus. We may expect the finding of many other species as restricted basins or isolated
valleys of the Andes and the temperate portions of South America are explored.

The genus is without doubt one of very long standing. In Central Chili and Patagonia it has
become differentiated into Hatcheria.

EIGENMANN: FRESH WATER FISHES. 249

Pygidium macret Eigenmann & Eigenmann, Proc. Cal. Acad. Sci. 2d
ser., II, 51, 1889 (Uspullatuo) ; id, South Am. Nematognathi 328,
1890 (Uspullatuo).

Elongate, rather compressd, especially backward. Head as broad as
long, snout rounded ; eye’ small, midway between tip of snout and end
of opercle ; none of the barbels reaching the gill-opening.

Gill-opening scarcely continued forward, joined to the isthmus for a
distance equal to half the width of the mouth.

Pectorals obliquely truncate, the first ray not produced; origin of dor-
sal some distance behind ventrals, equidistant from occiput and tip of
caudal; fourth or fifth dorsal ray highest, then gradually decreasing in
height to the last. Caudal emarginate, the upper lobe pointed, the lower
rounded ; anal inserted below the gth dorsal ray and terminating under
the 17th; ventrals inserted nearer tip of snout than to tips of middle cau-
dal rays, reaching tothe vent. Head 63 (7 in the total); depth6; D. 21.
10:

This account is drawn from a specimen (No. 8298) in the Museum of
Comparative Zodlogy. It came from Uspullatuo and is probably the type
of the species.

11. HATCHERIA MACULATA (Cuvier & Valenciennes).
(Plate XX XIII, Figs. 1, 1a and 12.)

Trichomycterus maculatus Cuvier & Valenciennes, Hist. Nat. Poiss.,
XVIII, 493, 1846 (Santiago) ; Guichenot in Gay, Hist. Chile Zool.,
i, 301, 1848: (Chili); Gunther, Cat) Fish. Brit. Mus) V,/273, 1864;
Girard, Proc. Acad. Nat. Sci. Phila., VII, 1854, 199 (Mapocho) ; Phil-
ippi, Mb. Akad. Wiss. Berl. 1866, Lae Delfin, Catalogo de los
Resces: de: Cinle) 20, 1907

Thrichomycterus maculatus Girard, in are U.S. Naval and Astronomical
Expedition 243, 1855 (Rio Mapocho).

Pygidium maculatum FEigenmann & Eigenmann, Proc. Cal. Acad. Sci.
2d ser., II, 51, 1889 (Rio Mapocho) ; id., South Am. Nematognathi
329, 1890 (Rio Mapocho).

Elongate, somewhat compressed; head as long as wide; caudal pe-
duncle long and slender. Eye small, midway between tip of snout and end
of opercle. Lips and lower surfaces of the head thickly covered with

250 PATAGONIAN EXPEDITIONS: ZOOLOGY.

warts. Gill-openings not continued forward to below the eye, the mem-
branes joined to the isthmus for a distance equal to one-third the width of
the mouth.

Pectorals rounded, the first ray not produced ; origin of dorsal in front
of the vent, but some distance behind the ventrals, equidistant from tip of
snout and tip of caudal, its last ray over the last ray of the anal. Caudal
long, truncate. Anal short and high, its height about equal to the length
of the caudal, its distance from the base of the caudal 33 in the length.
Origin of the ventrals equidistant from tip of snout and base of caudal,
their tips reaching beyond the vent.

Back and sides marbled with light and dark brown; fins pale, immacu-
late.

Head 53% (63 in the total); depth 73; D.15; A. 9. This account is
based on a specimen in the Museum of Comparative Zodlogy.

12. HATCHERIA PATAGONIENSIS Eigenmann sp. nov.
(Plate XX XIII, Fig. 2; and Plate XXXIV, Figs. 1 and 1a.)

Type. A specimen 118 mm. long. Rio Blanco, base of Andes. Lat.
47 30 -S.; long. 72 W. Hatcher; collector.

Cotypes, 8 specimens, 96-113 mm.

Elongate, slender ; head, 5; depth 8; caudal peduncle slender, its depth
nearly 3, in the head, about 4 in its length; upper maxillary barbel reach-
ing pectoral, lower to margin of interopercle, a broad lobe of skin join-
ing base of lower maxillary barbel to the lower lip; snout pointed, mouth
narrow, its width 34 in head, equal to interorbital; nasal barbels reach-
ing beyond eye; width of head but little less than its length; greatest
width of body behind the pectorals, 1.6 in the length of the head. Gill
opening not extending forward to below eye; origin of dorsal equidistant
from tip of caudal and posterior nares; base of dorsal equal to its dis-
tance from the caudal, its free surface emarginate, the anterior lobe rounded,
the posterior pointed; beginning of last third of dorsal not much more
than half as high as anterior lobe. Caudal moderate, emarginate, its lobes
rounded, .8 of the length of the head. Anal broadly rounded, its last ray
about under last ray of dorsal. WVentrals broad, their middle under origin
of dorsal, 14 in head, equal to height of anal. Base of pectoral hori-
zontal, closing edgewise to body, its lower part folded when appressed ;

EIGENMANN: FRESH WATER FISHES. 251

first ray sickle-shaped, slightly prolonged. Dark yellowish, more or less
regularly spotted with darker, dorsal and caudal and pectorals irregularly
blotched with black. D. (13) 15; A. 6.

Some of the cotypes are more robust in body; in one the anal is
blotched like the caudal; in some the spots form regular series along the
sides leaving lighter stripes between them.

13. HATCHERIA AREOLATA (Cuvier & Valenciennes).

(Plate XXXIV, Fig. 2.)

Trichomycterus areolatus Cuvier & Valenciennes, Hist. Nat. Poiss.,
XVIII, 492, 1846 (Coast of Chili); Guichenot, in Gay Hist. Chile,
Zool., II, 309, 1848; Giinther, Cat. Fish. Brit. Mus., V, 274; Delfin,
Catalogo de los Peces de Chile, 30, 1901; Philippi, Mb. Akad. Wiss.
Berlin, 1866, 714.

Pygidium areolatum Eigenmann & Eigenmann, Proc. Cal. Acad. Sci.,
2d ser., II, 51, 1889 (Rio Mapocho); id. South American Nematog-
nathi, 330, 1890 (Mapocho); Berg, An. Mus. Nac. Buenos Aires, IV,
143, 1895 (Coidola) ; Evermann & Kendall, Proc. U. S. Nat. Mus.
XXXI, 86, 1906 (Rio Comajo; tributary of Lake Traful; tributary
of Rio Limay).

Thrichomycterus maculatus Girard, in part, U.S. Naval and Astronomical
Expedition 243, 1855 (Mapocho).

Elongate, subterete. Lips and lower surfaces of head thickly covered
with small warts. Gill-openings continued forward to below the eye, the
membranes free from the isthmus. Upper maxillary barbels reaching to
the pectorals. Pectorals rounded, the first ray not prolonged ; origin of
dorsal slightly in front of the vent, equidistant from tip of caudal and
occiput, its last ray over the fourth ray of the anal. Caudal very
slightly emarginate. Distance of anal from the base of the caudal 5 in
the length. Origin of the ventrals equidistant between tip of snout and
middle of caudal; tips of the ventrals not reaching the vent.

Light brown, with purple longitudinal streaks.

Head 52 (6@ in the total); depth S3; D: 14; A. 8.

Based on a specimen in the Museum of Comparative Zodlogy.

252 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Order III, PLECTOSPONDYETL

Cope, Am. Assoc. Adv. Science, Indianapolis, 1871, 332.
As stated ante, p. 238, this order falls naturally into three groups of
families which are considered suborders and distinguished as follows :

a. Lower pharyngeals falciform, parallel with the gill arches, usually provided with teeth ; mouth
more or less protracted, usually bordered by the premaxillaries only ; jaws toothless ; brain
case produced between the orbits ; basis cranii simple ; no adipose. Eventognathi.

aa. Lower pharyngeals normal; mouth not protracted; teeth in the jaws or none; brain case
usually not produced between the orbits ; basis cranii double.

6. Not eel-shaped; vent submedian ; a dorsal fin and usually an adipose dorsal ; ventrals
abdominal ; mouth usually bordered by the premaxillary and maxillary, more rarely by
the premaxillary only. Fleterognatht.

66. Eel-shaped ; vent under the head or at throat ; dorsal obsolete ; ventrals wanting.

Gymnonott.

Of these suborders the Eventognathi belong to the northern hemisphere,
not a single species having reached South America through natural-
channels.

~The Gymnonoti are exclusively South American, reaching from Guate-
mala to La Plata. No species approach the Patagonian region.

Suborder HeEtTEROGNATHI.

The Heterognathi are composed of one, possibly two families, the Z7y-
thrinide, peculiar to South America and the Chavacide, in South
America and tropical Africa. The Avythrinide are not known to extend
south of Buenos Aires.

Family 1V.. CAARACTID 2.

Characint, Miiller, Archiv Naturgesch., 9, Jahrg., I, 323, 1843.
Characinide, Richardson, Encycl. Brit., 8th ed., XII, 245, 1856.
Characinide, Giinther, Cat. Fish, Brit. Mus. V, 278, 1864.
Charvacide, Gill, Mem. Nat. Acad. Sci. VI, 131, 1893.

The Characide flourish in the tropical fresh-waters; but one species
reaches as far north as the boundary of the United States. Of the Chava-
cide inhabiting the fresh-waters of tropical America, numbering about
100 genera and nearly 500 species, only two reach with certainty the
region under consideration. Both of these belong to the 7etragonopterine.

EIGENMANN: FRESH WATER FISHES. 253

A third one is recorded from south of Buenos Aires without a distinct
locality.'. It is without scales and forms a distinct subfamily closely allied
to the Zetragonopterine.

The relation of the Patagonian forms to those of other regions is shown
by the following keys to the subfamilies of Ciavacid@ and to the genera
of 7etragonopterine.

SUBFAMILIES OF CHARACID’

a. Teeth none ; an adipose fin.
6. Gill-rakers none; intestine very long. I. Curimatine, American.
66. Gill-rakers long, clupeiform ; fourth gill arch with raker-like organs behind.
II. Axoding, American.
aa. Teeth minute, mostly depressible ; nares close together ; gape short, mouth usually small ; an
adipose fin.
c. Last gill-arch normal.
ad. Jaws well developed, the mandibles firm ; teeth more or less spatulate, more or less
firmly joined to the jaws ; dorsal variable.
e. Scales cycloid ; teeth serrate or spoon-shaped ; gill-membranes free from isthmus

and from each other ; no predorsal spine. Ill. Hemiodontine, American.
ee. Scales ctenoid.

jf. Gill-membranes free from the isthmus. IV. Neolebiine, African.

ff. Gill-membranes joined to the isthmus. V. Distichodontine, African.

dd. Jaws very weak, mandibles transverse; teeth on the ‘margins of fleshy lips, freely
movable, spatulate; mouth subterminal, broad; dorsal short; gill-membranes
joined to isthmus; a movable, procumbent, predorsal spine.

VI. Prochilodine, American.
cc. Last gill-arch modified.

'GYMNOCHARACINUS Steindachner.
(Plate XXXIV, Figs. 3 and 3a.)

Gymnocharacinus Steindachner, SB. Akad. Wiss. Wien, CXII, 20, 1903 (type dergzz ).
This genus differs from all other Characins in being naked.

GYMNOCHARACINUS BERGII Steindachner.

Gymnocharacinus bergit Steindachner, SB. Akad. Wiss. Wien, CXII, 20, 1903. (A brook from
southern Argentina, exact locality not given.)

Head 4; depth 3%; D. 11; A. 13; V.6—7. Snout blunt, jaws equal ; maxillary not reach-
ing to middle of eye; width of head 1.5 in its length; eye 4 in head; interorbital nearly 3 ; base
of dorsal 2.4 in the head, its height 1.75 ; ventrals 2.4; base of anal less than twice in head ;
brown, a dark brown lateral band from the head to the caudal.

*The African genera are interpolated here although they very probably underwent a separate
evolution and their position in the key does not represent their relationship.

254 PATAGONIAN EXPEDITIONS: ZOOLOGY.

g. Dorsal short; lower jaw composed of three elements; teeth minute, in a single
series; gill-membranes grown to the isthmus; lower jaw with a longitudinal
extent; premaxillary rudimentary. VII. Chilodine, American.

gg. Dorsal elongate, with 17-23 rays; fourth gill arch with an accessory respiratory
organ; lower jaw composed of two elements; teeth minute, ciliform, in a single
series in each jaw ; anterior air bladder partly enclosed in a bony capsule.

VIII. Cttharinine, African.
aaa. Teeth well developed, fixed in the jaws ; an adipose fin or none.

h. Maxillary not rudimentary, not ankylosed with the premaxillary; premaxillaries not
coossified ; teeth compressed, truncate, notched, denticulate or serrate ; sometimes
canines also present. (See also Bramocharax.) Alimentary canal short.

7. Belly rounded before the ventrals ; teeth truncate, notched or serrate.

j. No scales; no adipose; teeth in premaxillary in two rows, in maxillary and
mandible in single series, narrow at base, broad at tip, 3-5 pointed; a long
fontanel in occiput. IX. Gymnocharacinine, south of Buenos Aires.

Jj. Body fully scaled.

k. Skull with a parietal, and usually in the American genera, with a frontal
fontanel ; teeth in the lower jaw not in two complete series.
7, Premaxillary and mandible with a single series of teeth.
m. Dorsal with 11 rays or fewer.

n. Large frontal and parietal fontanels. Gill-membranes united,
usually joined to the isthmus, free from it in some genera ;
intestine short; mouth small with little or no lateral ex-
tent; nares close together or remote.

X. Anostomatine, American.
nn. No frontal fontanel, parietal fontanel if present confined to
the occipital region. Gill-membranes sometimes slightly
united, usually free from each other and from the isthmus ;
nares remote ; mouth minute ; adipose present or absent ;
no teeth on maxillary. XI. Nannostomating, American.
nnn, Gill-membranes free from the isthmus and from each
other; nares close together, separated by a flap only;
fontanels large, separating the parietals and the frontals to
at least the middle of the eye; adipose always present ;
maxillary with teeth. XII. Aphiocharacine, American.
mm. Dorsal with 15 rays. XIII. Crenuchine, American.
W. Premaxillary with two or more series of teeth ; nares approximate.

o. Gill-membranes united, free from isthmus ; a single series of teeth
in the lower jaw. XIV. /euanodectine, American.

oo. Gill-membranes free from each other and from the isthmus.

p. Opercle prolonged ; premaxillary with two, mandible with
a single series of teeth ; belly not keeled ; gill-membranes
free from isthmus; rakers close together.

g. Adipose fin present; no rakers on lower limb of an-
terior gill-arch; anal long (31); dorsal in part in
frontof anal. XV. Déapomine, Rio Grande do Sul.

EIGENMANN: FRESH WATER FISHES. 255

qq. No adipose ; dorsal over anal, higher than long ; anal
long (20-31). XVI. Stevardiine, Trinidad.
pp. Opercle not prolonged; usually a single series of teeth in
the lower jaw, some genera imperfectly with two series.
XVII. Tetragonopterine, American and African.
kk. Skull without fontanels ; mandible with two series of teeth.
vy. An adipose dorsal ; walls of the air-bladder not cellular.
XVIII. Prabucinineg, American.
rr. No adipose fin; walls of the anterior portion of the posterior air
bladder cellular; premaxillary, maxillary, and outer series of the
teeth of the dentary tricuspid.
XIX. Lebiasinine, Pacific slope of Peru to Colombia.
it. Belly compressed before the ventrals ; no ventral spines; intestine short; gill-rakers
moderate ; nares close together.
s. Pectorals large, nearly or quite reaching anal, placed high or moderately so ;
gill-membranes free ; breast more or less expanded.

XX. Gasteropelicine, American.
ss. Pectoral small, placed low.

t. Body elongate ; anal basis nearly horizontal.
XXI. Agoniating, American.
tt, Body deep ; anal basis very oblique ; mouth very oblique ; pectorals on a

level with or below the level of the lower margin of the preopercle.
XXII. Stethaprionine, American.
wt. Belly trenchant and serrate, at least behind the ventrals ; intestine short ; body com-
pressed and deep.
z. All teeth in the jaws in a single series, trenchant or serrate ; gill-rakers mostly
short. XXIII. Serrasalmonine, American.
uu, Premaxillary teeth in two series, teeth tricuspid or truncate ; gill-rakers various.
XXIV. Myline, American.
hh. The minute maxillary ankylosed with the premaxillary.

v. Premaxillaries codssified; dentaries codssified ; teeth conical or compressed, uni-, bi-,
or tricuspid ; lateral line along middle of body; alimentary canal short; nares close
together; scales ctenoid.
zw. Fontanel minute or absent ; maxillary bordering the mouth.

XXV. Phagine, African.
ww. A large fontanel separating the parietals and frontals to the eyes; maxillary
not bordering the mouth; snout prolonged. XXVI. Jchthyoborine, African.

vv. Premaxillaries not codssified ; maxillary in part bordering mouth, its posterior,
toothless portion meeting the toothed part at an angle, snout not much prolonged ;
a few large triangular teeth in each jaw, nares close together ; lateral line below
the middle of body ; fontanel separating the parietals and sometimes in part the
frontals. XXVIII. Aydrocyonine, African.

hhh. Premaxillaries not codssified ; maxillary not rudimentary, not ankylosed with the pre-
maxillary ; the teeth conical or compressed (see Bramocharax) ; nares close together ;
gill-membranes free from isthmus ; alimentary canal short.

x. Fontanel present.

256 PATAGONIAN EXPEDITIONS: ZOOLOGY.

y. Belly compressed ; body long and narrow, knife-like anal long ; teeth unequal ;
a pair of large canines in the lower jaw in front, received in two grooves in
the palate; palate with minute teeth; gill-rakers reduced to slight denticles.

XXVIII. Cynodontineg, American.
yy. Belly rounded or the body not long and narrow.

z. No teeth on the palate ; snout comparatively short ; teeth all moderate in
size, sometimes strong canines ; gill-rakers strong ; nares approximate ;
gill-membranes free from isthmus. XXIX. Characine, American.
sz. Teeth on the palatines.

A, Snout very elongate ; palatine teeth minute, granular; anal short ;

intestine short. XXX. Hydrocynine, American.

AA. Snout moderate; some of the teeth canines; palatine teeth
conical, in a single series ; anal long ; intestine short.

XXXI. Acestrorhamphine, American.
x. No fontanel.

B. An adipose fin. Premaxillary with a backward projecting process between the
maxillary and the palatines, the process provided with small teeth ; no occipi-
tal crest; skull as in Erythrinine. Snout much prolonged.

XXXII. Sarcodacine, African.

BB. No adipose dorsal ; skull above more or less truncate behind, the supraoccip-
ital confined to the posterior surface and carinated by a rudimentary or obso-
lete vertical crest ; adipose fin none ; gill-opening wide, the membranes slightly
united, free from the isthmus. Dorsal in advance of the anal, usually over or
a little behind the ventrals; A. 10-13.

c. Gape very wide, premaxillary and dentaries with strong canines; lateral
line developed ; caudal rounded. XXXII. Erythrinine.
cc. Gape oblique ; maxillary with a few slender teeth, no canines ; lateral line
obsolete or developed anteriorly only ; caudal forked or emarginate.

Walls of the air bladder normal ; teeth all conical.
. XXXIV. Pyrrhulinine, American.

GENERA OF TETRAGONOPTERIN-E.

a. Jaws equal, the lower not included ; mandibulary teeth meeting the second of the premaxillary
series. Premaxillary with two series of teeth.
6. Mandibles without conical teeth in front.

c. Gill-rakers short, lanceolate ; anal long (29); lateral line complete ; maxillary with
conical teeth along half of its length; premaxillary and mandibular teeth as in
Tetragonopterus. Scissor Ginther.

cc, Gill-rakers setiform.
ad. Maxillary short, the snout and maxillary together less than half length of head.

Teeth of the lower jaw all alike in character and regularly graduated from in
front to the last tooth on the sides ; two teeth in the front row of the pre-
maxillary on each side; teeth of the second row multicuspid incisors with
contracted base, their anterior and posterior surfaces alike, convex, without
distinct ridges. Lateral line complete. Deuterodon Eigenmann,

EIGENMANN: FRESH WATER FISHES. 257

dd, Snout and maxillary forming at least half the length of the head ; lateral line
incomplete.

e. Maxillary with o-4 teeth; jaws equal ; premaxillary teeth five-pointed, in
an outer row of about 8 teeth and an inner row of about Io teeth;
maxillary usually with 2 teeth; anterior mandibular teeth stronger
than those of the premaxillary, 10-12 in number; sides of the lower
jaw with a long row of very small single-pointed teeth.

Creatochanes Gunther.
ee. Maxillary with teeth along its entire edge ; lateral line very short. Species
with dark lines along the sides.
jf. Anterior teeth of lower jaw all alike, tricuspid; anterior pair of pre-
maxillary teeth little larger than the rest. Hol/andichthys Eigenmann,
ff. Lateral teeth of lower jaw enlarged, canine-like, the fourth usually
largest ; middle teeth of premaxillary much enlarged.

Pseudochalceus Kner.
66. Mandibles with a pair of conical teeth behind the front series. Teeth otherwise all alike
imbricate, compressed, five-pointed; dorsal placed behind ventrals; scales large.
Toothed portion of maxillary-premaxillary border horizontal. Henochilus Garman.
aa. Lower jaw. included. Teeth of the front of the lower jaw strong, graduated, with ridges on
the anterior surface, those of the sides abruptly smaller (except in Hemibrycon) and incon-
spicuous ; upper jaw short, the lower jaw oblique when closed, its teeth pointing backward
and upward when the jaw is closed, frequently enclosed in the inner series of the upper jaw.

g. Premaxillary teeth in two series. (See also Sryconops.)

h. Lower jaw without a pair of conical teeth.
i. Maxillary with teeth along its entire length or nearly its entire length.
j. Lateral line interrupted. FHolopristis Eigenmann.
jj. Lateral line complete. Hemibrycon Gunther.
z. Maxillary with o-5 teeth at its anterior upper end.

&. Depth more than half the length; no predorsal spine; preventral region
flat. Tetragonopterus Cuvier.

kk. Depth less than half the length.

/, Anal and caudal naked.
m. Lateral line interrupted. Hemigrammus Gill.
mm, Lateral line complete.
z. American. Maxillary frequently with teeth. More than 3
scales between ventraland1.1. 9 Astyanax Baird & Girard.
nn. African. Maxillary without teeth. Petersiws Hilgendort.
“#, Anal or caudal scaled.
o. Anal naked; caudal scaled.
p. First and second series of premaxillary teeth in approxi-

mately parallel series. Menkhausta Kigenmann.
pp. First series of premaxillary teeth in a wavy line or imper-
fectly in two series. Bryconamericus Kigenmann.

oo. Anal and caudal scaled; anal long, convex. Postventral region
compressed, the scales not passing over it.
Markiana Eigenmann.

258 PATAGONIAN EXPEDITIONS: ZOOLOGY.

hh. Mandible with a pair of conical teeth behind the middle of the front series.
g. Both series of teeth of upper jaw simply compressed; dorsal beginning above
or behind ventrals.
rv. Lateral line complete. Micralestes Boulenger (African).
vr. Lateral line incomplete. Phenacogrammus Eigenmann (African).
gq. Inner series of teeth of the upper jaw tuberculate ; no maxillary teeth ; a series
of large, compressed multicuspid teeth in the lower jaw, similar to those in

Astyanax.

s. A parietal, no frontal fontanel. Brycinus Cuv. & Val. (African).

ss. A parietal and a frontal fontanel.
t. Origin of dorsal over or behind ventrals. Myletes Cuvier (African).
tt. Origin of dorsal in front of ventrals ; premaxillary teeth imperfectly
in three series. Bryconethiops Ginther (African).

gg. Premaxillary teeth in three or more series.

uz. Lower lip very broad, pendant. Othonophanes Eigenmann.

uu. Lower lip normal.
v. Mandible frequently with a pair of conical teeth behind the middle of the front
series except sometimes in Megalobrycon.

zw. Scales moderate, 35 in the lateral line ; anal with 20-23 rays.

Bryconethiops Ginther (African).
zz. Scales moderate or small, more than 45 in the lateral line; anal rays
more than 20.
x. Lower jaw with a second series of teeth, at least on the sides.
Brycon Miller & Troschel.
y. Anal rays more than 30. (Chalcinopsis Kner).
yy. Anal rays fewer than 30; belly not keeled.

¢. Lower jaw sometimes without conical teeth
(Megalobrycon Ginther.)
zz, Lower jaw with a pair of conical teeth. (Brycon.)
vy. Lower jaw without a pair of conical teeth in front, and without an inner series

of conical teeth on the sides.

A. Ventrals below anterior dorsal rays; anal with 32-38 rays; maxillary
without teeth ; premaxillary with two or three, mandible with a single
series of teeth. Bryconops Kner.

AA. Ventrals in front of dorsal or just under anterior rays. Anal short, of
12-22 rays; maxillary with a few blunt teeth; three series of notched
teeth in the premaxillary, a single series in the lower jaw.

Creagrutus Giinther.

7. CHEIRODON Girard.

Chetrodon Girard, Proc. Acad. Nat. Sci. Phila., VII, 1854, 199 ( fescéculus).
Type. Chetrodon pisciculus Girard.
Premaxillary, mandible and sometimes the anterior part of the maxillary
each with a single series of multicuspid incisors; dorsal short; opercle

EIGENMANN: FRESH WATER FISHES. 259

not prolonged; belly rounded; gill membranes free from each other and
from the isthmus. P

This genus is found in the Rio Magdalena, Rio San Francisco, in the
Amazon and south to the La Plata basin and from Santiago to Puerto
Montt on the Pacific slope. The relationship of Azscecudus, which enters
the Patagonian area, to the other species of the genus can be expressed
by the following key to the species.

a. Maxillary with a single tooth or none.

6. A black caudal spot ; a bright longitudinal band on the sides; maxillary without teeth in
the majority of individuals ; A. 20-27, usually 21-25 ; scales 5 or 6—30 to 36-5 or 6,
infrequently 7 scales above the lateral line; depth 2? to 24; head 33 to 4}; eye
3% to 4. interruptus Jenyns.

66. No caudal spot, or the spot very obscure; maxillary with a single tooth or none.

c. Anal 19-22; depth 2.6; head 3.6; eye 3 in head; scales 3-32-4; D. 10; an in-
distinct dark caudal spot ; a bright silvery longitudinal band. sonodon' Cope.
cc. Anal 12-15; depth 3-4; head 3-4; eye 3 in head; scales 6-33 to 36-5; D. 10-
12; a silvery lateral band, no black markings. anne McAtee.
aa. Maxillary with more than one tooth.
ad. Anal short, of but 14 rays; D. 10; teeth usually five-pointed ; a silvery lateral band
margined above with black ; depth 3-4 in the total length; head 5; eye 3 in head.
14. pisciculus Girard.
dd. Anal with 19-26 rays.
e. A black spot at base of caudal ; scales 30-32.

f. Anal 23-26; no humeral spot, an intense dark violet spot on the base of cau-
dal and continued to the tip of middle rays ; a light yellow spot just behind
the dark caudal spot on the upper and lower caudal rays; a conspicuous
violet stripe from behind the ventrals along the entire base of the anal, a branch
extending from the base of the first three or four anal rays obliquely to the
tip of the 6th and 7th rays and back along the remaining rays ; head 34-3};
depth 34-32 ; D. 9-10; lat. line 30 ; mouth very oblique, lower jaw project-
ing ; teeth numerous, very small and slender ; interorbital bones entirely cov-
ering the lower cheeks. natterert Steindachner.

ff Anal 21-24; no humeral spot, sometimes a rounded spot on each side of the
belly between the ventrals and anal; caudal spot continued on the middle

rays but not reaching the end; sometimes a black line extending forward ;
head 34-33; depth 23-3; eye nearly twice as long as the snout, 2.5-2.66
in the head; maxillary without teeth; D. 9; scales 32-34.

calliurus Boulenger.

'The dorsal profile of one of Cope’s types, kindly sent for examination by the Philadelphia
Academy, is more strongly arched than the ventral and the caudal spot appears under the lens
as a scattered series of contracted pigment cells. In specimens at hand of interruptus the ventral
profile is more strongly arched than the dorsal, and the caudal spot is large and conspicuous ;
otherwise the species are very similar.

260 PATAGONIAN EXPEDITIONS: ZOOLOGY.

iif. Anal 19; humeral spot when present not conspicuous; caudal spot not
continued to the tips of the middle rays; in some specimens the lower acces-

sory rays of the caudal spine-like; D. 11; lat. line 30-31; head 32-33;

depth 24-3 ; eye little more than 2 in the head. insignis Steindachner.

ee. No black caudal spot; scales about 35 in a longitudinal series, 9-10 in the lateral
line ; intermaxillary teeth 5, mandibular 7, maxillary 2-3 ; depth 3 in the length ;

D. 11; A. (2-20) 23; eye 3 in the head. piaba Litken.

14. CHEIRODON PISCICULUS Girard.

Chetrodon pisciculus Girard, Proc. Acad. Nat. Sci. Phila. VII, 1854, 199;
Girard, U. S. Naval and Astronomical Expedition, 249, pl. xxxiv, figs.
4—7, 1855 (Lagoons near Santiago, Chili); Eigenmann & Eigen-
mann, Proc. U. S. Nat. Mus., 1891.

Chivodon pisciculus Giinther, Cat. Fish. Brit. Mus. V, 332, 1864 (copied) ;
Steindachner, Zoo]. Jahrb. Suppl. IV, 328, 1898 (source of “Rio
Maulin, Lake Llanguihue at Puerto Montt).

This species is known only from the types and from a specimen from
Puerto Montt.

: 8. ASTYANAX Baird & Girard.

Astyanax Baird & Girard, Proc. Phila. Acad. Sci. VII, 1854; Girard, U.
S. Mex. Bound. Survey, 74, 1859 (avgentatius).

Pecilurichthys Gill, Ann. N. Y. Lyc. Nat. Hist., 54, 1858 (dvevoortiz).

Type. Astyanax argentatus Baird & Girard.

Premaxillaries with two series of teeth, the first series with several teeth
on each side; mandible with strong teeth in front and very small ones on
the sides, without conical teeth in front; teeth of premaxillary equal or
graduated, their crowns ridged and denticulate. Gill-rakers setiform.
No predorsal spine. Maxillary with o-10 teeth. Lateral line complete.
Form slender ; depth mostly more than 2 in the length.

This is one of the largest genera of Characins. Its representatives are
found from the borders. of the United States to the Rio Negro on the
east and Peru on the west, and from Para to the Cordilleras. Some of
its species have a restricted range, while others (4. vuti/us) have a range
nearly coextensive with that of the genus. Astyanax (forms like ratlus,
abramis and bimaculatus) is probably one of the oldest inhabitants of
South America. Its allomorphism and wide distribution would indicate its
long establishment. Very near relatives of this genus are found in Africa.

The only species known to enter the Patagonian region is Astyanax
vutilus (Jenyns).

a

EIGENMANN: FRESH WATER FISHES. 261

Giinther recorded this species under the name of 7efvagonofterus peten-
ensts from the Rio Negro. The only recorded locality of Aefenenszs is Lake
Peten in Yucatan. Mr. C. Tate Regan has reéxamined the Rio Negro
specimens and writes me: ‘‘I have no hesitation in referring the Rio
Negro specimens to 7etragonopterus rutilus.”’

Rutilus is found from Panama (possibly Mexico) to Patagonia.

The relation of this species to the other members of the genus is shown
by the following key :

a. Anal rays 30 to 49.
6. Lateral line 55. Anal 45; head 3.6; depth 2.63; lat. line 55. “Eye greater than snout
by one-third of its diameter, greater than interorbital by j—} of its diameter.
1. erythropterus (Holmberg).
66. Lateral line with 50 scales or fewer.
c. Anal 40-48.
d. Scales 42-50.

e. Anal 47-48.

jf. A few rudimentary teeth on the maxillary ; a silvery lateral band ; ven-
tral profile much arched ; snout pointed; A. 48. Lat. line 47.

2. spilurus Cuv. & Val.

ff. No teeth on the maxillary ; a very distinct silvery lateral band ; an

indistinct caudal and humeral spot ; maxillary reaching anterior mar-

gin of orbit; depth 2.3; eye 3 in the head; scales 9 or 10-47-10

oP ities Je\s ZU7fe 3. hauxwellianus (Cope).

ee. Anal 41-45.

g. Scales 10-45 to 50-8 or 9; asmall silvery lateral band ; a dark spot on
shoulder and another! on base of caudal; depth 24-22; A.41-45;
maxillary with a single, rather large caducous tooth; eye 2.4 in the

head, about equal to interorbital. Head 4.1-4.5.
4. pelegrint Eigenmann,
gg. Scales 8-42—7 or 8; A. 42; eye 3.7 in the head, 1.7 in the inter-

orbital ; head 3.7 in the length ; a caudal spot.
5. correntinus (Holmberg).

dd. Scales 37 or 38.

h. Several small teeth on the maxillary; head 33; depth 3 or somewhat
more than 3; snout 4 in the head; a silvery lateral band bordered above
by blue-green ; a round humeral spot ; caudal spot, when present, extends
to the end of some of the caudal rays. A. 43; scales 6-37 or 38-4.

6. bairdi (Steindachner).

hh, Maxillary with a single, small tooth. Scales in lateral line 37 or 38 ;
both a caudal and a humeral spot very distinct, round ; middle of caudal

fin thickly dotted with black ; maxillary not reaching orbit ; head nearly

4; depth about 2? ; A. 40-42. 7. tabatinge (Steindachner).

‘These spots are very obscure in alcoholic specimens, while in specimens first preserved in for-
malin they are conspicuous and a black band replaces the silvery band of the alcoholic specimens.

262 PATAGONIAN EXPEDITIONS: ZOOLOGY.

cc. Anal 30-41.
z. Lateral line 40-50.

j. A silvery lateral band ending in a black caudal spot, sometimes two lateral
spots in front; maxillary very short, without teeth, 2 in the eye; depth
24-3 ; head 4-44; eye 24-23 in head; D. 11, behind ventrals; A. 35-40.
Lat. line 44-47. 8. feste (Boulenger).

yj. A black band extends from base of caudal fin along its middle rays ; max-
illary extending to eye, its length at least 3 the length of the eye, with-
out teeth ; eye 22-3 in head ; depth 22-23 ; head 4; A. 34-40; scales 8
or 9-40 to 50-8 or g; last dorsal over first anal ray.

9. brevirostris (Ginther).

Jj. Maxillary with two small teeth, equal to eye in length; eye 2.6-2.8 in

head ; depth 2.7—3 ; head 3.5-3.6; scales 8-39 to 45-6 or 7; A. 28 or
29; a dark caudal spot, not extending to end of middle rays.
38. emperador, which see.

Jj. Caudal and humeral spots generally absent ; origin of dorsal over ventral ;

interorbital wider than eye; depth 3; head 4; scales 9-40 to 46-10;
A. 26-31. 37. cordove, which see.
zz. Lateral line less than 40, except sometimes in abramzs and bimaculatus.

k. Anal rays 36 or fewer, rarely as many as 39 in d¢maculatus, maxillary with
a few teeth or none.

/, Depth more than three in the length, the form long and slender ; max-
illary without teeth.

m. Humeral and caudal spots present, anal and ventral with broad
red margins ; eye 3% in the head, much smaller than the con-
vex interorbital. Maxillary extending to anterior margin of
eye. A. 27-31; scales 7-36-7. 27. humilis, which see.

mm. No caudal or humeral spots ; a silvery lateral band, most dis-

tinct posteriorly; head 34; eye 2} in the head. A. 30;

scales 5-35—3. Maxillary toothless. 10. astictus (Ulrey).

7. Depth 3 in the length. Middle rays of caudal covered by a broad
band which extends to the margin of the fin and bordered above and

below by yellow ; a silvery lateral band ; a humeral spot ; head not 4

in the length. A. 27-35. Scales 7-38-7. Maxillary witha single

tooth. 23. teniurus, which see.
il. Depth less than 3 in the length; a distinct caudal spot.

n. No humeral spot ; depth 2}~—23 ; eye 32-4 in the head ; maxillary
with 0-3 teeth, extending considerably beyond anterior margin
of eye ; a conspicuous black band on the caudal peduncle, be-
coming wedge-shaped on the caudal. A. 29-31. Scales 7 or
8-37 or 38-6. Il. maximus (Steindachner).

un, A distinct humeral spot.

o. Maxillary long; head 3.75; depth 2.75; eye 3 in the head,
1% in snout, equal to interorbital ; maxillary without teeth ;
a black humeral spot, a silvery lateral band, becoming
black on the tail and extending on the caudal; origin of

EIGENMANN: FRESH WATER FISHES. 263

dorsal over ventrals. A. 30; scales 8-37 or 38-7.
12. moort (Boulenger),
oo. Maxillary extending little if any beyond origin of eye.
p. Dorsal plain.
g. Humeral spot horizontally elongate.
ry. Scales in 19 or 20 rows, 10-43 to 47-8 or 9;
anal 28-32. Caudal plain or with an indis-
tinct spot ; humeral spot indistinct or wanting.
Dorsal distinctly behind the ventral; the
pectoral reaches the ventral. Depth 23;
head less than 4.
13. abramis (Jenyns),
rr. Scales in fewer than 17 rows.
s. Moderately compressed.

t. Teeth of the inner series of the pre-
maxillary with their posterior
surface convex, the denticles cor-
responding to the convexity
arranged in a U-shaped line.
Scales 6 to 8-30 to 40-5 to 8;
maxillary with o-4 teeth, extend-
ing somewhat beyond the front
margin of theeye. Depth 2-2.1 ;
head 4-4.2; eye 3 in the head.
A. 27-39.

14. b¢maculatus* (Linnzus).
tt. Teeth of the inner series of the pre-
maxillary alike in front and be-
hind, the denticles arranged in a
nearly straight line; scales 6—
40-6; maxillary with a single
tooth, extending to below origin
of pupil ; depth 2.4 ; head 4; eye
a little more than 3 in head.

Jals Gheh,
15. orthodus (Eigenmann).
ss. Greatly compressed. Scales 7 or 8-36
to 37-6 or 7; a humeral and caudal
spot; the pectorals extend beyond

‘The statement of Steindachner that the maxillary has minute teeth along its entire edge is
evidently due to a slip of the pen; A. drevoortit from Trinidad has the following characters :

Scales 7-37 to 40-7; very distinct humeral and caudal spots, the latter often extending to
the ends of the caudal rays; an obscure silvery band. Eye 3}; maxillary sometimes with a
single tooth. A. 29-34; var. “meatus from the La Plata system has more or less distinct series
of spots or lines following the rows,

264 PATAGONIAN EXPEDITIONS: ZOOLOGY.

origin of ventrals; maxillary toothless,
extending somewhat beyond front
margin of orbit; head 3}-3.5; depth
24-24; A. 38-41. Very greatly
compressed.
16. caucanus (Steindachner).
gq. Humeral spot circular or vertically elongate.
wz. Depth 2-2.2 ; scales 8 or 9-36
to 40-8 or 9; a humeral and
a caudal spot; maxillary
with one tooth; head 3.66;
depth 2-2.2; A. 38-39;
dorsal and ventral outlines
equally curved.
17. atratoensis (Eigenmann).
uu. Body more elongate, humeral
spot very conspicuous.
Anal rays 38-41. A silvery
lateral band, a humeral spot.
Dorsal a little behind the ori-
gin of the ventrals ; the an-
terior anal rays elongate.
Maxillary extending to near
the anterior margin of the
eye. Depth 2.4-2.8; head
3.5; scales 8-37 to 40-I0.
18. sti/be (Cope).
pp. Dorsal with a broad, oblique, dark band across the
middle, or plain. Scales 8-39-7; maxillary extends
a little beyond the origin of the eye; D. 11; A. 31;
depth 22; head 33. 19. dartlettii (Ginther).
aa. Anal rays 26-29, rarely 31.

v. A series of seven deep brown longitudinal bands. Maxillary extending little, if any,
beyond anterior margin of eye, with one tooth. Head 3.5 ; depth a little more than 2 in
the length. <A. 27; scales 5-31-4. (See also under dzmaculatus.)

20. steindachneri (Eigenmann).

vv. A single lateral band or none.

ze. Caudal without vertical band.

x. Scales 28-35.
y. A caudal spot.
z. Shoulder spot horizontally oval, well defined.

A. Scales in 18 or 19 rows; A. 28-32.1 13. abramis, which see.

AA. Scales in 11-16 rows. 14. dimaculatus, which see.

AAA, Scales 7-34-6. Lateral band black. Scales in 14 rows.

Dorsal fin behind the base of the ventrals. Maxillary with

EIGENMANN: FRESH WATER FISHES. 265

I tooth, extending a little beyond the anterior border of the
eye to end of first suborbital; interorbital space much
greater than the eye, very convex; eye 4 in head. Head
3%; depth 3; A.28. A humeral spot.
21. wappi (Cuv. & Val.).
AAAA. Scales 7~36-7 ; maxillary extending to eye; eye 3.66 in
head, much smaller than the convex interorbital; A. 2 7-31;
anal and ventral with broad red margins.
30. humilis, which see.
#z. Shoulder spot, if present, vertically elongate, sometimes faint.
&. Caudal spot band-like, continued on the middle caudal rays ;
scales in fewer than 15 rows.
C. A humeral spot.
D. Dorsal behind the ventrals.
£. Maxillary with 2-7 teeth; anal 27-32; depth
2}—3'; head about 4; maxillary equal to eye;
eye 3 in head, equal to intraorbital, Scales
737-9
22. rutilus nicaraguensis Eigenmann.
EE. Maxillary with o-2 teeth.
#. Maxillary extending distinctly beyond the
anterior margin of the eye. Head 4-43;
depth 2.25-3; A. 24~32; scales 6 or 7-30
to 39-44 to 6. 23. rutilus* (Jenyns).
/F. Caudal band bordered by yellow above
and below ; head not 4 in the length ;
maxillary with a single tooth; A. 27-
35; scales 7-38-7.
24. teniurus (Gill).
FFF, A, 24-32; origin of the dorsal behind the
base of the ventrals; maxillary with
from O-2 teeth, reaching anterior
border of the orbit; head 4; depth
2.25-2.75 ; scales 7 to 9-35 to 39-6
or 7. 25. rutilus eneus (Ginther).
FFFF, Depth 3.33; head 3.66; A. 29; scales
7-37-5; maxillary long, equal to
eye; eye 2.5 in the head; interor-
bitals 3.66. 26. cuvieri (Litken).
DD. Dorsal over ventrals; snout less than 4 in head ;
‘interorbital flattish ; A. 27, beginning behind the
dorsal ; depth 3 ; head 4; lat. 1. 37.
27. petenensis (Giinther).

* Rutilus jequitinhonhe has depth 3.

266 PATAGONIAN EXPEDITIONS: ZOGOLOGY.

CC. No humeral spot (Western Peru and Ecuador).
G. Dorsal behind ventrals; snout about 4 in the head ;
interorbital convex. A. 26-30, beginning under last
ray of dorsal; depth 3-33; head 33-4. Scales 6
or 7-36 to 40-7 ; 2 or 3 maxillary teeth. A silvery
lateral band which may become black on the caudal.
28. simus (Boulenger).
GG. Head 34-44; depth 23-3; interorbital much
more than diameter of the eye. Maxillary termin-
ates below the anterior margin of the eye, with 2
teeth ; origin of dorsal behind the ventrals. A.
26-30 ; scales 6-35 to 37-6 to 8.
29. peruanus (Miller & Troschel).
BB. Caudal spot not continued on the middle rays."

H. Scales in 15 rows, 7-36-7; A. 27-31; depth more than
3 in the length; eye 3.66 in the head, much smaller
than the convex interorbital ; a humeral spot and a cau-
dal spot. 30. humilis (Ginther).

HH. Scales in 10-14 rows ; lateral line 34-38.

I. Head 4 or less than 4 in the length.

J. A humeral spot; maxillary reaching anterior
border of eye, not to end of first suborbital ;
head 3.75-4; depth 2.4-2.7; A. 23-28;
scales 64 or 74-34 to 37-54 to 63. Twoor

three maxillary teeth.
31. fischeri (Steindachner).
J]. Maxillary ceases in front of the vertical from
the pupil, and end of first suborbital. Dorsal
considerably behind origin of ventrals ; pec-
torals reaching beyond ventrals. Eye about
equal to the slightly convex interorbital, 3 in
the length of the head. Depth 2}; head
3%; A. 26; scales 63-37 or 38-5; a
humeral spot. 32. caroline (Gill).
/I, Head 4 or less in the length. Scales 54-35-4.
Anal, dorsal and caudal fins with red markings.
Caudal and humeral spots indistinct. Head 4.2;
depth 3.75; A. 25-29. 33. phanicopterus (Cope).

yy. No caudal spot ;* a humeral spot.

K. Maxillary with three small teeth; eye 2.3 in head, snout 4.5, the
flattish interorbital 3 ; depth 2.66; head 3.6; A. 28; scales 5-
35-4 ; a silvery lateral band, a vertical humeral spot.

34. megalops (Eigenmann & Ogle).

1 Not examined in caroline and humilis.
2 Not stated for dahtensis.

EIGENMANN: FRESH WATER FISHES. 267

KK. Maxillary with numerous minute teeth, extending beyond the front
of the orbit. Snout shorter than eye. Eye 2.8-3. A. 29-30;
scales 5 or 6-32 or 33-4 to 5. Head 3-6; depth 2}~-2};
humeral spot elongate horizontally.

35. bahiensis (Steindachner).

KKK. Maxillary without teeth. Silvery band not edged with green

above ; origin of dorsal between ventral and anal ; pectorals

to origin of ventrals, ventrals nearly to anal. Head 43;

depth 33; eye 2,8 in head; D. 10; A. 27; scales 5-37-3.

36. alburnus (Hensel).

ax, Scales 9-40 to 46-6-8 ; longitudinal series of scales 14-17. Caudal and
humeral spots generally absent ; origin of dorsal over root of ventrals. Width

of interorbital greater than the diameter of the eye. Head 4; A. 26-30.

Interorbital very convex, 22 in head; eye equals snout, 32 in head; depth

of caudal peduncle about } of the depth. 37. cordove (Gunther)

+xx. Scales in 15 or 16 rows, 8-39 to 45-7; A. 28 or 29 ; head 3.6 or 3.5 ; depth

2.7-3 ; eye large, 2.6-2.8 ; interorbital 3.25 in head; maxillary as long as

eye, with two narrow teeth ; dorsal behind origin of ventrals ; humeral spot

faint ; caudal spot distinct, not continued on middle rays.
38. emperador (Eigenmann & Ogle).
aaa. Anal rays 16-25, rarely 26 and 27 in Jasciatus and eneus ; 3 to 5 series of scales below the
lateral line in all but enews (6 or 7) and fasciatus (44-7). Caudal without black cross
band.
ZL. Maxillary without teeth; two or three scales below the lateral line.
M. No caudal spot.

N. Depth 4.7 in the length; longitudinal series of scales 12. A broad
silvery lateral stripe ; no caudal spot. Maxillary toothless, rather wide,
extends little beyond anterior border of the orbit. Head 4.2; depth
Mey S JN Ala 39. /ongior (Cope).

NN. Depth 2.5-3.5 in the length.

QO. A silvery lateral band sharply edged above with a dark band. Dorsal
fin a little behind the ventrals. The pectorals not entirely reaching
the ventrals, the ventrals reaching the anal. Head 3.8; depth
3-3#; scales 5-32-3. <A. 21-22. 40. copet (Steindachner).

OO. The broad silvery lateral band not edged above with dark. Maxil-

lary extending nearly to the front margin of the eye; anterior
dorsal and anal rays elongate. Head 4.4; depth 3+; scales 4~35-
3.8 Ye\s UCL 41. diaphanus (Cope).

OOO. The narrow silvery band edged with greenish above. Origin
of the dorsal just behind the ventral ; the pectoral reaches to the
middle of the base of the ventral. Head 3-4-3.75 ; depth 2.8-2.86;
A. 24-25 ; scales 5-32 or 33-3. 42. collettit (Steindachner).

LL. Maxillary with teeth.
P. Middle caudal rays black.
Q. One to three teeth in the maxillary.

268 . PATAGONIAN EXPEDITIONS: ZOOLOGY.

R. A. 18-25. A silvery-gray lateral band, a dark caudal spot fading
out forward. Usually a humeral spot. Maxillary with two small
teeth. Head 4; depth 23-3; scales 7 or 8-37 to 40-3; eye 3 in
head. 43. meaicanus (Filippi).

RR. A. 24-27. Origin of the dorsal behind the base of the ventrals ;
-maxillary toothless, or with one or two teeth, reaching anterior

border of the orbit. Head 4; depth 24-2?; scales 7 to 9-35 to

39-6 or 7. 25. eneus (Gunther).
RRR. A. 18-25. <A dark caudal spot extending to the end of some of
the rays and fading out anteriorly; a silvery lateral band; an
indistinct humeral spot ; ventral and pectoral fins with red. Head

4, longer than its depth at the occiput; depth 23-3; scales 5

or 6—32 to 40-43 to 6 rarely 7. 44. fasciatus (Cuvier).

RRRR. A. 25-30; depth 2.6-2.66. 23. rutilus.

RRRRR. A. 3.16-19. Three or four scales below the lateral line.

Head 4-44; depth 24-3; scales 5-35 to 37-3 or 4. A.

3, 16-19. 45. fasciatus theringit (Boulenger).

QQ. Numerous minute teeth on the maxillary which extends either a little

or distinctly behind the front margin of the orbit. Eye small 3.5-4

in the head. A vertically elongate humeral spot and a longitudinal

caudal stripe extending to the end of the rays and fading out anteriorly.

i Head 34-37; depth 3-34; scales 5-33-4. A. 16-18.

46. jenynsii (Steindachner).
PP. Middle caudal rays not black.
S. Scales in lateral line 34-38.

T. A. 23-28; depth 2.4-2.7 ; scales 6} or 74-—34-37-54-6.

31. fischeri, which see.

TT. A. 19 or 20.'| Almost colorless; caudal rays sometimes dusky, a

grayish lateral band. Pectorals reaching % to ventrals, ventrals 3
to anal. Maxillary, with two teeth, reaching to eye. Head 4;
depth 3.2-3.75 ; scales 4 or 5-35 to 38-4.

47. mankhausi (Eigenm. & Kennedy).

TTT. Anal 17-18. Yellowish above, white on the sides; lateral band

plumbeous above, silvery below, a plumbeous humeral spot.
Head 4} to 44; eye 2} in head; depth 3.2-1.6. Lateral line
34-36. Dorsal band, large part of anal, the body near it and
the median spot of the caudal more or less red.

48. rubropictus (Berg).

TTTT. A. 18; scales 37 or 38; depth 35; head 43; plumbeous,

fins dusky ; pectorals reaching ventrals ; ventrals nearly to anal ;
caudal lobes rounded ; snout blunt, lower jaw distinctly shorter
than upper ; a faint humeral spot. Mouth very small, maxil-
lary not reaching eye; eye 34 in head, interorbital very con-
-vex, less than 3 in head; depth of caudal peduncle little less
than half the greatest depth.

49. eigenmanni Evermann & Kendall.
‘26 in one specimen.

EIGENMANN: FRESH WATER FISHES. 269

SS. Scales 5-31-3. Silvery lateral band; a diffuse caudal spot ; no humeral
spot. Head 33; depth 23. A. Io. 50. paucidens ( Ulrey).

15. ASTYANAX RUTILUS (Jenyns).
(Plate XXXIV, Fig. 4.)

Tetragonopterus rutilus Jenyns, Zool. Beagle, Fishes, 125, pl. 23, fig. 2, 1842 ;
Steind., Ichthyol. Notizen, IX, 10, pl. 1i, figs. 2 and 3, 1869 (Mon-
tevideo) ; Hensel, Wiegm. Archiv, 1870, 80; Steind. Siisswf. siiddstl.
Bras. III, 575, pl. ii, figs.“ 1 and 2, 1876 (Rio Parahyba; Rio Doce;
Montevideo ; Rio de Janeiro; Rio Jequitinhonha; Xamapa, Mexico);
Steind. Fischf. Cauca und Fliisse bei Guayaquil, 22, 1880 (Cauca) ;
Boulenger, Proc. Zool. Soc. Lond., 1887, 281 (Canelos); Boul., Ann.
& Mag. Nat. Hist. XIV, 173, 1887; E. and E., Proc. U.S. Nat. Mus.
XIV, 1891, 52 (Rio Grande do Sul); Cope, Proc. Am. Philos. Soc.
1894, 87 (Rio Grande do Sul); Ulrey, Ann. N. Y. Acad. Sci. VIII,
280, 1895; Perugia, Ann. Mus. Civ. Storia Nat. Genova, Ser. 2a,
X, 44, 1891; (Resistencia & Laguna Ibera; Candelaria; Buenos
Aires); Eigenm., Ann. N. Y- Acad. Sci. VII, 633, Feb. 1894 (Rio
Grande do Sul); Lahille, Revista Museo de la Plata, VI, 7, 1895
(Puerto Viejo; Arroyo de Gato; Dofia Flora; Dock Central; Isla
Santiago; Punta Lara); Boulenger, Boll. Mus. Torino XII, 1897
(Caiza; Mission de San Francisco; San Lorenzo) and XIII, 1898, 2
(Rio Peripa; Rio Zamora; Rio Santiago); Eigenmann & Norris,
Revista Museo Paulista, 357, 1900 (Taubaté ; Rio Tieté).

Tetragonopterus teniatus Jenyns, Zool. Beagle, Fishes, 126, 1842; C. &
V., XXII, 145, 1848; Mill. & Trosch., Fische Brit. Guiana, 635,
1848 (Ditches and swamps near coast) ; Giinther, Cat. Fish. Brit. Mus.
V, 329, 1864.

Tetragonopterus fasciatus C. & V., XXII, 149, 1848 (not of Cuvier) ;
Giinther, Cat. Fish. Brit. Mus. V, 22, 1864 (not synonymy) (Brazil ;
West Ecuador ; Huamachal; Rio Guacalate ; Rio Chisoy ; Vera Paz,
Mexico; Cordova; Central America; Guatemala) ; Giinther, Ann.
é@ Mac. Nat: Hist, July, 1880 (Lai Plata); E. & E: Proc. US. Nat.
Mus., XVI, 1893, 55; Vaillant, Bull. Mus. d’Hist. Nat. 1897, 221
(Chagres) ; id. l.c. 1899, 155 (Carnot).

270 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Tetragonopterus scabripinnis Kner (not Jenyns), Characinen, 39, 1859 (Xam-
apa, Mexico ; Irisanga) ; Giinther, Cat. Fish. Brit. Mus. V, 325, 1864
(in part).

Tetragonopterus microstoma Giinther, Cat. Fish. Brit. Mus. V, 323, 1864 ;
Giinther, Ann. & Mag. Nat. Hist. July, 1880 (Rio Plata).

Tetragonopterus eneus, Hensel (not Giinther); Wiegm. Archiv. 1870, 87
(Southern Brazil).

Tetragonopterus jequitinhonhe Steind., Siisswf. Siidéstl. Bras., III, 27,
pl. ii, fig. 3, 1876 (Jequitinhonha); E. & E., Proc. U. S. Nat. Mus.,
XIV; 191, 527 Ulrey, Ann NY. Acad. Sa. VITEs2380;40s05:
Eigenmann & Norris, Revista Museo Paulista, 357, 1900 (Piracicaba).

Habitat: Rio Negro, Patagonia to Mexico, in all streams of the east-
ern slope of South America; Western Ecuador.

Of this species Steindachner says: ‘Almost every river system pos-
sesses a peculiar variety of this species; according to age, sex, season ;
according to abundance or scarcity of food; according to the habitat in
cool or clear mountain brooks or deeper stagnant waters, the outlines of
the body vary, and in part also, the number of horizontal rows of scales
and of the anal rays.”

A single specimen has been recorded from the Rio Negro. It is pre-
served in the British Museum and is reproduced in fig. 4, plate XXXIV.

Order TV. HAPLOWEL

Cope, Proc. A. A. A. Sci., Indianapolis, 1872, 328 and 333; Gill, East
Coast Fishes, 1872, 14.

‘Air-bladder, if present, communicating with the digestive tract by a
duct. Opercle well developed. Pectoral arch suspended from the skull;
no mesocoracoid arch. Fins usually without, rarely with a few spines ;
ventrals abdominal, if present. Anterior vertebrz distinct, without
Weberian ossicles.”” Boulenger.

The boundaries of this order are differently conceived by different
authors and most of the constituent families are extralimital. It includes
three families with representatives in South America. Of these, one, the
Poeciliidze, which is dominant in Central America and whose species
straggle north through the temperate zone and south to Buenos Aires, is

EIGENMANN: FRESH WATER FISHES. 271

not known to have representatives in Patagonia. The three families may
be distinguished as follows :
a. Parietals separating the frontal from the supraoccipital ; post-temporal simple; precaudal

vertebra with autogenous parapophyses.
6. Margin of the upper jaw formed by the premaxillaries and the maxillaries ; basis cranii

simple ; no adipose dorsal fin. Galaxude.
66. Margin of the upper jaw formed by the premaxillaries only; basis cranii double; adipose
dorsal fin present. Aplochitonide.

aa. Frontals in contact with the supraoccipital ; precaudal vertebrae without well developed para-
pophyses ; border of mouth formed by premaxillaries only ; dorsal and anal without spines ;
scales cycloid, or with erect spines; no adipose fin; mouth protractile; ventral fins if present
with 5-7 rays. Pecilude.

Family V. G4LAXITIDA:.

Galaxie Miiller, Abhandl. Akad. Wiss. Berlin, 1844, 187.
Galaxide Giinther, Cat. Fish. Brit. Mus. VI, 208, 1866.

This family consists of two genera,’ Veochanna with a single species
inhabiting New Zealand, where it sometimes burrows in damp clay some
distance from water, Ga/axtas with 26 recognized species, of which six
are found in southern South America or adjacent islands, two at the Cape
of Good Hope, and the remainder in New Zealand, Tasmania and
Australia. One of them, affenuatus, is probably common to American,
Australian and New Zealand waters. The species range from the ocean
to an elevation of 6,000 feet. Some are evidently land-locked, while
others descend to the sea to spawn.

The Galaxias of the Cape of Good Hope is the only type of fishes
south of the tropic of Capricorn in Africa that is not found north of it.

Neochanna is distinguished from Galaxtas by the absence of ventrals.

g. GALAXIAS Cuvier.

Galaxias Cuvier, Régne Animal, II, 183, 1817 (¢vadt¢aceus).
Mesttes Jenyns, Voy. Beagle, Fishes, 118, 1842, sp.
Austrocobitis Ogilvie, Proc. Linn. Soc., New South Wales, XXIV, 1899,
158, (fide Regan).
Boulenger (Nature, Nov. 27, 1902) says concerning Ga/axias ; ‘‘Most
text-books and papers discussing geographical distribution have made
much of the range of a genus of small fishes, somewhat resembling trout,

1 Cromeria from the White Nile has been shown to be not related to the Ga/aviude.

272 PATAGONIAN EXPEDITIONS: ZOOLOGY.

the Galaxias, commonly described as true fresh-water forms, which have
long been known from the extreme south of South America, New Zea-
land, Tasmania and southern Australia. The discovery, within the last
few years, of a species of the same genus in fresh water near Cape Town,
- whence it had previously been described as a loach by F. de Castelnau,
has added to the interest, and has been adduced as a further argument in
support of the former existence of an Antarctic continent. In alluding to
this discovery, when discussing the distribution of African fresh-water
fishes in the introduction to my work ‘Les Poissons du Bassin du Congo,’
in 1901, I observed that, contrary to the prevailing notion, all species of
Galaxias are not confined to fresh-water, and that the fact of some living
both in the sea and in rivers suffices to explain the curious distribution
of the genus; pointing out that in all probability these fishes were formerly
more widely distributed in the seas south of the tropic of Capricorn, and
that certain species, adapting themselves entirely to fresh-water life, have
become localized at the distant points where they are now known to exist.
Although as recently as October last the distinguished American ichthy-
ologist D. S. Jordan, wrote (Science, XIV, p. 20): ‘We know nothing of
the power of Ga/axzas to survive submergence in salt water, if carried in
a marine current,’ it is an established fact, ascertained some years ago by
F. E. Clarke in New Zealand and by R. Vallentin in the Falkland Islands,
that Galaxias attenuatus lives also in the sea. In New Zealand it
periodically descends to the sea, where it spawns, from January to March
and returns from March to May. In accordance with these marine habits,
this species has a much wider range than any of the others, being known
from Chile, Patagonia, Tierra del Fuego, the Falkland Islands, New
Zealand, Tasmania, and southern Australia.

“T now wish to draw attention to a communication made by Captain
F. W. Hutton in the last number of the Transactions of the New Zealand
Institute (XXXIV, p. 198), ‘On a Marine Galaxias from the Auckland
Islands.’ This fish named Galaxtas bollansi, was taken out of the mouth
of a specimen of Merganser australis during the collection excursion to ~
the southern islands of New Zealand made in January, 1901, by His
Excellency the Earl of Ranfurly.

“It is hoped that by giving greater publicity to these discoveries the
family Galaxide will no longer be included among those strictly confined
to fresh waters, and that students of the geographical distribution of animals

EIGENMANN: FRESH WATER FISHES. 273

will be furnished with a clue to a problem that has so often been dis-
cussed on insufficient data.”

The species of this genus recorded from Patagonia have received elab-
orate consideration from Smitt. He found that the young differs greatly
from the adult and that three stages may be distinguished, ‘‘alevins,” less
than 45 mm., “fretins,” about 50 mm., and adult, over 60 mm. Species
based on these larval stages are scarcely recognizable. The difference
between the alevins and the fretins of the same species is greater than the
differences between two adult species. The only character distinguishing
the species in all ages is the length of the lower jaw in relation to the
distance of the dorsal from the tip of the snout.

He distinguishes the two species recognized by him as follows:

a. Head 18.8-20.1 per cent. of the length; distance of origin of dorsal from tip of snout 72-74.8
per cent. of the length; distance of origin of ventrals from tip of snout 48.5—52.1 per cent.
of the length; pectoral 11; A. 13-15; lower jaw 9.3-10.2 per cent. of the length from tip
of snout to dorsal. maculatus.

aa. Head 22.4-26.6 per cent. of the length; distance of origin of dorsal from tip of snout 53.8—
56.6 per cent. of the length; P. 13-15; A. 9-12; lower jaw 12.9-16.3 per cent. of the
distance from tip of snout to dorsal. alpinus.

The color of the alevins of both species is uniform yellowish or gray.
In the alevins of maculatus a series of pigment cells follows the dorsal
border posteriorly; another follows the ventral border and alimentary
canal; another series along the lateral line, another along the upper mar-
gin of the notochord; a black bar at base of pectoral; a black spot at
upper angle of preopercle and others on snout and top of head.

During the fretin stage the color of the two species is alike and
remains so till the species has reached a length of about 8.5 cm.

More recently Regan, Proc. Zool. Soc. London, 1905, II (April 1906),
considered the fretins of Smitt asa distinct species, gvace//imus of Canes-
trini. Smitt’s a/éznus he considers in part atfenuatus, in part maculatus
and in part platez.

Regan defines six American species of which a/fzvus is confined to the
alpine lakes of Hardy Peninsula, Tierra del Fuego, and smi to the
Falkland Islands. The remaining four are distinguished as follows :

a. Six or seven branchiostegals ; caudal emarginate ; origin of anal opposite or slightly posterior
to that of the dorsal.
6. Origin of ventral equidistant from tip of snout and base of caudal or nearer the former.

274 PATAGONIAN EXPEDITIONS: ZOOLOGY.

c. Length of head 5 (young) to 6} (adult) in the length; depth 53-10; snout a
little longer than eye (in the adult), smaller than interorbital ; origin of ventral
equidistant from tip of snout and base of caudal ; depth of caudal peduncle } to 2
in its length ; upper parts of head and body finely punctate with blackish and spotted
or marbled with dark purplish. Recorded specimens 55-170 mm. 16. attenuatus.

cc. Length of head 7-7} (young) in the length; depth 10-12; snout a little shorter
than eye, less than interorbital ; origin of ventrals nearer tip of snout than base of

caudal ; caudal peduncle 2} times as long as deep; some small blackish spots on
the head and on the upper part of the body; a line of black dots along the middle
of the side and one at the base of each of the unpaired fins. Recorded specimens
reaching a length of 55 mm. gracillimus.
66. Origin of ventral nearer to base of caudal than to tip of snout.

d. Maxillary extending to below anterior margin of eye or slightly beyond; depth
6-8; head 44-5?; snout nearly as long as eye which is 34-44 in the head;
interorbital 23-3; jaws equal; origin of ventral equidistant from middle of pec-
toral and origin of anal; caudal peduncle 13-2 times as long as deep; numerous
irregular blackish spots. Recorded specimens 73-120 mm. 17. maculatus.

aa. Eight or nine branchiostegals ; caudal truncate or emarginate ; origin of anal behind that of
dorsal ; origin of ventral considerably nearer to base of caudal than to tip of snout ; depth
5-6; head 44-4?; snout longer than eye, which is 5-73 in head; interorbital 23-22 ;
jaws equal; maxillary extending beyond origin — frequently to middle of eye; caudal
peduncle as long as deep or a little longer; head, body and fins covered with numerous
irregular dark spots. Recorded specimens 49-300 mm. 18. plate.

16. GALAXIAS ATTENUATUS (Jenyns).

Only the Patagonian synonymy and bibliography are given. The spe-
cies is also found in Tasmania, Australia, New Zealand.

Mesites attenuatus Jenyns, Voyage Beagle, Fishes, 121, pl. xxii, fig. 5,
1842. (Bay Islands, New Zealand.)

Galaxtas attenuatus Ginther (in part), Cat. Fish. Brit. Mus., VI, 210, 1866
(Falkland Islands; Peru?); Giinther, Zool. Coll. H. M. S. Alert, 21,
1881 (Puerto Bueno); Vaillant, Miss. Scient. Cap Horn, Poiss. C. 19,
1888 (Orange Bay); Perugia, Ann. Mus. Civico Storia Nat. Genova,
(2a), X, 54, 1891 (Lake and River of Porto Cook); Philippi, Verh.
Deutsch. Wiss. Ver. Santiago, Chile, III, 21, 1895. Delfin, Catalogo
Peces Chile, 33, 1901 (Chiloe); Dollo, Voy. Du S. Y. Belgica, pl. x,
fig. 5.

Galaxias maculatus (non Jenyns) Richardson, Voy. Erebus and Terror,
75, pl. xvi, figs. 14-117, 1847 (Falkland Islands); Smitt, Bih.
Svenska Akad. Handl. XXVI, IV, No. 13, 1901, p. 21, pl. ii, figs.
5-8 (Rio Pescado and Ultima Esperanza).

' This is undoubtedly the young of a¢tenuatus.

EIGENMANN: FRESH WATER FISHES. 275

Mesites gracilimus Canestrini, Arch. Zool. Anat. Fisiol. III, 1864, 100,
tav4y fe2te (Chile):

Galaxias gracillimus Ginther, Cat. Fish. Brit. Mus. VI, 213, 1866; Philippi,
Verh. Deutsch. Wiss. Ver. Santiago, Chile, II], 21, 1895; Delfin,
Catalogo Peces Chile, 34, 1901; Regan, Proc. Zool. Soc., London,
1905, II, 370 (Falkland Islands).

?Galaxias minutus Philippi, Wiegemann’s Archiv. XXIV, 1858, 309;
Verh. Deutsch. Wiss. Ver. Santiago, Chile, III, 21, 1895 (Valdivia) ;
Delfin; Lc s4n100%

? Galaxias punctatus Philippi, Wiegemann’s Archiv., XXIV, 1858, 310;
Verh. Deutsch. Wiss. Ver. Santiago, Chile, III, 21, 1895; Delfin,
l.c. (Valdivia and Puerto Montt.)

Galaxias alpinus (non Jenyns) Smitt, 21, pl. i, figs. g-12, (Rio Azo-
pardo; Rio Pescao; Rio Gallegos; alpine lakes of the Sierra Toro
and Lago Toro basin); Delfin, I. c., 34, 1901.

Habitat: (Peru?) Puerto Montt south to Falkland Islands and Aus-
tralia, Tasmania, New Zealand. No specimens were secured by Hatcher.

17. GALAXIAS MACULATUS (Jenyns).
(Plate XXXV, Fig. 4.)

Stomias variegatus Lesson, Voy. La Coquille, II, 142, 1830 (Malouines).

Mesites maculatus Jenyns, Zool. Voyage Beagle, Fishes, 119, pl. xxii,
fig. 4, 1842 (Fresh-water brook in Hardy Peninsula, Tierra del Fuego ;
higher tributaries of the Rio Santa Cruz).

Galaxias maculatus Cuvier & Valenciennes, Hist. Nat. Poiss. XVIII, 355,
1848 (Malouines); Giinther, Cat. Fish. Brit. Mus., VI, 212, 1866;
Vaillant, Miss. Scient. Cap Horn, Poiss. C. 18, 1888 (Fresh-water of
Orange Bay); Perugia, Ann. Mus. Civico Storia Nat. Genova (2@) X, 54,
1891 (Lake and River of Porto Cook) ; Philippi, Verh. Deutsch. Wiss.
Ver. Santiago, III, 21, 1895 ; Steindachner, Zool. Jahrb., 1898, Suppl.
IV, 328 (Rio Pescado near Punta Arenas); Delfin, Catalogo Peces
Chile, 33, 1901 (Puerto Montt; Valdivia); Dollo, Voy. Du S. Y.
Belgica, 80, pl. x, fig. 4, 1904 (Lapataia, Beagle Channel); Regan,
Proc, Zool. Soc, ond, 1905, II, 370 (Alert: Bay); Orange “Bay; ;
Falkland Islands, Esterode Penco) : Evermann & Kendall, Proc. U.
S. Nat. Mus. XXXI, 1906, 91 (Lake Nahuel Huapi, one of the
sources of the Limay River; elevation 2,500 feet).

276 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Galaxtas coppingeri Giinther, Proc. Zool. Soc. London, 1881, 21 (Alert
Bay); aud, Zool.“Goll: Hi: M.S. Alert, 21, 1636p (AlerteBay:)
Galaxtas alpinus (non Jenyns) Smitt, Bih. Svenska Akad. Handl. XXIV,

IV, No. 5, 1898, p. 56, pl. v, figs. 40-402.
Habitat: Falkland Islands, Tierra del Fuego, Patagonia.

18. GALAXIAS PLATEI Steindachner.
(Plate XXXV, Figs. 1, 2, 2a, 3 and 3a.)

Galaxias platet Steindachner, Zool. Jahrb., 1898, Supplement IV, Zweites
Heft, 329 (Rio Pescado near Punta Arenas); Delfin, Catalogo Peces
Chile, 34, 1901. |

Galaxias alpinus (part) Smitt, Bih. Svenska Akad. Handl. XXVI, IV,
No. 13, p. 9, pl. \i11,-1901.

?Galaxias grandis Philippi,’ Verh. Deutsch. Wiss. Ver. Santiago, Chile
III, 19, 1893 ; Delfin, Catalogo Peces Chile, 33, 1901 (Punta Arenas).

? Galaxias delfini Philippi, Verh. Deutsch. Wiss. Ver. Santiago, Chile III,
19, 1893; Delfin, Catalogo Peces Chile, 33, 1901 (Punta Arenas).

? Galaxias titcombt Evermann & Kendall,? Proc. U. S. Nat. Mus. XXXI,
1906, 92 (Rio Traful, Argentina).

' Philippi described two species of Ga/axias as follows :

Galaxias grandis Philippi, p. 19, based on a specimen shrivelled by strong alcohol and
from which all fins but the caudal had been removed, froma lake north of Punta Arenas. Length
without caudal 292 mm., with caudal 330 mm.; depth 3.4 in the length, width 3.56; depth of
caudal peduncle 10.8, its width 32.4 in the length; head 4.5. Head low, flat above and little
arched; mouth ‘small, eye 9 + in the head; four ridges on preopercle. Dorsal profile more
strongly arched than in other species of Ga/axias, ventral profile but little arched. Head light
gray, body yellowish, lighter below; numerous spots 11% mm. in diameter everywhere on the
body, fainter ones on upper part of head.

Galaxias delfini Philippi, p. 19, based on a specimen preserved in such strong alcohol that
it had to be soaked in water before its characters could be made out. From a lake north of
Punta Arenas. Length without caudal 192 mm.; depth 5.8 in the length; width less than
depth ; depth of caudal peduncle 1.2 in the length; head a little over 4.5 ; depth of head greater
than in grandis but equally flat; eye 9 in head; dorsal and ventral profiles nearly straight ; base
of dorsal 8.7 in the length, its height 7; back light yellow, sides light gray ; back with numerous
partially confluent brown spots; fins unspotted. D. 8; A. 12.

* This species, which may be distinct, was published as this report was going through the press.
It was described as follows :

Head 4 in length without caudal; depth 5.94; eye 4.23 in head; snout 3.92; D. 10; A. 11;
snout bluntish; eye moderate, slightly shorter than snout; dorsal outline arching slightly from
occiput, thence nearly straight to front of dorsal ; height of dorsal 8.56 in length without caudal,
its base about 2.20 in head, the first rays when depressed not reaching tip of last rays; distance

EIGENMANN: FRESH WATER FISHES. 277

Habitat: Chili to Punta Arenas and Rio Chico.

Twenty-four specimens, ranging from 49-96 mm., were collected by
Hatcher in the Rio Chico, 15 miles below Rio Belgrano, and eight be-
tween 48 and 113 mm. near Punta Arenas. The oldest specimen is finely
mottled with irregular light cross streaks. The light becomes more marked
in the specimen 90 mm. long. Of two specimens 87 mm. long one is
colored like the largest, the other has dark blotches and cross streaks on
a yellow background. With a decrease in the size there is an irregular

decrease in the amount of pigment which is in cross streaks and irregular
blotches.

Family VI. 4APLOCHITONID:.

Haflochitonide Giinther, Cat. Fish. Brit. Mus. V, 381, 1864.

The family is composed of two genera, the naked 4f/ochiton with two
species in Chili and Patagonia, and the scaled Profotroctes with three
species in New Zealand, Queensland and South Australia.

The species of Alochiton have also received detailed consideration
from Smitt.

10, AAPLOCHITON jJenyns.

A plochiton Jenyns, Voy. Beagle, Fishes, 131, 18.
Fartonella Cuvier & Valenciennes, Hist. Nat. Poiss. XXII, 507.
Haplochiton Ginther, Cat. Fish. Brit. Mus. V, 381, 1864.

SPECIES OF APLOCHITON.
a. D. 12-13; A. 15-16.
6. Head 22.5-25.9 per cent. of the length; width of head 11.3—12.4 per cent. ; olivaceous,
darker above, a series of narrow dark cross bars. 19. sebra.

from tip of snout to origin of dorsal about 1.46 in length without caudal ; height of anal about
8.46 in same length, its base about 2.20 in head, the tips of first rays not reaching tips of last
when depressed ; distance from tip of snout to anal origin about 1.32 in the length without caudal ;
pectoral short, rounded; ventral very short, about 2.5 in head, the distance from its origin to
base of pectoral about 3.35 in length without caudal, and distance from its origin to point of anal
about 5.25 in same length; caudal deeply emarginate.

Color, very pale gray, slightly more dusky on back from thick minute punctulations ; irregular
groups of black dots on side extending not quite to belly, giving a clouded effect and the appear-
ance of broken and entire crossbars ; belly pale, with very few dots in front of ventral ; a row of
black dots from base of each ventral to each side of vent; fins pale, with some punctulations, head
thickly punctulated above, on snout, and on side about to level of upper jaw, abruptly pale below.

Type. — A specimen 5.62 inches long, collected December 13, 1903, by Mr. John W. Titcomb,
from Rio Traful near Lake Traful, Argentina.

278 PATAGONIAN EXPEDITIONS: ZOOLOGY.

66. Head 20-21.1 per cent. of the length ; width of head 9.2-9.5. Lighter, no bars.
20. tentatus.

19. APLOCHITON ZEBRA Jenyns.
(Plate XX XV, Figs. 5 and 52.)

Aplochiton zebra Jenyns, Zool. of the Voyage of the Beagle, Fishes,
131, pl. xxiv, fig. 1 (Falkland Islands).

Hafplochiton zebra Giinther, Cat. Fish. Brit. Mus. V, 381, 1864 (Port
Louis) Gtinther, Zool. Coll. H.'M. S. Alert 1681, 22 (ast: Bay,
freshwaters at Tom Bay); id. Rep. Voy. Challenger, Shore Fishes,
23, 1889; Smitt, Bih. Svenska Akad. Handlinger, XX VI, Afd. IV,
No. 13, 1901, 1 (Rio Tres Pasos, a tributary of Lago Toro); Delfin,
Catalogo Peces Chile 32 (Puerto Montt; Rio Renaico; Puerto Otuai).

Frartonella gayi, Cuvier & Valenciennes, Hist. Nat. Poiss. XXII, 508, pl.
640.

Flabitat: Puerto Montt, south to the Falkland Islands.

20. APLOCHITON TAENIATUS Jenyns.
(Plate XX XV, Figs. 6 and 62a.)

A plochiton tentatus Jenyns, Zool. of the Voy. of the Beagle, Fishes, 132,
1842, pl. xxiv, fig. 2 (Mouth of freshwater streams, Gore Sound,
Tierra del Fuego).

Flaplochiton tentatus Giinther, Cat. Fish. Brit. Mus. V, 382, 1864 (Gore
Sound); Smitt, Bih. Svenska Akad. Handl. XXVI, Afd. IV, 1, 1901
(Lago Toro) ; Delfin, Catalogo Peces Chile, 32, 1901 ; Dollo, Voy. du
S. Y. Belgica, 81, 1904 (Lapataia, Beagle Channel).

FHlabitat: Southern Patagonia and Tierra del Fuego.

Order V. ACANTHOPTERI.

Anterior vertebrze unmodified ; no mesocoracoid and no interclavicles ;
mouth border formed by premaxillary; shoulder girdle attached to skull
by a post temporal; anterior rays of dorsal and anal usually spinous; air
bladder usually without duct.

Including the majority of marine fishes and many fresh-water species.

EIGENMANN: FRESH WATER FISHES. 279

Suborder PERCESOCES.

Cope, Trans. Am. Philos. Soc. XIV, 456.

Ventral fins abdominal, with one spine and five soft rays; gills and
mouth normal.

Two of the families of this suborder have representatives in South
American fresh waters. They may be distinguished as follows:

a. Ribs attached to strong parapophyses ; teeth small or wanting’; gill rakers long and slender ;
pectoral fins inserted high up.

6. Pelvic bones free or connected with the clavicle by ligament ; carnivorous ; head elongate ;
vertebrae more than 35; dorsal spines slender, flexible, 1-8 in number; stomach not
gizzard-like. Atherinide.

26. Pelvic bones suspended from the postclavicles ; teeth feeble or absent, limnophagous ;

head short and broad ; vertebrae about 24; stomach gizzard-like ; intestines very long.
Mugilide.

The Mugilide are chiefly marine, some enter or are permanently
established in fresh water, none are recorded from Patagonian fresh waters.

Family VII. 47HERINIDA:.

The Atherinide, like the Mugilide are chiefly marine. A number of
genera or species have become established in fresh water; thus species
of Atherina and Mentdia are established in fresh water in Cuba and
North America. Lad/desthes is a genus confined to the fresh waters of the
eastern slope of the United States. Lethostole and Chirostoma are fresh-
water fishes of Mexico. /vof/stius is found in the Peruvian Andes at an
altitude of 12,000 feet ; Gastropterus on the Pacific slope to an elevation
of 7,500 feet.

Along the coast of Chili and Patagonia and entering its rivers or
established in its lakes are found several species of A¢herinopsis and
Menidia.

The genera represented in the waters of Patagonia may be distinguished
as follows:

a. Premaxillaries freely protractile, broad posteriorly, their edge strongly curved ; jaws nearly
equal, the lower slightly shorter, forming a pointed muzzle ; scales large, smooth ; first dorsal
rounded ; belly rounded ; anal about 14—19 ; teeth minute, in bands; dorsal and anal naked.

Menidia.
aa, Premaxillaries not protractile, the skin of upper jaw mesially continuous with that of the
forehead ; teeth simple, pointed, in villiform bands ; vomer with teeth or not. -A¢herinopsis.

280 PATAGONIAN EXPEDITIONS: ZOOLOGY.

11. MENIDIA Bonaparte.

Menidia Bonaparte, Fauna Italica 1836 (smenzda).
? Chivostoma Swainson, Fishes 1839, 243 (Aumboldtianum).
Argyrea Dekay, New York Fauna, Fishes, 141, 1842 (sofa/a).
Atherinoides Bleeker, Verhand., Batay. Genootsch. Japan, XX, 40, 1853
(vomerina).
? Atherinichthys Bleeker, 1. c., 40 (Aumboldtianum).
? Heterognathus Girard, Proc. Ac. Nat. Sci. Phila. 1854, 198 (Aum-
boldtianum).
? Lethestole Jordan & Evermann, Bull. U.S. Nat. Mus. 47, 792, 1896 (esfor).
E:islopsarum Jordan & Evermann, Check List Fishes, 330, 1896 (jordanz ).
If only the long jawed forms of Chzvostoma of Mexico were consid-
ered, they could readily be generically distinguished from AZenzdia. But
Meek has shown that while some species have a projecting lower jaw,
others have the jaws equal and still others have the upper jaw projecting.
These, according to Meek, intergrade to such an extent that no generic
dividing line can be drawn between Mexican species. Inasmuch as the
short jawed forms appear to be true A/enzdzas it is a question whether the
genus Chzvostoma should not be united with A/enzaia.
Smitt and Fowler have made out that Menzdia laticlava is synonymous
with 4. vegius. Steindachner on the other hand found that they are
generically distinct.

21. mauleana.

aa. Head 5.5-6; depth 6.6—7; D. viii, 10; A. i, 18; Lat. line 86. 22. laticlava.
aaa.’ Head 5.5; depth 5.5; D. vii, 9; A. i, 16; Lat. line 68-70; eye a little more than 4.4 in
head. 23. hatcheri sp. nov.

aaaa, Head 5; depth 8.5; D. vi-viii, 10; A. i, 18-i, 20; Lat. line 100; eye 3-5 in head, a
conspicuous silvery lateral band; each scale of the back with one or more dots ; origin
of spinous dorsal over tips of ventrals. patagoniensis sp. Nov.

21. MENIDIA MAULEANA (Steindachner).

Chirostoma mauleanum Steindachner, Ann. Naturh. Hofmus. V, II, 1896,
213 (Pichi Laguna, a branch of Lake Llanquihue) ; id., Zool. Jahrb.
1898, Suppl. IV, 313.

This species is known only from the types.

“Hatcher collected three small specimens, probably at Sandy Point (Punta Arenas), Straits of
Magellan.

EIGENMANN: FRESH WATER FISHES. 281

22. MENIDIA LATICLAVA (Cuvier & Valenciennes).

Atherina laticlava Cuvier & Valenciennes, Hist. Nat. Poiss. X, 473,
1835 (Valparaiso; Laguna de la Taguatugua) ; Guichenot, in Gay
Hist. Chile, Zool. II, 252, 1848, and plate iv, fig. 1, 1854.

Atherinichthys laticlava Giinther, Cat. Fish. Brit. Mus. III, 402 (Valpa-
raiso; Falkland Islands; Port Louis); Cunningham, Trans. Linn.
Soc. London, XXVII, 471, 1871 (Punta Arenas, Port Famine, Port
Gallant, all on Str. of Magellan); Vaillant, Miss. Scient. Cap Horn,
Poiss. C. 22, 1888°(Rio Pesca); Perugia, Ann. Mus. Civ. Stor. Nat.
di Genova, Ser. 23, X (xxx), 620, 1891 (mouth of the Rio Negro) ;
Delfin, Catalogo Peces Chile, 46, 1901.

23. MENIDIA HATCHERI spec. nov.
(Plate XX XVII, Fig. 4.)

Type. Asingle specimen 219 mm. to base of caudal. Lake Pueyrredon.

Elongate; head 5% ; depth 5% ; D. vi-1, 9; A. 1, 16; lat. line 68-70.
Snout 3.75 in head, eye 4.4, 1.6 in interorbital; mouth small, the lower
jaw included when closed; snout and premaxillary equal to length of
eye; snout and maxillary equals length of snout, the maxillary not reach-
ing the eye. Teeth in narrow bands in each jaw, persistent, those of the
outer jaw slightly enlarged; no teeth on vomer.

Spinous dorsal nearer caudal than tip of snout, its origin about over
tip of ventrals, its tip over anus; origin of second dorsal over second fourth
of anal, its tip not reaching tip of anal. Caudal widely forked; anal
emarginate, its last ray a little longer than eye, considerably longer than
some of the preceding rays, its longest anterior ray equal to half the dis-
tance between tip of snout and origin of pectoral, equal to length of ven-
trals ; ventrals reaching very little more than half way to anal; pectorals
short, about 6% in the length, reaching half way to middle of ventrals.

A silvery lateral band, margined with plumbeous above; each scale of
head and body with a band of chromatophores along the margin, those
of the back more numerous than below ; all the fins dusky, the membranes
being densely pigmented. Scales cycloid.

282 PATAGONIAN EXPEDITIONS: ZOOLOGY.

12. ATHERINOPSIS Girard.

A therinopsis Girard, Proc. Ac. Nat. Sci., Phila. VII, 1854, 134 (calfor-
uiensts).

Basilichthys Girard, Proc. Acad. Nat. Sci. Phila. VII, 1854, 198, and U.
S. Naval and Astronomical Expedition, 238, 1855 (secrolepidotius).

The genus Baszlichthys is distinguished from A/herinopsis solely on
the relative length of the jaws. Girard says: “The genus Baszlichthys
will be characterized by the protrusion of the upper jaw beyond the lower.”’
It was based on Atherina microlepidota, by Steindachner considered a
synonym of the 4¢herina regia of Humboldt.

The Atherinopsts is defined as having ‘‘no palatine teeth, with both
jaws equal, and the snout more or less rounded.” Until specimens
can be directly compared, the genus Basz/ichthys may be united with
Atherinopsts.

There is but a single species of this genus in Patagonian waters. If,
as Steindachner thinks, the species is confined to the Pacific coast from
Valparaiso northward, the reference to this species south of that point
may belong to Menidta laticlava.

24. ATHERINOPSIS REGIUS (Humboldt).

Atherina regia Wumboldt, Rec. Observ. Zool. Anat. Comp. II, 187, 1835
(Callao; Lima); Cuvier and Valenciennes, Hist. Nat. Poiss, X, 474,
1835.

? Atherinichthys regia Smitt, Bihang Svenska Akad. Handl. XXVI, Afd.
IV, No. 5, 93, pl. ix, figs. 30-31, 1898 (in part).

Atherinopsts rvegius Steindachner, Dk. Ak. Wiss. Wien, LXXII, 39, 1902
(Rio Tambo, South Peru).

? Atherinopsts regia Hatcher, these reports, Vol. I, 280, 1903 (Inlet of the
Gallegos river).

Atherina microlepidota Jenyns, Voy. Beagle, Fishes, IV, 68, pl. xvi, fig.
1, 1842 (Valparaiso) ; Guichenot, Gay Hist. Chile, Zool. II, 253, 1848.

Basilichthys microlepidotus Girard, Proc. Acad. Nat. Sci. Phila. VII, 1854,
198; id. U. S. Nav. and Astron. Exped., 238, plate xxx, figs. 8 and
9, 1855 (Valparaiso; Mapocho, a tributary of the Maypu); Ever-
mann & Kendall, Proc. U. S. Nat. Mus. XXXI, 1906, 97 (Lakes
Nahuel Huapi and Traful; Nirihuahay, tributary of Rio Limay).

EIGENMANN: FRESH WATER FISHES. 283

Atherinichthys microlepidota Giinther, Cat. Fish. Brit. Mus. III, 403, 1861 ;
Kner, Novara, Fisch. 222, 1869; ? Perugia, Ann. del Mus. Civ. di
Stor. Nat. Ser. 2a, X (XXX), 620, 1891 (Mouth of the Rio Negro) ;
Berg, Ann. Mus. Nac. Buenos Aires, IV, 28, 1895; Delfin, Catalogo
Peces Chile 47, 1901.

Atherina laticlava Cope, Proc. Am. Philos. Soc. XVII, 1878, 44 (Callao).

Habitat : (2? Montevideo to) Valparaiso and northward in salt and fresh
water.

No specimens of this species were collected by Hatcher. He says of it:

‘One species, A/herinopsis regia, of particularly fine flavor, is especially
common at certain seasons along the east coast. So abundant were these
fish in the inlet of the Gallegos river for several days during the spring
of 1898 that on one occasion they were brought in great schools by the
incoming tide and left stranded in such quantities on the shingle bed of
the north coast that, for a distance of twelve miles, or over, the beach
extending from North Gallegos to two or three miles above Killik Aike,
the shingle was covered to an average width of ten yards and to the depth
often of several inches with dead fish.”

Suborder PErRcomorputi.

Cope; Proc, A. Av A. S:, 1371, Indianapolis; 341.
A dominant group of fishes of fresh and especially of salt water.
Ventrals thoracic, with a spine and 4 or 5 rays ; lower pharyngeals usually
separate ; gills 4, a slit behind the fourth; nares double on each side ;
scales ctenoid ; post temporal slender, divided at tip and not codssified with
skull ; bones of jaw distinct ; pectoral actinosts normal ; vertebrae 24—40.

Family VIII. SERRANIDAZ.

Maxillary usually not sheathed by the preorbital (sheathed in Patagonian
species) ; anterior vertebrae without transverse processes ; all or most of
the ribs inserted on the transverse processes when these are developed ;
anal spines 3; pseudobranchiz well developed ; vomer with teeth ; lateral
line extending to base of caudal ; anal shorter than dorsal; head not every-
where covered with rough scales ; second suborbital with an internal lamina
supporting the globe of the eye; entopterygoid present; six or seven
branchiostegals.

284 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Of this widely distributed family, whose species are chiefly marine, two
genera with five species are found in South American fresh waters. Two
of the species are found in Patagonia, two others are confined to Central
Chili, and the fifth is found near Buenos Aires. The genera may be distin-
guished by the following characters :

a. Two dorsals ; pectoral asymmetrical, upper rays longest ; ventrals below ora little behind base

of pectoral. Tongue smooth ; scales ciliate; vertebrae 33-35.
6. Maxillary with supplemental bone ; ventral below posterior edge of base of pectoral ;

palatine toothed ; scales small. Percichthys.
66. Maxillary without supplemental bone ; ventral a little behind base of pectoral ; palatine
toothless ; scales larger. Percilia.

13. PERCICHITHYS Girard:

Percichthys Girard, Proc. Acad. Nat. Sci. Phila., VIF, 1854, 197 and U.S.
Naval and Astronom. Exped. 230, 1855. (dvucha.)
Type. Perca trucha Cuvier & Valenciennes.

SPECIES OF PERCICHTHYS.

a. Depth less than length of head.
6. Sides of head and body usually spotted ; highest dorsal spine less than half the length of

head in adult ; scales 8 to 10-59 to 67—22 to 25. 25. trucha.
66. Sides of head plain ; highest dorsal spine more than half the length of head in the adult ;
lateral line 66-70. 26. vinciguerre.

aa. Depth equals length of head.
c. Maxillary extending to below anterior border of eye ; 10-13 spines on lower border of

preopercle. Scales 10 or 11-54 to 58—Ig or 20. 27. melanops.
cc. Maxillary extending a little beyond the margin of the eye ; 4-7 spines on the lower limb
of preopercle ; scales 8-66 to 72—23. 28. altipinnis.

25. PERCICHTHYS TRUCHA (Cuvier & Valenciennes).
(Plate XXXVI, Fig. 3, and Plate XX XVII, Fig. 2.)

Perca trucha Cuvier & Valenciennes, Hist. Nat. Poiss. IX, 429, 1833
(Rio Negro, Chile) ; Guichenot, in Gay, Hist. Chile, Zool. II, 146,
1848 (Chile) ; id. Atl. Ichthiol. I, v, fig. 1, 1854.

Percichthys trucha Girard, Proc. Ac. Nat. Sci. Phila. VII, 1854, p. 197,
and U. S. Naval and Astronom. Exped. II, 230, 1855 ; Giinther, Cat.
Fish. Brit. Mus. I, 61 ; Jordan & Eigenmann, Bull. U. S. Fish. Com.
VIII, 427, 1890 (Santiago and Curico, Chili) ; Boulenger, Mem. Soc.
Se Chile, IV, 10, 180947; 1d.°Cat. Fish. (Brit. Mus. jaime) eiees-

e

EIGENMANN: FRESH WATER FISHES. 285

Steindachner, Zool. Jahrb. 1898, Suppl. IV, Zweites Heft, 281 (Pichi-
Laguna, an arm of Lake Llanquihue at Puerto Montt) ; Delfin, Cata-
logo Peces Chile, 58, 1901; Evermann & Kendall, Proc. U. S. Nat.
Mus. XXXI, 1901, 100 (Rio Negro; Rio Limay; Lake Nahuel
Huapi; Lake Traful).

Perca levis Jenyns, Zool. Beagle, Fishes I, pl. 1, 1842 (Rio Santa Cruz).

Percichthys tevis Giinther, Cat. Fish. Brit. Mus. I, 61; Kner, Novara
Fische II, 1865 (Valparaiso); Vaillant, Miss. Scient. Cap Horn,
Poiss. C. 31, 1888 (Santa Cruz); Jordan & Eigenmann, Bull. U. S.
Fish Com. VIII, 428, 1890; Perugia, Ann. del Mus. Civ. Stor. Nat.
Genova, Ser. 2a, X (XXX) 609, 1891 (Tierra del Fuego; Laguna
del Rio Negro; Rio Santa Cruz).

Percichthys chilensis Girard, Proc. Acad. Nat. Sci. Phila., VII, 1854, 197;
id, U. S. Naval and Astron. Exped. 231, pl. xxix, figs. 1-4, 1855
(Rio de Maypu, near Santiago, Chili); Philippi, Mon. Berl. Acad.
1866, 708.

Flabitat: Rio Negro and Valparaiso and southward. Four speci-
mens were taken by Hatcher in the Rio Santa Cruz; the locality of seven
others is not indicated.

The following description is based on these specimens and on 12 speci-
mens from Santiago and Curico, Chill.

Body elongate, deepest below first dorsal spine; maxillary reaching to
below pupil, or little beyond anterior border of eye, 25 to 3 inhead. Max-
illary and mandibular teeth in similar bands, broadest in front and taper-
ing backwards; teeth on vomer in a triangular patch; palatine bands of
teeth much longer than those on vomer, separated from the latter, and
placed almost at right angles to them. Head covered with scales forward
to the anterior nostril, a short linear naked area, always present, at or
near the base of the supraoccipital keel. Scales on cheeks in twelve to
fifteen irregular series; scales on opercle as large as those on body, in about
six series. Eye large, 13 to 13 in snout, 5 to 53 in head, 1 to 1% in in-
terorbital space. Mouth terminal or subterminal, the lower jaw slightly
included. Profile straight or very slightly concave from tip of occipital
process to premaxillary process, thence abruptly decurved.

Preorbital with strong teeth directed downward and backward, strong-
est in young examples, largest near posterior angle of maxillary ; entire
vertical margin of preopercle with fine teeth, largest below; lower mar-

286 PATAGONIAN EXPEDITIONS: ZOOLOGY.

gin of preopercle with larger, wider-set teeth, the anterior ones directed
forward ; posterior half of free edge of interopercle and lower half of
subopercle with very fine teeth, which become more or less obsolete with
age; opercle with a strong spine and a blunt or rounded point above it.

Gill-rakers short, chubby, about two-thirds the length of the pupil, 6+
13; inner side of the gill-rakers covered with short, stout teeth.

Distance of first dorsal spine from tip of snout, 2% in length of body.
First dorsal spine less than half the length of the second, the second from
one-half to two-thirds length of the third spine, which is the highest, 25
in the head in adult, 13 in young, the spines decreasing in height to the
ninth; the spinous and soft dorsals connected. Caudalin the young
slightly emarginate, the upper part truncate, lower rounded. First anal
spine inserted under second dorsal ray, the spines graduated, the second
strongest, highest ray 2-2% in head. Ventral inserted below the base
of lower pectoral rays, the second divided ray longest, 23 to 14 in head.
Pectoral 2 to 1% in head.

Scales of body of about uniform size, becoming very much smaller on
breast and top of head. Scales strongly ctenoid on sides.

Small scales on the caudal membrane at its basal third. Anal and
dorsal without scales.

Color olivaceous yellow, with peppery black dots, aggregated in spots
on the back; the scales along base of dorsal and the upper half of cau-
dal peduncle with a brownish spot at their base, spots forming more or
less regular longitudinal lines; membrane of soft dorsal with minute
brownish dots, aggregated in places to form rather large spots. Membrane
of caudal dusky; anal with brownish dots along the middle of the mem-
branes ; pectorals and ventrals with similar, but fewer, spots. Head 3% —
32 in length, to base of caudal; depth 4 to 42; D. IX—XI, 11 or 12; A.
III, 8-10. Scales 9-59 to 67-17. This species reaches a length of 480
mm.

26. PERCICHTHYS VINCIGUERR Perugia.
(Plate XXXVII, Fig. 3.)

Percichthys vinciguerre Perugia, Ann. del Mus. Civ. Stor. Nat. Genova,
Ser. 2a, X (XXX), 610, 1891 (Rio Santa Cruz); Boulenger, Cat.
Fishes, I, 120, 1895.

Through the courtesy of Dr. R. Gestro I have received one of the

EIGENMANN: FRESH WATER FISHES. 287

types of Percichthys vinciguerre Perugia and also a specimen of nearly
the same size of P. devis = trucha. The species is very closely allied if
not identical with ¢rvwcha. The characters on which Perugia laid stress as
distinguishing this species are not present in the type sent me. Other
characters which readily distinguish this specimen are different according
to Perugia, in other specimens of vzxuczguerre. The only characters dis-
tinguishing the type of wxceguerre sent from the specimen of “vucha of
nearly equal size are the following :

a. Head and sides plain in color; depth 4% in length to caudal; lat. 1. 70 (about); highest

dorsal spine more than half length of head. VINCLZUEITE.
aa, Sides of head, above line back from angle of mouth, and sides of body spotted ; depth 4 in the
length ; lat. 1. 65 ; highest dorsal spine less than half length of head. trucha.

The extent of the maxillary made use of by Perugia is not diagnostic.
Taking the distance from tip of snout to end of maxillary with dividers
and comparing it with distance of tip of snout from eye, it is found that in
the three specimens of ¢vucha collected by Hatcher and the ¢vucha and
vinciguerre collected by the Italian expedition :

1. Hatcher specimen ; maxillary reaches middle of eye.

2. Hatcher specimen; maxillary reaches to anterior margin of pupil.

3. Hatcher specimen; maxillary reaches to anterior margin of pupil.

4. P. Zevrs collected by Italian expedition ; maxillary reaches posterior
margin of pupil.

5. P. vinciguerre ; maxillary reaches to middle of pupil.

The number of spines on the lower limb of the left preopercle of the
five specimens are respectively 8, 11, 8, II, II.

The scales in the lateral line are 61, 65, 59, 65, 70.

The greatest depth of body in the length to base of caudal is respec-
tively 4, 4, 4, 4, 45, the specimen of winciguerre being distinctly more
slender than the others. I am inclined to think that the depth ‘334 ” of
Perugia must be a misprint for 434.

In a specimen collected by Hatcher with a length of 210 mm. the
highest dorsal spine is equal to half the length of the head ; in a specimen
160 mm. long it is equal to the head less the opercle.

Elongate slender, deepest in space between base of ventrals and first
dorsal spine; maxillary reaching to below middle of pupil, 3 in the head;
teeth as in /vucha; squamation of head similar to that of fvwcha but the
occipital crest and parietals striate and naked (in this specimen only ?) ;

288 PATAGONIAN EXPEDITIONS: ZOOLOGY.

scales of cheek irregular in size, those near edge the smallest; opercular
scales as large as those on body. Eye 1% in snout, 13 in interorbital,
65 in head; mouth terminal, jaws equal, profile slightly concave from
middle of occipital process to tip of maxillary process.

Preorbital serrate; vertical limb of preopercle finely serrate, lower limb
of preopercle with strong antrorse spines, more or less aggregated in
groups of two or three.

Opercular spine as in ¢vwcha, subopercle and interopercle finely serrate.

Gill-rakers about ¢ the diameter of pupil, 9 + 13.

Distance of first dorsal spine from tip of snout 27 in the length; fourth
dorsal spine highest, more than half length of head; caudal emarginate ;
ventrals 13 in head, pectorals 134. Scales as in fvucha.

Sides yellowish wethout spots aside from the chromatophores giving the
darker color of the back; left dorsal and caudal very faintly spotted, other
fins plain.

D. x, 11 or 12; A. 11, 9; scales 7% to 9-66 to 70-18.

27. PERCICHTHYS MELANOPS (Girard).
(Plate XXXVI, Figs. 2, 2a and 26.)

Percichthys melanops Girard, Proc. Acad. Nat. Sci. Phila. VII, 1854, 197
(Rio de Maypu, Chili); id.,U. S. Nav. and Astron. Exped. 233, pl.
xxx, figs. I-5, 1855 (Rio de Maypu, Chili); Giinther, Cat. Fish. Brit.
Mus. I, 61, 1859; Jordan & Eigenmann, Bull. U. S. Fish Com.
VIII, 428, 1890; Boulenger, Mem. Soc. Sc. Chile IV, 13, 1894; id.,
Cat. Fish. Brit. Mus. I, 120, 1895; Delfin, Catalogo Peces de Chile,
(Rio Maipo, Central Chili; Colima).

Percosoma melanops Gill, Proc. Acad. Nat. Sci. Phila. 51, 1861.

Perca pocha Philippi, Arch. Wiegm. 1863, 210 (Santiago, Chili).

Percichthys pocha Jordan & Eigenmann, Bull. U. S. Fish. Com. VIII,
428, 1890 (Curico, Chili).

The southernmost record for this species is Curico, Chili, about 100
miles north of the source of the Neuquen on the Pacific slope.

Body ovate, deepest below first dorsal spine. Maxillary reaching
scarcely to vertical from anterior margin of orbit, 3 in head. Teeth of
lower jaw in a band widest near tip, and tapering to a single series behind,
some of the lateral teeth longer than the others; teeth of upper jaw in a

EIGENMANN: FRESH WATER FISHES. e 289

broader band, those of the sides not in a single series, teeth all about
equal; teeth on vomer in a very narrow, crescent-shaped patch; those on
palatines in a band much narrower and shorter than that on the vomer.
Mouth oblique, the jaws subequal, the lower slightly included. Head
scaled forward to the anterior nostril; scales on cheek in about eight
series, those on opercle about as large as those on body, in about six
series. Profile straight, from anterior margin of orbit to tip of occipital
crest, rounded in front and behind. Eye 1% in snout, 4% to 5 in head;
interorbital area a little wider than eye. Preorbital minutely serrate, the
serree weaker than in P. frucha. Preopercle with minute teeth on its
vertical border, the teeth near the angle sometimes very much enlarged,
sometimes little enlarged, more numerous than in P. ¢vucha; serration
of the subopercle and preopercle scarcely visible; opercular spine placed
higher than in P. /vucha, its tip sometimes incompletely two or three
parted; a bluntish projection on opercle above the spine. Gill-rakers
very short, about equal to one-third diameter of eye, 6 + 11. Distance
of first dorsal spine from snout 27 to 27 in the length. Height of dorsal
spines variable, the first always less’ than half as high as the second, the
third or fourth dorsal spine highest, 2 to 3 in head, the spines decreasing in
height to the ninth; spinous and soft dorsal connected. Caudal truncate
when spread out, emarginate when closed. Anal inserted below the be-
ginning of the soft dorsal, its spines graduated, the second thickest; highest
ray about half as long as the head; ventral 17 in head; pectoral 1% to
17 in head.

Scales on the body of about equal size, less strongly ctenoid than in
P. trucha, reduced on breast and head. Lateral line much more strongly
curved than in P. frucha.

Color brownish, golden-yellow below, everywhere with brownish dots ;
those on the lower half of the body scattered with usually a light (blue ?)
center, a dusky spot at the base of each scale on the sides; all the fins
dusky, with reddish brown dots; those on the base of the soft dorsal
sometimes aggregated into spots. Head 3 to 3% in length to base of
eatidals. depth:3) tors 477 Dax 11! or. 725) AN 1, 9 6r 10.1 Scales 10: oF
11-54 to 58-19 or 20.

This species reaches a length of 180 mm.

290 PATAGONIAN EXPEDITIONS: ZOOLOGY.

28. PERCICHTHYS ALTIPINNIS Regan.
(Plate XX XVII, Fig. 1.)

Percichthys altipinnis Regan, Revue Suisse de Zoologie 13, 390, 1905
(Rio Colorado, Buenos Aires). Dorsal x—xi-1, 11; Anal 11, 8
or 9; scales 8-66 to 72-23.

Depth 3.5-3.66; head 3.6-3.75; eye 5-5.33 in the head, interorbital
4.33-4.66 in the same. Maxillary extending a little beyond the margin
of the eye; preorbital finely serrate ; vertical limb of preopercle finely ser-
rate, the angle armed with 2 or 3 spines and with 4—7 antrorse spines on
the lower limb; a strong spine at the opercle; 3-5 spines at the clavicle
above the base of pectoral; supraclavicle fine serrate; third dorsal spine
longest, measuring # of the length of the head; second anal spine longest,
2 of the length of the head. Olivaceous.

This species is known only from the types. It is the most northern of
the eastern species.

14. PERCILIA Girard.

Ferciha Girard, Proc. Acad. Nat. Sci. Phila... VII, 1854, 197, and U.S
Naval and Astron. Exped. 235 (g7d/ssz).

Type: Percilia gillisst Girard.

This genus does not belong to the tropical American fauna. Nor has
it been taken in Patagonian waters. Its affinities are Patagonian. Its
southernmost recorded habitat is the Rio Itata directly opposite the source
of the Neuquen, the northern source of the Rio Negro. I have not seen
the species and have drawn on Boulenger’s account of it.

29. PERCILIA GILLISsI Girard.
(Plate XXXVI, Fig. 1).

Percitia gillisst Girard, Proc. Acad. Nat. Sci. Phila., VII, 197, 1854 (Rio
de Maypo); Girard, U. S. Nav. and Astron. Exped. 236, pl. xxix,
figs. 5-9, 1856 (Rio de Maypo); Giinther, Cat. Fish. Brit. Mus. I,
255, 1859; Boulenger, Mem. Soc. Sc. Chile, IV, 15, 1894; Boulenger,
Cat. Fish. Brit. Mus. I, 121, 1895 (Itata); Delfin, Catalogo Peces
Chile 59, 1901 (Maipu and Itata; Pline).

Perca segethi Philippi, Wiegm., Arch. 1863, 211 (Santiago, Chili).

Percilia gracilis Philippi, Mb. Akad. Wiss. Berl. 1866, 708 (Rio Reine,

EIGENMANN: FRESH WATER FISHES. 291

Santiago, Chili); Jordan & Eigenmann, Bull. U. S. Nat. Mus. VIII,
430, 1890.
Habitat: Chili from Itata to Valparaiso.

D. vui-1x, 1, 10-11. A. 1, 8-10. Scales 4 or 5-36 to 40-9 to 11;
lat. line 1, 33-35.

Depth of body 5% to 4 times in total length, length of head 3% to
3% times. Snout slightly overlapping lower jaw, as long as the diameter
of the eye, which is 3% to 4 times in length of head, and a little ex-
ceeds interorbital width; maxillary reaching to below anterior border of
eye, the width of its distal extremity about % diameter of eye; head
naked above, cheeks and opercles scaly. Gill-rakers very short, 10 to 12
on lower part of anterior arch. Dorsal originating behind vertical of
axilla, spinous portion once and a half to once and two thirds as long as
second; third or fourth dorsal spine longest, # to % length of head, and
longer than longest soft rays. Pectoral % to 34 length of head. Sec-
ond anal spine longest. Caudal truncate. Brownish above, the scales
often edged with black ; yellowish beneath ; young with eight or nine more
or less distinct dark cross-bars and the caudal barred with blackish ; lips,
border of gill-membranes, pectoral and border of dorsal fins red.

This species reaches a length of 95 mm.

LIST OF NOMINAL SPECIES WITH THEIR IDENTIFICATIONS.

Perca trucha Cuv. & Val., 1833, Percichthys trucha.
Atherina regia Humboldt, 1835, Atherinopsis regius.
Atherina laticlava Cuv. & Val., 1835, Menidia laticlava.
Arius papillosus Cuv. & Val., 1840, Diplomyste papillosus.
Perca levis Jenyns, 1842, Perichthys trucha.
Atherina microlepidota Jenyns, 1842, Atherinopsis regius.
Mesites maculatus Jenyns, 1842, Galaatas maculatus.
Mesites attenuatus Jenyns, 1842, Galaxias maculatus.
Mesites alpinus Jenyns, 1842, Galaaias alpinus.
Aplochiton zebra Jenyns, 1842, Aplochiton zebra.
Aplochiton teniatus Jenyns, 1842, Aplochiton teniatus.
Petromyzon mordax Richardson, 1845, Caragola mordax.
Trichomycterus arcolatus Cuv. & Val., 1846, Hatcheria areolata.
Trichomycterus maculatus Cuv. & Val., 1846, Flatcheria maculata.
Trichomycterus tnermis Guichenot, 1848, Nematogenys tnermts.
Farionella gayu Cuv. & Val., xxii, 1849, Aplochiton zebra.
Geotria australis Gray, 1851, Geotria australis,

Caragola lapicida Gray, 1851, Caragola lapicida.

292 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Velasia chilensis Gray, 1851, Geotria chilensis.
Percichthys chilensis Girard, 1854, Percichthys trucha.
Percichthys melanops Girard, 1854, Percichthys melanops.
Percilia gillissi Girard, 1854, Percilia gillisst.
Chetrodon pisciculus Girard, 1854, Cheirodon pisciculus,
Galaxias minutus Philippi, 1858, ? Galaxias maculatus.
Galaxias punctatus Philippi, 1858, ? Galaxias maculatus.
Arius carcharias Leybold, 1850, Diplomyste papillosus.
Perca pocha Philippi, 1863, Percichthys melanops.
Perca segethi Philippi, 1863, Perctha gillisst.
Petromyzon anwandtert Philippi, 1864, Caragola mordax.
Petromyzon acutidens Philippi, 1864, Caragola acutidens,
Mesites gracillimus Canestrini, 1864, Galaxias maculatus.
Percilia gracilis Philippi, 1866, Percilia gillisst.

Arius villosus Philippi, 1866, Diplomyste papillosus.
Arius squalus Philippi, 1866, Diplomyste papillosus.
Arius micropterus Philippi, 1866, Diplomyste papillosus.
Arius synodon Philippi, 1866, Diplomyste papillosus.
Petromyzon macrostomus Burmeister, 1868, Exomegas macrostomus.
Galaxias coppingeri Ginther, 1881, ? Galaaias maculatus.
Percichthys vinciguerre Perugia, 1891, Percichthys vinciguerre.
Galaxias delfini Philippi, 1895, ? Galaxias alpinus.
Galaxias grandis Philippi, 1895, ? Galaxias alpinus.
Chirostoma mauleanumt Steindachner, 1896, Memadia mauleana.
Macrophthalmia chilensis Plate, 1897, Geotria chilensis.
Galaxias platei Steindachner, IgQOl, Galaxias platet.

Geotria macrostoma gallegensis Smitt, IgOl, Exomegas gallegensis.
Gymnocharacinus bergu Steindachner, 1903, Gymnocharacinus ber git.
Percichthys altipinnis Regan, 1905, Percichthys altipinnts.

Galaxias titcombi Evermann & Kendall, 1906, ? Galaxias plate,

THE ARCHIPLAT A-ARCHHELENIS THEORY.

I. STATEMENT OF THE THEORY.

In volume IV, page 319, e¢ seg., of these Reports, Ortmann discusses
von Thering’s Archiplata-Archhelenis theory from the standpoint of marine
invertebrate fossils. Inasmuch as the evidence for this theory is mainly
derived from the fresh-water fauna, I propose to discuss the application
of the distribution of the fresh-water fishes of tropical America to this
theory. Before entering into the details I shall outline the theory in von
Ihering’s own words,' culled from his resumé in Science (new series, XII,
857-864, 1900.) :

“The study of the fresh-water fauna, and especially of the Unionidz
of South America, gave me as a practical result the separation of the two
sub-regions ‘Archiplata’ and ‘Archamazonia.’ The first contains Chili,
Argentina, Uruguay and Southern Brazil, the second central and Northern
Brazil (Archibrazil) and Guyana, Venezuela, etc. (Archiguyana). Archi-
plata contains numerous genera of Mollusca, Crustacea, etc., that are com-

1 Ortmann has misread von Ihering in putting the line separating Archiplata and Archhelenis in
the Amazon valley. Ortmann says :

“ Archiamazonas or Archhelenis were separated completely from Archiplata by a broad stretch
of sea, which extended across the present continent, where is now the valley of the Amazon River,
the Cordilleras not being yet formed, and thus a broad communication existed between what is
now the Atlantic and Pacific Oceans.” * * *

Ortmann has been led into this mistake no doubt by the probable northward extension of
Archiplata along the present line of the Cordilleras to central Peru, where Archiplata most nearly
approached, not Archhelenis, but the Archipelago of the northern Cordilleras more nearly asso-
ciated with Archhelenis.

‘Communication was possible in a certain dégree between the shores of Archiplata and Archi-
amazonas. This fact is most plainly seen in the presence of Navidad fossils in corresponding
deposits of northern Peru; the Navidad beds were apparently deposited near the northwestern
extremity of Archiplata, and the Payta Tumbez beds near the southwestern extremity of Archi-
amazonas: at these places we have probably these two continents at their points of nearest ap-
proach to each other, and an exchange of marine forms may have been possible. The exchange,
however, was rendered more difficult by climatic conditions, and especially, although possible in
a certain degree between the Ecuadorian and the Chilian province, it was hard for the Caribbean
* and Ecuadorian fauna to migrate further south, into the Patagonian region.” Ortmann, I. c.
Ortmann includes the greater part of Archhelenis, z. ¢., Archibrazil, in Archiplata.

293

294 PATAGONIAN EXPEDITIONS: ZOOLOGY.

mon to Chili and the La Plata district, such as Uuzo0, Chilina, Parastacus,
glea, etc., including many species and even their parasites ( Zemoce-
phala), which are identical on both sides of the Andes. This contrasts
sharply with the Archamazonian fauna, as tropical genera extend to Rio
Plata and Rio Negro, which are completely wanting in Chili and Peru.
In Ecuador, however, the Cordilleras form no such zodgeographical division,
due certainly to differences in the geological history of both parts of the
Andes. * * * These facts point out that the invasion of the Archama-
zonian element into Archiplata is quite a recent one. The intrusion of the
Archamazonian element is Pliocene or post-Tertiary, and the Andes
formed a barrier insurmountable to fresh-water crabs and mussels as well
as to fishes, chelonians and alligators.

“Tt is evident that the two faunal elements of South America corre-
spond to geographical districts which were separated by the ocean during
the greater part of the Tertiary. The intermixture of the two elements,
and especially the intrusion of Bolivian ants, land snails, etc., in Eastern
Brazil is by no means finished, but is a fact which we observe to-day. * * *

“In relation to the ancient connection of Africa and Archamazonia I
have given arguments (1890) in favor of a Mesozoic ‘archiatlantic conti-
nent,’ which existed during the earlier Tertiary.

‘At first because of some paleontological facts noted by Schlosser, I
believed that this continent could have transmitted Eocene mammals from
South Africa to Europe, an idea now defended by Ameghino and Osborn ;
. but in 1893 I modified my opinion and set forth the hypothesis that no
Eocene placental mammals had existed either in Archamazonia or in
Aéthiopian Africa. The ancient continent uniting Archamazonia with
Africa I named Archiatlantica in 1890, using in 1892 the term Helenis,
and in 1893 that of Archhelenis, with the purpose of preventing confusion
with the ‘Atlantis’ a hypothetical land bridge between South Europe and
Central America proposed by Unger.

“The intimate relations between the fresh-water faunas of Africa and
Brazil, and the colossal difference which exists between the fresh-water
faunas of Archamazonia and Archiplata, prove that both territories during
the greater part of the Tertiary were separated quite as completely as the
two Americas. In this case the mammalian fauna of Patagonia may have
reached Ecuador or Columbia by means of the upheaval of the Andes,
but not Brazil, and both Brazil and the A2thiopian region may have been

EIGENMANN: FRESH WATER FISHES. 295

without mammals and especially placental mammals, during the Eocene.
When, toward the close of the Eocene this land-bridge was submerged,
there already existed many types that have been conserved until our
time. * * * Archamazonia, after the loss of the connection with Africa,
consisted of Archiguyana and Archibrazil. * * * While the distribution
of the existing types of mammals is a result of changes in geography
during Tertiary time, the most fundamental facts in the distribution of the
fresh-water fauna dates from the Mesozoic epoch. The fresh-water fauna
of Chili preserved such a remnant of the Cretaceous fauna almost intact,
and even the connection between the two Americas has not at all modi-
fied the South American fresh-water fauna. * * * There is a further dif-
ference in the distribution of mammals and fresh-water mussels. The
former migrate on land-bridges in most directions, the fresh-water fauna
generally in only one, due to the opportunity given by the currents.
Thus although there was an invasion of Cyprinid fishes into Africa there
was no corresponding emigration of A:thiopian types. * * *

“The fresh-water fauna is not only older but also much more conserva-
tive than the distribution of mammals. One of the most striking examples
of this is given by the history of Africa. While the characteristic mam-
mals are Neogene immigrants, and Lydekker proceeds quite correctly in
making Africa an annex only of the Holarctic region, thus establishing
his Arctogzea, with relation to the fresh-water fauna, Africa is a part of
South America, somewhat modified by the Neogene invasion of Cyprinid
fishes. If as regards mammals Africa belongs to Arctogzea, with relation
to the fresh-water fauna it belongs to the Archhelenic region.”

It is seen from the above that von Ihering divides South America into
two distinct faunal areas, that he attributes the division to an ancient sep-
aration of the continent into Archiplata and Archamazonia, that he places
the separating sea north of Uruguay and that he divides Archhelenis into
Archiguyana and Archibrazil by the eastern Amazon valley. Derby has
objected to the Uruguayan sea because there is no geological evidence
for it. The fresh-water fishes demand a more southern line on the east
than northern Uruguay, and, considering the fresh-water fishes only, there
is no hesitation in placing the line in the La Plata~Paraguay valley and
allowing it to extend to the north-west corner of Archiplata, wherever that
may have been, probably in central Peru.

296 PATAGONIAN EXPEDITIONS: ZOOLOGY.

II. ICHTHYOGEOGRAPHICAL REGIONS.

America, south of the Tropic of Cancer, contains four (or five) distinct
faunas. These faunas are the Transition, the Mexican, the South Amer-
ican and the Patagonian. Only the last two are intimately concerned
with the Archiplata-Archhelenis theory.

The northernmost, or Transition Fauna, is characterized larch by intru-
sive elements from both the north and the south. It is found on the At-
lantic slope from the tropic to the Isthmus of Tehuantepec and on the
Pacific slope of this isthmus. It is of comparatively recent origin and
not an independent fauna but a mixture of the advance guards of the
North American and South American faunas.

The Mexican fauna is unique, and occupies a narrow strip including the
valley of the city of Mexico and the Lerma basin, draining to the west,
and the Rio San Juan, a tributary of the Panuco, draining to the east.
While containing intrusive elements from the north, it contains none from
the south, and its fauna is so distinct from either that there is slight hesi-
tation in considering it as equivalent to the North American, South Amer-
ican and Patagonian faunas. It is a very old fauna, but has at no time
influenced the South American fauna.

The third, the South American Fauna, is sharply divisible into the
Brazilian and Andean. The Brazilian occupies the rivers from Southern
Mexico to Buenos Aires and from the Atlantic to the Pacific, exclusive of
the high Andes and Chili. This fauna is the richest in species in the
world. From this region about 10 per cent. of all the known fishes have
been recorded. The Andean, from 3,000 to 5,000 feet and over above
sea level, while possessing some forms in common with the Brazilian, is
quite distinct. The species inhabiting this region, while derived from the
Brazilian fauna, have become so far modified that they would cause
surprise if found at Manaos; those of the lower portions of the Pacific
slope would not.

The Patagonian fauna, in distinct contrast to the South American, is
one of the poorest in the world. It occupies the Rio Negro basin and
-everything lying south of it and a line joining it with Valparaiso. Its
fauna is distinct from the Brazilian and has been considered in detail,
pp. 227-292.

For convenience these faunal areas may be enumerated as the following
“regions” of unequal value.

EIGENMANN: FRESH WATER FISHES. 297

Transition.
Mexican.
Brazilian.
Andean.!

Patagonian.

ee ean

Ill. THE TRANSITION REGION.

While there are absolutely no North American intrusive elements in
South America, a number of Mississippi Valley forms have reached as far
as the Isthmus of Tehuantepec; only a Lepzsosteus, Carpiodes meridio-
nalis and possibly Aplodinotus grunniens extend farther.

On the Atlantic slope the North American fauna gradually gives place
to or shades into the tropical fauna between the Rio Grande on the north
and the Isthmus of Tehuantepec on the south. The Rio Grande fauna is
distinctly northern, that of the Papaloapam on the south is as distinctly
tropical, although the former contains a few southern elements and the
latter a few northern ones. On the western slope, the southern element
makes its appearance farther south than on the eastern and is separated
from the northern realm by the Mexican fauna. The Mexican and Transi-
tion faunas have recently been made the subject of profound and exhaus-
tive study by Meek and I can not do better than to summarize his results.

In this discussion the Petromyzonide, Bagrine, Catostomide and
Cyprinide, being unquestionably derivatives from the north, are credited
to the north, although some of the genera are peculiar to the Mexican
or Transition regions. For similar reasons members of the Crchkide,
Characida, Pimelodine et al. are considered South American, although
some of the genera are peculiar to or dominant in Mexico and Central
America. Agonostomus and the Pacilide, which are dominant in Mexico,
Central America and the West Indies, are considered Transitional, although
they extend both into North America and South America.

In the Rio Grande basin on the boundary of the United States there are
known to occur sixty-one species of fishes. Of this number three species,
Astyanax mexicanus, Cichlasoma pavonaceum and Herichthys cyano-
guttatum, are distinctly tropical, belonging as they do to Central or South
American genera; forty-four are distinctly northern, and of the remainder,
two are tropical Gobies also found on the American Gulf coast, and eleven

' The Titicacan basin probably should constitute a “ region ” distinct from the Andean north of
Titicaca.

298 PATAGONIAN EXPEDITIONS: ZOOLOGY.

are Paciliide, some of them distinctly northern and some distinctly south-
ern, but none of them causing any surprise.

Farther south, near the tropic on the Atlantic side, is the basin of the
Rio Panuco. Nearly half of its fauna, 15 species out of a total of 32,
belongs to the northern fauna. Five species are distinctly tropical. In
the following list of species from this river the northern forms are marked
with an N, the southern with an S and the Mexican with an M, the transi-
tion forms witha T. Marine derivatives are marked Me. The Rio Panuco
ison the same degree of latitude north of the equator that the Parahyba
is south. The Parahyba harbors over 50 species, all of them equatorial ;
of the 32 found in the Panuco less than 16 per cent. are equatorial.

Tue Panuco System (after Meek).

N Lefisosteus osseus (Linnzeus).

N Lepisosteus tristechus (Bloch & Schneider).
N Sctalurus furcatus (Le Sueur).

N Lctalurus punctatus (Rafinesque).

N Ameiurus australis Meek.

N Ameturus mexicanus Meek.

N Carpiodes tumidus Baird & Girard.

N Carfpiodes labiosus Meek.

N Aleansea tincella (Cuvier & Valenciennes).
N Hybognathus rasconts (Jordan & Snyder).
N Aztecula mexicana Meek.

N Wotropis forlonensis Meek.

S Astyanax mexicanus (Filippi).

N Dorosoma exile Jordan & Gilbert.

N Stgnalosa mexicana (Giinther).

N Fundulus heterochtus (Linnzus).

T Cyprinodon eximius Girard.

T Gambusia affints (Baird & Girard).

M Goodea towert Meek.

M Goodea atripinnis Jordan.

M Platypecilus variatus Meek.

T Pecilia latipunctata Meek.

T Pacilia sphenops Cuvier & Valenciennes.
T Mohenesta latipinna Le Sueur.

EIGENMANN: FRESH WATER FISHES. 299

T Xiphophorus montezume Jordan & Snyder.
Me Pomadasys templet Meek.

S Cichlasoma steindachnert Jordan & Snyder.
S Cichlasoma bartonti (Bean).

S Herichthys cyanoguttatum (Baird & Girard).
S Neotroplus carpintts Jordan & Snyder.'

T Philypnus dormitor (Lacépéde).

T Chonophorus tatasica (Lichtenstein).

This number of species was taken from near tide level to an elevation
of almost 2000 feet.

On the Pacific slope the Rio Presidio, a short, insignificant stream, at
Mazatlan is near the tropic. North of it no South American elements
are found.

Its fishes are:

T Pecilia butlert Jordan.

T Peciha presidionis Jordan & Culver.
Me Szphostoma starkst Jordan & Culver.
Me Z7hyrina crystallina Jordan & Culver.
T Agonostomus monticola Bancrott.

S Cichlasoma beant (Jordan).

T Philypnus dormitor (Lacépéde).

T Dormitator maculata (Bloch).

T Lleotris pictus Kner & Steindachner.
T Chonophorus tarasicus (Lichtenstein).
Me Achirus mazatlanus (Steindachner).
Me Achirus fonsecensis (Giinther).

Only Cechlasoma beant is distinctly tropical American. Agonostomits is
Central American; the rest are types that would cause no surprise
anywhere in a coastwise stream between the United States and South
America.

The Rio Presidio, the Rio Mezquital to the south and the Rio Yaqui
to the north, are on the slope of the Sierra Madre opposed to the slope
drained by the Rio Grande. Confining ourselves for the present to the
consideration of replacement of the northern by the southern fauna, we
shall return to a comparison of their fauna with that of the Rio Grande.

' This species is included in the synonymy of the one preceding it in the systematic part of this
report.

300 PATAGONIAN EXPEDITIONS: ZOOLOGY.

The Rio Mezquital south of the Presidio contains :

N Amewurus pricet (Rutter).

N Pantosteus plebeius( Baird & Girard).
*N Moxostoma austvinum Bean.

N Fybognathus episcopus (Girard).

N Leuctscus nigrescens (Girard).

N Notropis ornatus (Girard).

M Characodon garmanit Jordan & Evermann.
*M Characodon furcidens Jordan & Gilbert.
T Cyprinodon latifasciatus Garman.

*M Chirostoma mezquital Meek.

N Ltheostoma pottsw (Girard).

As Meek points out, all but three of these species (marked *) belong
to the Rio Grande basin and were probably captured from it. There are
no distinctly South American elements in this river system as far as ex-
amined; it belongs to the north. But no collections have been made
below the Mexican plateau and the tropical elements would naturally
occur in the lowlands.

Skipping the Rio Grande de Santiago and its tributary the Lerma, whose
basin embraces the greater part of the Mexican plateau, the Rio Balsas
belongs distinctly to the transition zone containing 7etvagonopterus mext-
canus and Cichlasoma tstlarius as (S) South American intrusives ; /s¢/a-
vius, Astecula and Notropis as (N) North American intrusives, the first
two of which have become generically distinct from their North American
relatives and might therefore be classed as transition genera. The rest
are transition genera (T) with the one exception of the Mexican Goodea.

This river is almost equal in size to the Lerma. It contains only about
one-fourth as many fishes. Its basin is larger than the basin of the Panuco.
It was fished from about 1400 feet to 6000 feet in elevation. In this
area the number of species is less than half the number found in the
Panuco.

N /stlarius balsanus Jordan & Snyder.
N Aztecula vittata (Girard).

N WNotrofpis boucard? (Giinther).

S Astyanax mexicanus (Philippi).

T Gambusia gracilis (Heckel).

M Goodea whitet Meek.

EIGENMANN: FRESH WATER FISHES. 301

T Platypecilus nelsoni Meek.

T Pecilia sphenops Cuvier & Valenciennes.
T Melaniris balsanus Meek.

T Agonostomus monticola (Bancroft).

S Cichlasoma tstlanum (Jordan & Snyder).
T Chonophorus tarasicus (Lichtenstein).

Near the Isthmus of Tehuantepec, on the Atlantic side, is the basin of
the Papaloapam with the following species. In this river the northern
element has approached the vanishing point, while the southern element
is increasing. Collections were made from tide level to an elevation of
about 3500 feet.

Me Galeichthys aguadulce Meek.

S Rhamdia oaxace Meek.

S Rhamdia brachyptera (Cope).

N Carpiodes meridionals (Giinther).

S Astyanax mexicanus (Filippi).

S Astyanax eneus (Giinther).

S Hemigrammus compressus Meek.

N Dorosoma anale Meek.

N Dorosoma extle Jordan & Gilbert.

N Szgualosa mexicana (Giinther).

T Cynodonichthys tenuis Meek.

T Pseudoxiphophorus bimaculatus (Heckel).
T Gambusia bonita Meek.

T Paragambusia nicaraguensis (Giinther).
T Belonesox belizanus Kner.

T Platypecilus maculatus Ginther.

T Heterandria lutzst Meek.

T Paecilia sphenops Cuvier & Valenciennes.
T Aiphophorus hellert Heckel.

Me Zy/losurus marinus (Walbaum).

Me Menidia tsa Meek.

T Agonostomus monticola (Bancroft).

Me Centropomus mexicanus Bocoutt.

Me Pomadasys starri Meek.

Me Pomadasys templet Meek.

S Cichlasoma salvint (Giinther).

302 PATAGONIAN EXPEDITIONS: ZOOLOGY.

S Cichlasoma hedricki Meek.

S Cichlasoma parma (Giinther).

S Cichlasoma etgenmannit Meek.

S Cichlasoma melanurum (Giinther).
S Cichlasoma nebulifer (Ginther).

S Thorichthys hellert (Steindachner).
S Thorichthys elliot? Meek.

T Philypuus dormitor (Lacépéde).

T Dormitator maculatus (Bloch).

T Chonophorus tatasicus (Lichtenstein).
Me Achirus fasciatus (Lacépéde).

Only four out of the total of thirty-seven species are northern, while
thirteen are distinctly tropical, of which the genera Czch/asoma and Tho-
vichthys are largely middle American.

The list may be contrasted with the list of the combined Geronimo and
Verde, directly across the headwaters of the Papaloapam and emptying
into the Pacific. They were fished near tide water and at an elevation of
nearly 5000 feet.

S Astyanax eneus (Giinther).

T Fundulus oxace Meek.

T Gambusta fasciata Meek.

T Anableps dovit Gill.

T Heterandria lutzi Meek.

T Paecilia sphenops Cuvier & Valenciennes.
Me Mug cephalus Linnzus.

S Cichlasoma evermanni Meek.

T Philypnus dormttor (Lacépéde.).

Of these, but two are South American, the rest are transitional.

The tropical element has reached farther north on the Atlantic side than
on the Pacific, or in any cross section of the country there are more South
American elements in the former than in the latter. None of the South
American types have reached the plateau of Mexico. They skirt along
the coastal plain on both slopes to an elevation of perhaps 2000 feet and
in the Rio Grande basin to an elevation of even 3000 feet.

EIGENMANN: FRESH WATER FISHES. 303

IV. THE MExIcan REGION.

North of the Rio Balsas, and therefore north of the transition zone
is the Rio Grande de Santiago. It is the largest river of the Pacific
slope within the tropics or between the Rio Colorado and Tierra del
Fuego. It is also by far the richest in species, the Rio Balsas com-
ing next in size and but little further south containing but 12 species as
compared with the 48 of the Rio Grande de Santiago. Many of the
species of this region (C/zvostoma) are indirect derivatives from the sea.
The only fish in this, the largest Pacific slope river, that suggests South
America, is Czchlasoma beant, and this occurs only near the coast. Its
contact with North America is shown by such genera as Lampetra, Ametu-
rus, Moxostoma, Algansea, Notropis and Fybopsts.

This basin above the falls separating the coastal section from the Mexi-
can plateau, together with the Valley of Mexico and the headwaters of the
San Juan, a tributary of the Panuco, which rises in the Mexican plateau,
harbors a fauna that is distinct from both the fauna of North America
and that of South America or even southern Mexico. Of this Meek says:
“From the area which includes the valley of Mexico, the headwaters of the
San Juan del Rio and the Lerma basin, there are at present fifty-four spe-
cies of fishes known (inclusive of coastal plain species), only two of which,
Moxostoma austrinum Bean and A ztecula vittata (Girard) |to these should
be added Crchlasoma bean? | have been taken in any other river basin.
These fifty-four species belong to twenty-one genera, eight of which are
peculiar to this region.” Three of the other genera are confined to the
neighborhood and have probably emigrated from it, six are northern
genera, of which only one extends farther south than this area. ‘The
Lerma river system is far from being thoroughly explored, but apparently
its fishes are quite as distinct and characteristic as if the fauna were insular.”

The relation of each species in the following list to the northern (N),
southern (S), or Marine (Me), faunas, whether peculiarly Mexican (M) or
belonging to the coastal plain (C), is indicated by the respective symbols.
N Lampetra spadicea Bean.

N Ameturus dugest Bean.

N Moxostoma austrinum Bean.

N Xystrosus popoche Jordan & Snyder.

N Algansea tincella (Cuvier & Valenciennes).
N Algansea dugest Bean.

304 PATAGONIAN EXPEDITIONS: ZOOLOGY.

N Algansea rubescens Meek.

N Algansea lacustris Steindachner.

N Falcula chapale Jordan & Snyder.
N A2tecula lerme Jordan & Snyder.

N Notropis cahentis Jordan & Snyder.
N Hybopsts altus (Jordan).

M Zoogoneticus cuitzecensis (B. A. Bean).
M Zoogoneticus dugest (Bean).

M Zoogoneticus vobustus (Bean).

M Zoogoneticus maculatus Regan.

M Zoogoneticus diazt Meek.

M Characodon multiradiatus Meek.

M Characodon eisent Rutter.

M Characodon variatus Bean.

M Characodon laterals Giinther.

M Chapatichthys encaustus (Jordan & Snyder).
T Gambusia infans \Voolman. _

M Goodea luitpoldi (Steindachner).

M Goodea atripinnis Jordan.

M Skiffia multipunctata (Pellegrin) «

M Skiffia lerme Meek.

M Skiffia variegata Meek.

M Skeffia btlineata (Bean).

T Pecilia occidentalis (Baird & Girard).
M Chirostoma jordant \Voolman.

M Chirostoma arge (Jordan & Snyder).
M Chivostoma barton Jordan & Evermann.
M Chirostoma attenuatum Meek.

M Chirostoma labarce Meek.

M Chirostoma patzcuaro Meek.

M Chirostoma zirahuen Meek.

M Chirostoma chapale Jordan & Snyder.
M Chirostoma grandocule Steindachner.
M Chevostoma promelas Jordan & Snyder.
M Chirostoma sphyrena Boulenger.

M Chirostoma luctus Jordan & Snyder.

M Chirostoma humboldtianum (Cuvier & Valenciennes).

itn ee

EIGENMANN: FRESH WATER FISHES. 305

M Churostoma lerme Boulenger.

M Chirostoma ocotlane Jordan & Snyder.
M Chivostoma estoy Jordan.

CT Agonostomus monticola (Bancroft).
CS Cichlasoma beant (Jordan).

It is evidentfrom a glance at the letters showing the relationship of
these species that the Mexican fauna has not been qiguereed by the South
American fauna. Nor has it influenced in any way the South American
fauna.

V. THE ANDEAN REGION.

The Andean Region includes the high Andes on both slopes from
Venezuela and Colombia to Chili.

It is poor in species at any given point, but some of the genera have a
large number of local adaptations or species. This region is distinctly
marked off into three provinces.

1. The Northern includes the high lands of northern Peru, Ecuador,
Colombia and Venezuela. This is the richest in species and distinguished
by the genera 4rges, Cyclopium, Prenadilla and the high development of
Chetostomus. ts fauna is largely an ancient derivative from the low-
land fresh-water fauna of Archiguiana,

2. The Titicacan, including the basin of Titicaca and neighboring
streams, and possibly the land-locked basins of Bolivia, concerning which
nothing is known, is distinguished by the genus Ovestzas and the absence
of the genera distinguishing the Northern province. Its fauna is largely
an ancient derivative from the ocean.

3. The Southern is the poorest in species, characterized by the absence
of everything but a few species of Pygzdiam, a genus which extends the
entire length of the Andean Region.

Concerning the physical features of the Titicacan province Agassiz
says (Bull. Mus. Comp. Zool. III, No. 11, 1875):

“Tn fact, the geography of the whole of the west coast of the Andes to
the north of Chili seems to point to a former condition of things such as
we now find on the west coast of Chili. The plains to the southward of
Santiago, bounded by the coast range to the westward, and the Andes to
the east, gradually pass to the condition of the coast now prevailing at
Conception Bay, and south of it — the coast range forming the archipel-
ago, the Andes forming the coast range, and the plains of the more

306 PATAGONIAN EXPEDITIONS: ZOOLOGY.

northern regions becoming changed to bays; the immense basins suc-
ceeding each other towards the north which form the so-called Desert of
Atacama, the nitrate-beds, the llanos of the coast, the pampas of Peru,
through which the rivers flowing to the west have cut deep valleys with
more or less marked terraces, showing the different periods of ascent in
the elevation of the continent. These plains are everywhere found either
between a coast range and the base of the eastern talus of the Andes, or
extending from the summit of the shore terrace, if we may so call it, gen-
erally at a height of from 1200 to 3000 feet, sloping to the second ter-
race, with its base at an average height of from 6000-7000 feet, and
then followed by a second and third more or less indistinct terrace until
we reach the main elevated plateau or basin which lies between the eastern
and western slope of the Andes. All these basins show more or less dis-
tinctly the trace of their former marine origin, so that, if we are to judge
from the presence of strictly marine forms, the successive terraces devel-
oped on a magnificent scale on the west coast of the Andes, with the
interlying basins, we have a fair presumption that the elevation of the
Andes to their present height has taken place at a comparatively recent
date, and during their upheaval the present nitrate district and saline
deposits were left as large lagoons during a considerable period, to judge
from the great thickness of the deposits found within their basins, all
denoting the presence of a comparatively quiet inland sea.

“Lake Titicaca itself must have, within a comparatively very recent
geological period, formed quite an inland sea. The terraces of its former
shores are everywhere most distinctly to be traced, showing that its water-
level must have had an elevation of 300 or 400 feet at least higher than
its present level. This alone would send its shores far to the north in
the direction of Pucara, forming a narrow arm reaching up to S. Rosa.
Lake Arapa is probably only an outlier of the ancient lake, as well as
several of the small lakes, now at a considerable distance from the west
shore. The immense plain of Cabanillas, extending north beyond Lampa
to Juliaca, only 100 or 120 feet above the lake at its highest point, was
one sheet of water. The terraces of the former shores are still very dis-
tinctly seen. The eastern shores did not probably differ greatly from the
present outline, though the peninsula of Achacache was probably an
island. The bay of Puno must have been connected with the plains of
Llave, and those back of Juli; while from the lower lake, back of Aygache,

EIGENMANN: FRESH WATER FISHES. 307

the lake formed huge inlets or deep bays, now represented only by the
nearly dry river-beds flowing into the lake at Aygache, Corilla and Guajui.
The sluggish Desaguadero must have been a strait of considerable width,
with large islands; and this long lake, connecting Lake Titicaca with
Lake Aullagas, must have equalled in extent the upper lake, the upper
lake, at that time, extending across the Isthmus of Yunguyo, leaving the
Peninsula of Copacabana, as a large island, connected with the lower lake
by a broad pass between the hills to the west of Copacabana, and those
to the west of Yunguyo. The plains, now laid bare at the northern
and western shores of Lake Titicaca, give us an excellent idea of the
appearance the whole basin of the lake would present if entirely dry.
The number of lakes and basins, great and small, which formerly covered
the elevated plateau of the Andes, must have been very great; but we
now find only here and there a small sheet of water. The former lakes
are only represented by the more or less extensive pampas, forming basins
at great altitudes, showing plainly that the whole of this district is receiv-
ing a much smaller waterfall than in former times, but probably not within
historic times, if we take into consideration the position of some of the
most ancient ruins of Bolivia (at Tiahuanaco), which are only about 75
feet above the present level of the lake.’’!

The area to which Lake Titicaca belongs probably remained separate
from Ecuador until long after the latter was colonized from Archiguiana,
and the Brazilian fauna did not gain access to the Titicacan region until
after the elevation of the lake had become too great for most of the Bra-
zilian types of fishes.

I give below a list of the species of the Northern and Titicacan prov-
inces with the localities from which they are recorded. The Northern
province is characterized by several genera, some of them undoubted
derivatives of the neighboring lowlands marked ‘‘/’’ and by others on the
western slope derived from the sea marked “Me.” Rfamadiaand Pygidium
have a wide distribution. The Titicacan province is characterized by the
extravagant genus Ovestias, which has developed from a Fundulus-like
species received when the lake was still an arm of the sea.’

'See also page 372.

* The crustacean fauna is even more evidently of marine origin. The genus Orchestia, with
its seven Titicacan species, is still preéminently a marine genus.

303 PATAGONIAN EXPEDITIONS: ZOOLOGY.

1. Vorthern.

| Rhamadia pentland: Cuvier & Valenciennes : Monterico, Tullumayo, Rio
Huambo.

Pygidium vivulatum (Cuvier & Valenciennes) : Ucayale.

Pygidium paeyanum (Cope): Arequiba, Peru.

Pygidium meride (Regan): Merida and Rio Albirregas, near Merida,
Venezuela.

Pygidium tenia (Kner): western slopes of Peru.

Pygidium laticeps (Kner): western slopes of Peru.

Pygidium knerid (Steindachner): Canelos and Rio Zamora, Peru.

Pygidium ergenmannt (Boulenger) : Cumbaca.

Pygidium dispar Tschudi: eastern slope of Peru at an altitude of

14,000 feet.

Pygidium nigromaculatum (Boulenger): Colombia.

Pygidium taczanowski (Steindachner) : Rio de Huambo; Rio de Tortora.

Pygidium quechuorum (Steindachner) : Arequipa, Peru.

| Evemophilus mutise¢ Humboldt: Bogota.

| Chetostomus —all (14) species of this genus.

| Arges—all (15) species of this genus (to an elevation ofat least 10,700 feet).

/ Cyclopium — all (4) species of this genus.

| Astroblepus grixalvit Humboldt: Popayan, Rio Cauca.

| Creagrutus —all species.

Me Protistius semotilus Cope:' Peruvian Andes at an elevation of 12,-
000 feet).

Me Gastropterus archeus Cope:1 Arequiba, at an elevation of 7,500 feet.

Me Oreogobius rosenbergti Boulenger: Paramba, northwest Ecuador, 3,-
500 feet.

2. Titicacan.

Pygidium vivulatum (Cuvier & Valenciennes).
| Ancistrus montana (Regan): eastern Andes of Bolivia.
| Ancistrus triformts (Cuvier & Valenciennes): eastern Andes of Peru
and Bolivia.
Oresttas —all (11) species of this genus.
From the Southern Province there are known only a very few species of
Pygidium.
_ | These two genera occur in the latitude of Titicaca, south of the other species. I am not
sure that they do not belong to the Titicacan province.

EIGENMANN: FRESH WATER FISHES. 309

VI. THE BRAZILIAN REGION.

The Brazilian region is divisible into several conspicuously marked
provinces; the Central American, including the area from Veragua to the
Isthmus of Tehuantepec; the Pacific, including the Pacific slope from
Panama to Peru; the Amazonian, including all the lowlands drained by
the Amazons; the southeastern, including the streams from Cape San
Roque to Santos and the Trinidad Province. Less marked provinces are
the La Platan, the Guianan and the Magdalenan, the latter including the
Chagres. All of these provinces are distinguished from the Amazonian
chiefly by negative characters. The number of peculiar genera varies
greatly in the different provinces.

The question of the character of the different provinces is largely a
question of what Amazonian types are absent. We shall take up these
provinces in detail emphasizing the Central American, Pacific and Mag-
dalenan, as these are of chief interest in discussing later the origin of
the Pacific slope fauna, and the East Brazilian and Guianian, because these
are the ones principally concerned in the Archiplata-Archhelenis theory.

Manaos is the centre of the great Brazilian region and the variety of
the fauna of any province, other things being equal, varies inversely with
its distance from Manaos. The Magdalena and the San Francisco at
opposite sides of Manaos have essentially the same types of fishes.

1. Zhe Central American Province.

This province extends from the Isthmus of Darien to the Isthmus of
Tehuantepec.

The most comprehensive account of the Central American province
has been given by Giinther (An Account of the Fishes of the States of
Central America, Trans. Zool. Soc. VI, 1866). Recently Miller has made a
detailed study of the Motagua basin. This province will amply repay
further study.!. Giinther has shown that this province is distinguished
from the South American by the absence of most South American fami-
lies. A few Pimelodine of the Nematognathi and a few forms of the
Tetragonopterine and of the Characinz are all there is of the one thou-
sand species of the Ostariophysiz found in the Brazilian region. The
genera Cichlasoma and Aiguidens are dominant in all the streams.

‘The proof of this had been read before Regan’s Pisces of the Biol. Cent. Amer, appeared.

310 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Giinther has shown that the short and separated water courses are for the
most part inhabited by local adaptations or species peculiar to each water
system, or two or three of the neighboring systems. For instance, all but
two or three of the inhabitants of Lake Peten are peculiar to this small lake.

For this reason he has been able to divide this province into the follow-
ing subprovinces :

A. Fresh waters north of the Lakes of Managua and Nicaragua, empty-
ing into the Pacific. From this subdivision he recorded 22 species, nearly
all of them confined to this region. Only 4 of these species extend north
as far as the Isthmus of Tehuantepec. Eight are South American, one,
Lepisosteus, is northern, nine are transitional or middle American and the
rest are Marine.

B. The fresh waters north of the Lakes of Managua and Nicaragua,
emptying into the Atlantic. From this subregion he recorded 28 species
nearly all of which are peculiar. Eighteen are South American in their
relationships. Only four species, Pemelodus salvint, Belonesox belizanus,
Xiphophorus hellert and Ameturus meridionalis (from the Usumacinta) are
found north of the province; only the last is a northern intrusive. The
Rio Motagua contains no North American elements. This subprovince
also contains Gymnotus carapo and Symbranchus marmoratus, the South
American species having the farthest range northward.

C. Lake Peten with 14 species, of which all but two were supposed to
be peculiar.

D. Lake Managua with 6 species.

E. Lake Nicaragua with 9 species.

F. The fresh waters south of the Lakes of Managua and Nicaragua to
the Isthmus of Darien. (The region about Panama is unmistakably South
American, as distinct from Central American.)

2. Pacific Province.

The Pacific province includes all the fresh waters emptying into the
Pacific from Panama to the southern boundary of Peru, except the head-
waters above an elevation of about 3000 feet. It is a narrow strip, nowhere
exceeding 200 kilometers in width, and in places less than 50 kilo-
meters wide. North of 5° of south latitude it is abundantly supplied with
rivers. South of this point the rivers become smaller and of less impor-
tance. The rivers are the Mamoni and Chepo in Panama, Rio San Juan

EIGENMANN: FRESH WATER FISHES. 31!

and Rio Patia Mira in Colombia, the Esmeralda, Daule, Guayas and Tumbey
in Ecuador, and the Chia, Jequetepec, Santa, Rimac and others in Peru.
None of those in Colombia have been examined, and but few of those
in Peru.

The headwaters of these rivers are all near the headwaters of At-
lantic slope streams. In northern Ecuador and in Peru they are separated
by a high ridge. Of special interest are the San Juan and Patia. In
northern Peru there are four main chains of the Cordilleras, besides the
coast range. The valleys between these are occupied by large rivers, the
Ucayale, Huallaga and Marafion, all ultimately draining into the Amazon.
In Ecuador there are two very high chains with frequent cross-ridges divid-
ing the longitudinal trough into parks or highland basins, some drained
to the east and some to the west.

In Colombia there are again four chains. The eastern one divides to-
wards the north, one branch sweeping around to the east toward Caracas
and Trinidad, another branch extending to the coast west of Lake Maracaibo.
Between this and the central chain flows the Magdalena which occupies
much the same position in regard to the Cordilleras that the Ucayale or
Huallaga occupies, but empties into the Caribbean Sea. Between the
central and western Cordilleras flows the Cauca, a large tributary of the
Magdalena. South of the Cauca the Rio Patia has cut through the west-
ern Cordilleras, its headwaters coming from the same trough in which
the Cauca flows.

Of still higher interest is the trough between the western Cordillera
and the coast or border Cordillera.

Sievers, p. 478, states that the west Cordillera of Ecuador is younger
than the central Cordillera, and the coast range still younger. Its height
varies from 800-1800 m. A longitudinal valley separates the coast
range from the western Cordillera. In this flows the Atrato towards the
north, emptying into the gulf of Darien and the San Juan towards the
south, emptying into the Pacific Ocean. The water-shed separating the
two scarcely reaches a height of 100 m. above sea-level, the distance of
the height of land above their headwaters is naturally much less than this."
I shall return to these rivers again.

The fauna of this province, so far as known, is very poor. In con-

‘A gentleman, for several years a missionary in Colombia, has informed me that the Indians of
the Atrato have a legendary beliet in a former interoceanic connection through the Atrato valley.

312 PATAGONIAN EXPEDITIONS: ZOOLOGY.

sidering it two things may be borne in mind: 1, that the province, while
very long, is very narrow and contains no large river system, which is so
necessary to develop or harbor a diverse fauna. The sources of the
Paute, a tributary of the Marafion are only 60 km. from the bay of Guay-
aquil. In North America the large Colorado, Sacramento or Columbia
basins of the Pacific slope contain fewer species of fishes than the un-
known Beanblossom creek of the Mississippi basin: 2, that no river has
been thoroughly explored and some of the principal ones, the San Juan
and Patia for instance, not at all.
With very few exceptions, the species of this province belong to genera
with a very wide distribution; on the Atlantic slope, all but one of these
genera reach as far north as the Magdalena or Panama. Few of the species
would have occasioned surprise if they had been discovered at Manaos.
There are remarkably few peculiar genera, which indicates a compara-
tively recent separation from the Atlantic slope fauna. The species
recorded from this province are listed below and marked ‘“*” for species
peculiar, ‘“**’’ genus peculiar, ‘‘/’’ species found also in the Chagres or
Magdalena, ‘‘Me”’ marine. Most of the species are recorded from Ecua-
dor, few are found in Peru, and Colombia has not been explored.
| Rhamdia cinerascens (Giinther): Rio Daule; Rio Peripa; Guayaquil and
Esmeraldas ; Chagres.

| Rhamdia wagnert (Giinther): Atlantic and Pacific slope of Panama;
Mamoni near Chepo.

| Pimelodus clarias (Bloch): Rio Bayano—eastern slope to Buenos Aires.

| Pimelodella modesta (Giinther): Eastern Panama; Esmeraldas.

* Pimelodella elongata (Giinther).

* Pimelodella grisea Regan: Durango; Sapayo and Vaqueria Rivers, in
northwest Ecuador.

* Pimelodella yuncensts Steindachner: Pacasmayo.

**® Cefopsogiton occidentalis (Steindachner) : Guayaquil.

Species of Pygidium, Arges, Cyclopium and Chetostomus considered

under the Andean Region may descend into this province.

* Pygidium punctulatum (Cuvier & Valenciennes) : Callao.

* Pyoidium pardus (Cope): Jequitepeque; Callao Bay.

* Plecostomus spinosissimus Steindachner: Guayaquil; Rio Vinces; Rio
Daule.

* Plecostomus feste Boulenger: Rio Vinces; Rio Peripa.

EIGENMANN: FRESH WATER FISHES. 313

* Femiancistrus annectens (Regan): St. Javier and Rio Durango, west
Ecuador.

* FHlemiancistrus aspidolepis (Giinther): Pacific slope of Panama.

* Chetostomus fischeri Steindachner: Rio Mamoni near Chepo.

* Chetostomus dermorhynchus Boulenger: Rio Zamora; Rio Santiago ;
Rio Peripa; Rio Vinces.

* Chetostomus microps Giinther: Western slope of Ecuador.

* Toricaria jubata Boulenger: St. Javier 60 feet; Rio Durango 350 feet.

* Loricaria variegata Steindachner: Mamoni.

| Loricaria uracantha Kner & Steindachner: both slopes of Panama.

| Sturisoma panamensis (Eigenmann & Eigenmann): Rio Magdalena;
Pacific slope of Panama.

* Sturisoma frenata Boulenger: St. Javier 60 feet; Salidero 350 feet;
Rio Durango 350 feet.

| Hophas mucrolepis (Giinther): Rio Chagres; Rio Daule; Rio Vinces ;
Guayaquil.

** Tebrasina bimaculata Cuvier & Valenciennes: Callao; Rio Vinas;
Lima) Remac- Eten:

* Curimatus troschelit Giinther: Guayaquil; Rio Vinces.

| Curimatus magdalene Steindachner : Magdalena; Mamoni.

* Curimatus boulengert Eigenmann: Rio Vinces.

* Prochilodus humeratis Giinther: Rio Peripa, western Ecuador.

* Prochilodus caudofasciatus Starks: Perene.

*#* Saccodon wagnert Kner & Steindachner : Ecuador.

** Saccodon craniocephalum Thominot : Rio Guayaquil.

* Leporinus leschenaulti Cuvier & Valenciennes : Guayaquil.

* Prabucina astrigata Regan.

** Pseudochalceus lineatus Kner: western slope of the Andes of Ecuador.

* Astyanax feste (Boulenger): Rio Vinces and Mirador, Ecuador.

| Astyanax rutilus (Jenyns) : western Ecuador: Rio de la Plata to Mexico.

* Astyanax stmus (Boulenger): western Ecuador.

Astyanax peruanus (Miller & Troschel): Rio Remac; Pacasmayo ;

Lake Amatitlan, Rio Eten and Payta, Peru.

* Hlemibrycon polyodon (Giinther) : Guayaquil.

* Brycon alburnus Giinther: Rio Peripa and Rio Vinces, western Ecuador.

| Brycon dentex Giinther: Esmeraldas and Rio Peripa to Lake Nica-
ragua and Rio Usumacinta.

314 PATAGONIAN EXPEDITIONS: ZOOLOGY.

| Brycon atrocaudatus Kner: Magdalena; Paramba, western Ecuador,
Payta and Eten, Peru.
>? Creagrutus peruanus Steindachner: Monterico and Huambo, Peru.
| Creagrutus miilleri Giinther: Cauca; Canelos; Andes of western
Ecuador.
* Gasteropelecus maculatus Steindachner : Mamoni.
| Luctocharax insculptus Steindachner : Mamoni; Magdalena.
| Eigenmannia humboldti (Steindachner) : Mamoni ; Magdalena.
| Gymnotus equilabiatus Humboldt: Guayaquil ; Magdalena.
| Symbranchus marmoratus Bloch: Panama ; entire eastern slope.
Me Stolephorus poevi (Kner & Steindachner) : Rio Bayano.
Me Pristigaster effulgens Regan : Rio Vaqueria.
Fundulus guatemalensis Giinther : western Ecuador to Amatitlan and
Guacalate.
* Aplochetlus peruanus Regan: Perim.
* Pocilia feste Boulenger: Hot Springs, S. Vicenta, Santa Elena, Ecuador.
Me Zylosurus fluviatilis Regan: Western Ecuador.
* Me Centropomus atridorsalis Regan: Rio Vagueria, near la Tola, N. W.
Ecuador.
* diguidens sapayensis Regan : Rio Sapayo, Ecuador.
* iguidens rivulata (Giinther): Rio Sapayo; Rio Peripa; Rio Vinces.
| Cichlasoma godmanit Giinther: west slope of Panama; Rio Cahabon,
Atlantic slope.
* Cichlasoma feste Boulenger: Rio Durango, Ecuador.
Me * Awaous giinthert Regan: western Ecuador.
Me * Awaous transandeanus Giinther : western Ecuador.
Me * Gobordes peruanus Steindachner.
Me * Chasmodes maculipinna Regan: Rio Durango.
Number of species, 65; number peculiar, 42; on Atlantic slope, 18
of which in Amazon, 3.
These are for the most part lowland species belonging to lowland genera,
7. e., not such as are found in mountain streams or that are likely to have
reached the Pacific slope by crossing the high Andes. Pygedium and
Chetostomus are exceptions to this. If all of the purely freshwater species
could be transferred to Manaos or Tabatinga there are very few that would
occasion surprise if taken in the collector’s net.

EIGENMANN: FRESH WATER FISHES. 315

3. Magdalena Province.

The fauna of the Rio Magdalena, as far as known, is of peculiar interest,
in connection with the fauna of the Pacific Province. The Madalena basin
is Y-shaped. The right (thick) arm and stem form the Rio Magdalena
proper, draining the area between the eastern and central Andes of south-
ern Colombia. The left arm of the Y is the Cauca, which drains the
area between the central and the western Andes of Colombia. The
central Andes appear to stop at the junction of the arms of the Y.

The Magdalena has an elevation of 1600 m. at San Augustin, 1000
m. at Timana and 400 m. at Neiva. There are rapids at the 5th degree
of north latitude about Pescaderias. In its lower course it is a much-
branched and divided stream. Its length is about 1350 km. and it drains
300,000 square kilometers. The Cauca at Popayan has an elevation
of 1740 m. It is marked with rapids at Quilichao 1070 m., and again,
below Cartago, goo m., and between Antioquia and Caceres.

Nearly all that is known of the fauna of this river has been contributed
by Steindachner in three papers, in which the following fresh-water species
are enumerated (exclusive of mountain forms).

Species *, and genus **, peculiar to the Magdalena; species /, and
genus //, peculiar to Magdalena, Panama and western slopes of Ecuador.
A, also found in the Amazon.

* Potamotrygon magdalene Steindachner.

A Pseudopimelodus zungaro (Humboldt).

A Rhamdia sebe (Cuvier & Valenciennes).

A Pimelodus clartas (Bloch).

* Pimelodus grosskopfit Steindachner.

A Pseudoplatystoma fasciatum (Linnzus).

A Sorubim ftima (Bloch & Schneider).

* Doras longipinnis Steindachner.

Trachycorystes tnsignis (Steindachner).
Trachycorystes magdalene (Steindachner).

* Agenetosus caucanus Steindachner.

A Agenetosus dentatus Kner.

? Pygtdium, probably several species in upper reaches of the rivers.
A Floplosternum thoracatum Cuvier & Valenciennes.
A Plecostomus emarginatus Cuvier & Valenciennes.
* Panaque cochliodon (Steindachner).

316 PATAGONIAN EXPEDITIONS: ZOOLOGY.

* Pterygoplichthys undecimal’s (Steindachner).

* Loricaria filamentosa Steindachner.

* Loricaria magdalene Steindachner.

| Sturisoma panamense (Eigenmann & Eigenmann).
A Hoplias matabaricus (Bloch).

| Curimatus magdalene Steindachner.

* Curvimatus mivartit Steindachner.

A Prochilodus rubrotentatus Schomburgk.

* Prochilodus magdalene Steindachner.

* Prochilodus longirostris Steindachner.

A Leporinus eques Steindachner.

* Leporinus muyscarus Steindachner.

A Leforinus obtusidens (Valenciennes).

A Leforinus striatus Kner.

A Leforellus vittatus (Cuvier & Valenciennes).

* Chetrodon insignis Steindachner.

A Astyanax bimaculatus (Linnzus).

| Astyanax rutilus (Jenyns).

* Astyanax caucanus (Steindachner).

* Brycon rubricauda Steindachner.

| Brycon atricaudatus (Kner).

* Othonophanes labiatus (Steindachner).

| Creagrutus miilleri Giinther. (Cauca and Ecuador).
* Chalcinus magdalene Steindachner.

* Poeboides dayit Steindachner.

Cynopotamus magdalene Steindachner. (Magdalena and Paraguay).
Salminus affinis Steindachner. (Magdalena and Ecuador).
#8 cestrocephalus anomalus (Steindachner).

** Grlbertolus alatus (Steindachner).

|| Luctocharax tnsculptus Steindachner.

A Hypopomus brevirostris Steindachner.

?| Eigenmannia humboldti (Steindachner).

| Gymnotus equilabiatus Humboldt.

A Synbranchus marmoratus Bloch.

* Rivulus elegans Steindachner.

* Givardinus caucanus Steindachner.

E-quidens pulchrum (Gill). (Rio Chagres to Trinidad.)

EIGENMANN: FRESH WATER FISHES. 317

* Cichlasoma krausstt Steindachner.
* Geophagus steindachnert Eigenmann & Hildebrandt.
| Geophagus crasstlabris Steindachner.

Undoubtedly this is but a small part of the fauna of this basin. As our
knowledge stands at present, this basin lacks very largely those Amazonian
genera that are also absent from the coastal plain of southeastern Brazil.
See later List. Thirty per cent. of its known species are Amazonian.

4. The Amazon Province.

East of the Cordilleras, and therefore east of the Magdalena basin, is
found the most extensive and intricate fresh water system in the world.
It forms a network of rivers practically uninterrupted, extending from the
mouth of the Orinoco through the Cassiquiare, Rio Branco, Rio Negro,
Rio Madeira, Rio Guaporé, Rio Paraguay, Parana and La Plata to Buenos
Aires, through 45° of latitude and from Para to within a few miles of the
Pacific coast, throtigh 30° of longitude at the equator.

The connection between the Orinoco and Branco through the Cassi-
quiare has been definitely known since Humboldt traversed it. The fol-
lowing is from the International Bureau of American Republics, ‘‘ United
States of Brazil,” p. 79:

“Another remarkable phenomenon of the Paraguay is the mingling of
its principal head waters with those of the affluents of the Amazon. An
affluent of the Jauru River is sufficiently near the Guaporé River to be
connected with the latter by a canal. The Aguapehy, another tributary
of this river, is separated from the Alegre by a narrow isthmus 5 kilo-
meters wide. In the eighteenth century an attempt was made to open up
a canal here, and owing to the abundant rains a large canoe of 12 oars
succeeded in passing from the one river to the other. One of the gover-
nors of the state also endeavored to open up a canal 10 kilometers long
in another part of the isthmus, but on account of the small amount of
trade it was never completed. This would connect Montevideo and Para
by a continental waterway 8300 kilometers long. In the near future it is
probable that railways will take the place of the canal. There are many
places on the edge of the plateau, farther to the east, where a simple cut
of afew meters would connect the tributaries of the Amazon with those
of the Paraguay, transforming eastern Brazil into an island. There is a

318 PATAGONIAN EXPEDITIONS: ZOOLOGY.

space of but 100 meters between the Estivado, a small tributary of the
Tapajoz, and the Tombador, which empties into the Cuyaba.”

Cuyaba on the headwaters of the Paraguay is not more than 200 m. in
elevation and the junction of the Cassiquiare with the Orinoco has an
elevation of but 280 m. One could therefore traverse the 8300 kilome-
ters by inland water without ever exceeding an elevation of 280 m.

The Amazon itself, where it leaves the Andes, has an elevation of but
180 m.; indeed nearly 60 per cent. of all South America consists of flat
lowlands, drained by this system of rivers.

The Amazon draining more than 7,000,000 square kilometers, the La
Plata about 3,100,000 and the Orinoco about 1,000,000, these rivers
which as far as the fishes are concerned form asingle system, drain an
area 3% times as large as the Mississippi basin.

The lowlands through which these main rivers flow are the youngest
part of South America. The parts that first arose out of the sea and
became populated with fresh-water fishes were probably two land areas.
The one embraces the highlands of Guiana and Northern Brazil, the
other the highlands of Brazil east of the Araguay and south of the falls
of the Tapajos. In Tertiary times the Cordilleras arose out of the ocean
on the west. The basins left as open seas between the three land areas
later, in late Tertiary, became dry land, in part by elevation and in part
by delta formation by the rivers coming from the surrounding land
masses. The rivers coming from the Andes became stocked from the
surrounding ocean and /afey from one of the older eastern land areas.
The main part of the later colonization very probably did not take place
until the Andes had become an effective barrier to the distribution of fresh-
water fishes.

These interior rivers, chiefly the Amazons, colonized from the north-
east and southeast became themselves the seat of unparalleled adaptive
radiation and centers of distribution, as we shall see.

The Amazons are unequalled in richness and form a province distinct
from the Orinoco and Guiana region to the north and the La Plata on
the south. The La Plata basin is well distinguished from the Amazonian,
but by negative characters only. !

‘T insisted on this distinction in 1891. Perugia registered his ‘“ disgosto”’ with this scheme,
because the fishes from the La Plata basin were nearly all Amazonian, as I had pointed
out.

EIGENMANN: FRESH WATER FISHES. 319

Neither the Orinoco nor the Paraguay contains any fishes that would
cause the slightest surprise if they were found at Manaos or any other
part of the Amazon. Every collection made in the Paraguay in recent
years has brought to light more types hitherto considered Amazonian and
has increased the number of genera and species that the two systems
have in common. Very little is yet known of the Orinoco.

The list of Amazonian species is too long to give here. The Amazon-
ian species have therefore been marked A in the general list of the species
forming Part III of this report. They number about 700.

5. The Guiana Province.

The Guiana Province, including one of the oldest land masses, is drained
by the Cachipur, Oyapoc, Cayenne, Mana, Maroni, Surinam, Corentine,
Essequibo (Mazaruni, Cuyuni, Rupununi and other tributaries of the Esse-
quibo), Demarara, Berbice, by the Orinoco and its eastern and southern
tributaries, the Caroni, Caura, Ventuari and by the Rio Branco, and
the northern tributaries of the Amazon east of the Branco.

The feature distinguishing this region is the known or reported connec-
tions between neighboring streams. The Cassiquiari connects the Orinoco
with the Rio Negro. The Atabapo is said during the rainy season to be
connected with the Guaina, and the Rupununi of the Essequibo basin
with the Tacutu of the Rio Branco basin. It is said that the Essequibo
is also connected with the Rio Trombetas through the Apini and the
Oyapoc, Cachipur and Araguary with the tributaries of the Yari which
empties into the Amazon.

The lower course of the Essequibo of British Guiana at least is con-
nected with the lower Orinoco by natural canals, so that these form part
of the Orinoco-Amazon-La Plata meshwork and contain the same types.
The Eastern Guianas (French) have a less varied fauna.

The fauna of this region is of the greatest importance to theoretical chorol-
ogy, since this is one of the two old land-masses, and since it was by a
continuation of this area that South America is supposed to have been
connected with Africa.

Our knowledge of the fish fauna of this region is derived from Miiller
& Troschel’s account of the fishes of British Guiana; Bleeker’s Silure de
Surinam; Vaillant’s notes on the fishes of French Guiana, and his account
of the Berbice and the general work of Cuvier & Valenciennes, Giinther,
Eigenmann & Eigenmann, Regan and Pellegrin.

320 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Unfortunately little is known of the fauna of the streams of the table
land of Guiana, the region where the remnant of the original fauna may
be expected to persist. Schomburgk’s collections made in these streams
were largely lost. He states that it was rich in species (over 30).

André (‘A Naturalist in the Guianas,”’ p. 205) says: “ In fact the falls of
Para [of the Caura] appear to constitute an effective barrier between dis-
tinct forms of river life,’ and that the fauna above the falls is different
from that below.

The most promising field for scientific results, if not number of species,
seems to me to be offered by the rivers of this region, which should be
explored adove and de/ow falls that are impassable barriers for the ascent
of fishes.

The rivers of this region, exclusive of the northern tributaries of the
Amazon, concerning which not much is known, contain a total of about
298 recorded species. Of these about 60 per cent. are also found in the
Amazon; as of these about 16 species are from the Rio Branco basin and
not from the other streams and the Rio Branco belongs to the Amazon
system, this number must be reduced by 16, which still leaves over 50
per cent. of the fauna identical with that of the Amazon.

LisT OF GUIANA FISHES WITH THEIR KNOWN DISTRIBUTION.

In the following list the letter A indicates that the species has been re-
corded from the Amazon. L indicates that the species has been recorded
from the La Plata, S that the genus is also found in the San Francisco,
* that the species is peculiar to the province, ** that the genus is peculiar
to the province.

EIGENMANN:

FRESH WATER FISHES.

a2

Takutu.
Lake Amucu.
Orinoco.
Barima.
Waini.
Pomeroon.
Essequibo.
Cuyuni,
Mazaruni.
Rupununi
Rewa
Cuyuwini.
Demerara.
Berbice.
Corentine.

Surinam,

Maroni.

Mana.
Cayenne and
Mahuey.
Carsevenne.
Lunier
Carnot.
Cachipour.

* Paratrygon strongylopterus (Schomburgk).
AL Potamotrygon hystrix (Miiller & Henle).
* Potamotrygon a’ orbignyt Castelnau.
A Potamotrygon reticulatus Giinther.
L Potamotrygon motoro (Miiller & Henle).
A Ellipesurus spinicauda Schomburgk.
* Bunocephalichthys hypsiurus (Kner).
A Bunocephalus gronovit Bleeker.
A Platystacus cotylephorus Bloch.
* P. nematophorus Bleeker.
A P. aspredo (Linneus).
* P. sicuephorus (Cuv. & Val.).
* P, filamentosus (Cuv. & Val.).
A P. tibicen (Temmink).
A Callophysus macropterus (Lichtenstein).
A Pinirampus pirinampu (Spix).°
“AL Luctopimelodus pati (Val.).
SA Pseudopimelodus raninus (Cuv. & Val.).
AL Ps. sungaru (Humboldt).
S* Rhamdia laukidi Bleeker. *
A Rh. schomburgkti Bleeker.*
A Rh. sebe (Cuv. & Val.).°
* Rh. foina Miller & Troschel.
A Rh. multiradiatus (Kner.).
S* Rhamdella notata (Schomburgk).
AL Heptapterus mustelinus (Val.).
* Acentronichthys surinamensts (Bleeker).
SAL Pimelodus ornatus Kner.
AL Pimelodus clartas (Bloch).
A Pimelodus altipinnis (Steind. ).
A Geldiella eques (Miiller & Troschel)."
SA Pimelodella cristatus (Miiller & Troschel).*
A P. wesselii (Steind.).
AL P. gracilis (Valenciennes).
A Platynematichthys punctulatus (Kner).
A Phractocephalus hemiliopterus (Bloch & Schnei. ):°
A Brachyplatystoma vaillanti (Cuy. & Val.).
AL Hemisorubim platyrhynchos (Cuv. & Val.).
SA Pseudoplatystoma fasctatum (Linneus).
A Ps. tigrinum (Cuv. & Val.).”
AL Sorubim lima (Bloch & Schneider).
A Sorubimichthys planiceps (Agassiz).
SA Doras dorsalis Cuv. & Val.
* D. albomaculatus Peters.
* D. helicophilus Giinther.
* D. dentatus Kner.
AL D. costatus (Linnzeus).”
AL D. armatulus Cuv. & Val.
A D. cataphractus (Linnzus).
A D. affinis Kner.
* D. castaneo-ventris Schomb.”

X X | Rio Branco.

x XX

x XX

SA Oxydoras niger (Val.).™

x
x X
Xx X

x X| xX

upper

x X
x
ui xX -v

xX KK XXX
ux

upper <

Pee eral naliemey aiesaliate

x

“XK XK XK

x

x KX

XK PRK OK

' French Guiana. ? British Guiana.
® British Guiana. TTn all rivers of Guiana.
"British Guiana. 2 Rio Pasawiri.

iT Guiana. * Guiana.
8 Rio Mahu. ° All rivers of Guiana.
13 All rivers of British Guiana.

> Guiana.
1 Rivers of Guiana.

322 PATAGONIAN EXPEDITIONS:

ZOOLOGY.

Rio Branco.
Takutu

Lake Amucu.

Orinoco.

Barima.

Waini.

Pomeroon.

Essequibo.
Cuyuni

Mazaruni.

Rupununi
Rewa

Oxydoras carinatus (Linn.).
Trachelyopterus coriaceus Cuy. & Val.
Centromochlus oncina (Schomb.).*
C. heckelit Philippi.
C. perugie Steindachner.
C. aulopygius Kner.
Trachycorystes glaber (Steind. ).
T. obscurus (Giinther).
T. ceratophysus (Kner.).
T. galeatus (Linnzus).
TZ. robustus (Giinther).
SA Pseudauchenipterus nodosus (Bloch).?
A Auchenipterus nuchalis (Spix).
* Agenetosus tnermis (Linnzus).
* A. armatus Lacépéde.
A A. dentatus Kner.
* A. porphyreus Cope.
A A. dawalla (Schomburgk).
AL A. brevifilis Cuv. & Val.
* A. axillaris Giinther.
4k Flelogenes marmoratus Giinther.
A HHypophthalmus edentatus Spix.
A Cetopsis cecutiens (Lichtenstein).
S Pygidium tenia (Kner.).
A Vandellia plazaii Castelnau.

AL Callichthys callichthys (Linneus).
AL foplosternum littorale (Hancock).
A Hf. thoracatum (Cuv. & Val.).°

A Corydoras punctatus (Bloch).
SAL Plecostomus plecostomus (Linneus).*
A P. verres (Cuv. & Val.).
A P. emarginatus (Cuv. & Val.).
* Hemiancistrus medians (Kner).
AL Cochliodon cochliodon (Kner).
* Hemiancistrus schomburghkii (Gimther) .°
* HT, megacephalus (Giinther).
Prerygoplichthys etentaculatum (Spix).
SAL P. multiradiatus (Hancock).
L P. barbatus (Cuv. & Val.).®
* P. depressus (Giinther).
* P. guenthert (Regan)."
L Xenocara gymnorhynchus (Kner).
A Ancistrus dolichoptera (Kner).
* A. temminckit (Cuv. & Val.).
AL A. hoplogenys Giinther.
AL A. cirrhosus (Val.).
* Pseudacanthicus serratus (Cuv. & Val.).
* Ps. fordit Giinther.
A Acanthicus hystrix Spix.
Loricaria filamentosa Steind.
SA LZ. maculata Bloch.
AL Z. typus (Bleeker).

be
cos 9 eee

x X

x X X

xXxXxXX xX

x

x

x

x X

Berbice

| Cuyuwini.
Demerara.

Corentine.

Surinam.

Maroni.

Mana.

Cayenne and
Mahuey.

Carsevenne.

Lunier.

Carnot.

4

Cachipour,

x X

x X

x
x

x

x XK K XK x KKKKKKKXKX

Xx X x X

x X

x

x X

x X

xX

xX X

1 Rio Padawiri.

Muller & Troschel. 5 British Guiana.

* British, French and Dutch Guiana.

3 Curassarrasa.
° British Guiana.

* Probably as HZ. commersont

7 British Guiana.

EIGENMANN : FRESH WATER FISHES. 323

Rio Branco.
Takutu
Lake Amucu.
Orinoco.
Barima.
Waini.
| Pomeroon.
Essequibo.
Cuyuni
Mazaruni.
Cuyuwini.
Demerara.
Berbice
Corentine.
Surinam.
Maroni.
Mana.
Cayenne and
Mahuey.
Carsevenne. |
Lunier
Carnot
Cachipour.

x
x
x

AL Loricaria cataphracta Linneus.
* Z. brunnea Hancock.
* L. platyura Miiller & Troschel. x
* Harttia platystoma (Giinther). Xx
A Neoplecostomus granosus (Cuv. & Val.).
SAL Hoplias malabaricus (Bloch).' SEVEN EW 2< 9) Cy eo eles
SAL Hoplerythrinus uniteniatus (Spix). x
A Erythrinus erythrinus (Bloch & Schneider).
* Erythrinus longipinnis Giinther. x
* Pyrrhulina filamentosa Cuv. & Val. x x
* Elopomorphus orinocensts Steind. :
AL Psectrogaster ciliatus (Miiller & Troschel).
SAL Curimatella alburnus (Miiller & Troschel).
SA Curimatus spilurus Giwmther. x x
* C. microcephalus E. & E.
A C, nerii Steindachner. x
C. cyprinotdes (Linnzus).
C. schomburgkit Giinther. x
Gs
e

x

x

x

xX XK X
x X
xX X

xX X

essequibensis Giinther.
Semitapicis latior Spix.
Prochilodus rubroteniatus Schomb.
P. insignis Schomb.
P. binotatus Schomb.
* P. laticeps Steind.
A P. brama (Cuv. & Val.).
A Cenotropus labyrinthicus (Kner). x
* Chilodus punctatus (Miiller & Trosch.).’ x
SAL Hemiodus unimaculatus (Bloch). Xx
A HZ. semiteniatus Kner.
A HZ. immaculatus Kner. x
* Anisitsia keppleri (Giinther).
A A. notatus (Schomburgk).
* Nannostomus beckfordit Giinther. x
SA Anostomus anostomus (Linnzus). x
* Anostomus orinocensis (Steind. ).
AL Schizodon fasciatus (Spix.). X| xX
SAL Characidium fasciatum Reinhardt.
AL Leporinus striatus Kner.
SA Z. nigroteniatus (Schomb.). x
L. miillert Steind.
L. megalepis Giinther.
L. frederict (Bloch). x
L. affinis Giinther.
Li,
we

xX -*%XXX
x
x

xX X

x X X

x X

MIS PROS 0K OOK OK

. Aypselonotus Giinther.
. fasciatus (Bloch).
L. margaritaceus Giinther. : x
A L. maculatus Miill. & Trosch.*
A Piabucina uniteniatus Giinther.
KE Scissor macrocephalus Giinther. x
SA Hemigrammus unilineatus Gill. x
* HT, riddlet Meek. i x
* H. micropterus Meek. x

‘All rivers of Guiana. *Savanna swamps of British Guiana, *British Guiana. *Awaricuru, ° British Guiana.

324

PATAGONIAN

EXPEDITIONS:

ZOOLOGY.

| Rio Branco,

Takutu.
Lake Amucu.

A Hemigrammus belotti (Steindachner ).
L Ctenobrycon hauxwellianus (Cope).
AL Tetragonopterus argenteus Cuvier.
AL TZ. chalceus Agassiz.
* Poptella longipinnis (Popta).
A Fowlerina orbicularis (Cuv. & Val.).
* Astyanax polylepis (Giinther).
S* 4. spilurus (Cuv. & Val.).
AL A. abramis (Jenyns).
AL A. bimaculatus (Linnzus).
* A. wappi (Cuv. & Val.).
AL A. fasciatus (Jenyns).*
AL Cvreatochanes melanurus (Bloch).
* C. affinis Giinther.
* Moenkhausia grandisquamis(Miiller & Troschel)
AL MW lepidurus (Kner).
* M. oligolepis (Giinther).
* M. chrysargyreus Giinther.
* M. ovalis Giinther.
*K Chalceus macrolepidotus Cuvier.
S* Brycon longiceps Steind.
* B. schomburgkti Miiller & Troschel.
* B. pesu Miiller & Troschet.
* B. falcatus Miiller & ‘Troschel.
* B. luctdus (Kner).
LA Thoracocharax stellatus (Kner).
A Gasteropelecus sternicla Linneus.
SAL Chalcinus angulatus Spix.
A C. giinthert Garman.
A C. elongatus Giinther.
*k Agoniates halecinus Miiller & Troschel.
A Piabucus dentatus (Keelreuter).
A Crenuchus spilurus Giinther.
A Pygopristis denticulatus (Cuvier).
S* Pygocentrus niger (Schomburgk).®
AL P. piraya (Cuvier).
A P. scapularis (Giinther).
AL P. natterert (Kner).
AL Serrasalmo gymnogenys Giinther.
SAL Serrasalmo marginatus Val.
AL SS. spilopleura Kner.
* S. rhombeus (Linneus).
A Catoprion mento (Cuvier).
A Acnodon oligocanthus (Miiller & Troschel).
A Piaractus macropomus (Cuvier).
A Metynnis hypsauchen (Miiller & Troschel).”
AL Jf, maculatus (Kner).
SA AGjleus setiger (Miiller & Troschel).
* M. schomburgkii (Jardine ).”
A M. torquatus (Kner).
* MM. ellipticus (Giinther).
* M. knerii (Steindachner).

~v

Orinoco.
Barima.

YESS]

x X

x Xv xX

Waini.

| Pomeroon.

| Essequibo,

Cuyuni.
| Mazaruni.
| Rupununi

Rewa.
| Demerara.

Berbice.
| Corentine.

Surinam.

Maroni.

Mana.
Cayenne and

Mahuey.
Carsevenne.

Lunier.

Carnot.
| Cachipour.

| Cuyuwini.

KISKE XK
x

x
XOX aw

~X a

xxx X
~xX

x
x

xX X

x X
x X

xT x

x

t Dutch Guiana. ? British Guiana.
5 British Guiana. ® British Guiana.
9 British Guiana. ” Lake Tapacumea.

° British Guiana.
7 Savanna swamps.
1 Below the cascades.

* Ditches and swamps near coast.
8In upper courses of streams of British Guiana.
2 Zuraima and Savanna swamps.

EIGENMANN :

FRESH WATER FISHES.

325

A Myleus discoideus (Kner).

A M. rhomboidalis (Cuv.).*

* M. rubripinnis (Miiller & Troschel).
AL MW. asterias (Miiller & Troschel).
AL Mylossoma aureus (Spix).

AL Characinus gtbbosus (Gronow).
SA Reboides affinis (Giinther).
R. microlepis (Reinhard).

A Cynopotamus humeralis Val.

A Exodon paradoxus Miller & Troschel.
AL Salminus hilarit Cuv. & Val.

SA Acestrorhynchus falcirostris (Cuv.).

A A. microlepis (Schomburgk ).?

AL A. falcatus (Bloch).
A Aydrocynus cuviert (Agassiz) .
A #. ocellatus (Schomburgk).
AL Rhaphiodon vulpinus Spix.
AL Aydrolycus scomberotdes (Cuv.).
SAL Sternarchus albifrons Linneus.

* Sternarchogiton sachst (Peters).

* Sternarchorhynchus oxyrhynchus (M. & T.).

A Rhamphichthys rostratus (Linneus).
AL R. marmoratus Castelnau.

* Hypopomus artedi Kaup.*

S* Higenmannia microps (Boulenger).
AL 2&. virescens (Val.).
> SAL Giton fasciatus (Pallas).
SAL Gymnotus carapo Linneus.

A Electrophorus electricus (Linnzus) .°
AL Synbranchus marmoratus Bloch.

A Stolephorus spinifer (Cuv. & Val.).

AS. clipeoides (Swainson).

A Pterengraults atherinoides (Linnzus).®

A Lycengraulis grossidens (Cuvier).®
AL Jlisha flavipinnis (Valenciennes) .°

A Pristigaster cayanus Cuvier.®

A Osteoglossum bictrrosum Agassiz.

A Arapaima gigas Cuvier.

* Rivulus geayi Vaillant.

A R. urophthalmus Giinther.

R. micropus Steind.
R. obscurus Garman.
A Anableps anableps (Linneus) .*
A A. microlepis Garman.
AL Pecilia vivipera Bloch & Schneider.
me * Zlosurus mtcrops (Giinther).
meA 7. almeida Quoy & Gaimard.
AL Potamorrhaphis guianensis Schomburgk.
me * Polycentrus schomburgkii Miiller & Troschel.

A Plagioscion squamosissima (Heckel).°

P. heterolepis (Bleeker).

A P. surinamensis (Bleeker).

X | Rio Branco.

x

BEX

x X

Takutu.

|

x X

| Lake Amucu.

a 8 calea |g | gl ol | cl. S neellan 5
S| s|-2]8| 8 |) 8/8 e¢/-S]/ 8/2/68] 8lal 4/5 oa] 8] 3/4] 6
Sree (isi tole loll) betel hoig
Pe Ne) SS Te | Re x x
x
x x
xX
lowerX xX|X
x
x
Xx
x x
x
Xx
X | upperX
x<X x
x x<
x
x
x x
x x
x
x
s x x
x x
£ x
x xX
xX xX x x
x xX x X|X
SSE) Ee MGS Ste te lits x<
x x x x x
x
x
x
x K
x Rowan
x
x
x
SRN RY ? x
PRB) Oe ? x x
P| Pte} 2, s x
x
x
S| SK
x
x ?
x
x L

' All rivers of British Guiana.
British Guiana. ° Entering rivers.

* British Guiana.
™ Upper courses ?

3 British Guiana. * French Guiana.
8 Mouths of rivers of British Guiana.

5Tn all fresh waters of
* Rivers of Guiana.

Takutu.
Lake Amucu.
Orinoco.
Barima.
Waini.
| Pomeroon.
Essequibo.
Cuyuni
Mazaruni.
Rupununi
Rew2
Cuyuwini.
| Demerara.
Berbice.
Corentine.
Surinam.
Maroni.
Mana
Cayenne and
Mahuey.
Carsevenne.
Lunier.
Carnot.
Cachipour.

SA Pachyurus schomburgkit (Giinther).
A Pachypops furcreus (Lacépéde).
A P. trifilis (Miiller & Troschel).*
A Chetobranchus flavescens Heckel.
A Cichla ocellaris Bloch & Schneider.
A Cichla temensis Humboldt.
A UWaru amphiacanthotdes Heckel.
A Acaropsis nassa (Heckel).
AL Astronotus ocellatus (Agassiz) .
A Aquidens vittata (Heckel).
A 4. tetramerus (Heckel).
* 4. geayl (Pellegrin).
* 4, maronit (Steindachner), x x|xX
42, guianensis (Regan).?
*K MVannacara anomala Regan. x
A Cichlasoma bimaculatum (Linneus). x X| xX
A C. temporale (Giinther).* x x
A C. festivum (Heckel). x
A C. severum(Heckel).*
A C. psittacum (Heckel).
A Crenicara punctulata Giinther.
L Batrachops semifasciatus Heckel. |X
* B. punctulatus Regan.
A B. reticulatus Heckel.
A Crenicichla saxatilis (Linneus).
C. wallacit Regan.
C. multispinosa Pellegrin.
C. vittata Heckel. XK
C. lenticulata Heckel. x |
C. lugubris Heckel. |x x |
C. johanna Heckel.® 1X
Geophagus surinamensis Bloch. x
G. camopiensis Pellegrin.®
G. jurupari Heckel."
G. cupido Heckel. x
A Satanoperca pappaterra (Heckel). |
* Pterophyllum altum Pellegrin. x .
A Colomesus psittacus Bloch & Schneider.

xxXxXxX XX | Rio Branco.

x

2 BE AOS
x
x
xX
x X

xX
x

x X XK XK XK

326 PATAGONIAN EXPEDITIONS: ZOOLOGY.

>
x

>
POX e SP ee

' Along the coast of British Guiana. * Guiana. * Guiana. * Guiana.
° British Guiana. ° French Guiana. 7 Guiana.

EIGENMANN: FRESH WATER FISHES. 327

6. Trinidad Province.

The fauna of Trinidad is of peculiar interest. Trinidad is an island
off the delta of the Orinoco, approximately 40 by 50 miles. Its geograph-
ical affinity is with Venezuela rather than with Guiana. Its fresh-water
fauna consists of three elements:

1. West Indian, 16+per cent. (about). This element consists chiefly of
members of the Gobiidae which enter fresh water throughout tropical
America.

2. Amazonian, 33 per cent. This element consists, with few excep-
tions, of species ranging from the La Plata basin to Trinidad. The ex-
ceptions are found in the Amazon and Trinidad. All are marked A.

3. Autochthonous, 33 per cent. This element consists either of local
adaptations, *, of genera with a very wide distribution, or of distinct
genera marked **.

Me Selenasfis herzbergi (Bloch). Trinidad to Brazil.

Me Zachisurus laticeps Giinther. British Guiana, Trinidad.

A Rhamadia sebe knert Steindachner. Paraguay to Trinidad.

* Pseudauchenipterus guppyt (Regan). Trinidad.

* Trachycorystes pasee (Regan). Trinidad.

A Anctstrus cirrhosus Cuvier & Valenciennes. Rio Grande do Sul to
Trinidad.

* Hlemiancistrus trinttatis (Giinther). Trinidad.

A Plecostomus robin Cuvier & Valenciennes. La Plata to Trinidad.

A Plecostomus plecostomus (Linnzus). Eastern South America.

A Caliichthys calhchthys (Linnzeus). La Plata to Trinidad.

A FHofplosternum tittorale Hancock. La Plata to Trinidad.

A Floplosternum thoracatum (Cuvier & Valenciennes). Eastern South
America.

* Corydoras eneus Gill. Trinidad.

A floplhas malabaricus (Bloch). La Plata to Magdalena.

A Floplerythrinus unittentatus (Spix). La Plata to Trinidad.

* Curimatus argenteus Gill. Trinidad.

* Odontostilbe pulcher (Gill). Trinidad.

A Hemigrammus unilineatus Gill. Trinidad to Bahia.

* Hemibrycon tenturus (Gill). Trinidad.

* Flemibrycon guppyt (Regan). Trinidad.

328 PATAGONIAN EXPEDITIONS: ZOOLOGY.

A Astyanax bimaculatus (Linnzus). La Plata to Panama.
** Stewardia albipinnes Gill. Trinidad.
A Gymnotus carafo Linneus. La Plata to Rio Motagua.
A Symbranchus marmoratis Bloch. La Plata to Cuba.
* Haplochilus hartt (Boulenger). Trinidad.
* Acanthophacelus guppi (Giinther). Trinidad.
W Doryhamphus tineatus Valenciennes.
W Agonostomus monticola (Bancroft). West Indies, Trinidad.
Me Mugil brasiliensis Agassiz.
Me Mugil trichodon Poey.
W Centroponius ensiferus Poey.
W Centropomus undecimalrs Bloch.
Polycentrus schomburgkit Miller & Troschel. Trinidad.
Eiquidens pulchrum (Gill.) Chagres, Trinidad.
A Cichlasoma bimaculatum (Linnzeus). Paraguay to Trinidad.
A Crenicichla saxatilis (Linnzus). Amazon to Trinidad.
W Dormitator maculatus Bloch.
W Philypnus dormitator (Lacépéde).
** Evorthodus breviceps Gill. Trinidad and Surinam.
W Chonophorus banana (Cuvier & Valenciennes).
* Gobius fasciatus (Gill). Trinidad.

7. Southeast or East Brazilian Plateau.

The southern land-mass, Archamazona, the Brazilian plateau, is drained
by the Tocantins, Rios San Francisco, Doce, Jequitinhonha, Parahyba
and other coast streams and the upper tributaries and head-waters of the
Parana. The rivers leave this mass over important falls. The Parana
is part of the La Plata system, and as far as it may be considered, will
be taken up with the La Plata.

This old land-mass was doubtless more extensive in the past. The
isolated ranges or peaks in the territory west of the Araguay are the prob-
able remnants of the formerly more extensive plateau. The falls in the
Tocantins, Xingu, Tapajos and Madeira mark the probable northern and
western boundary of this plateau. However, west of the Rio San Fran-
cisco, the streams all drain into the Amazon and their fauna could not
have retained its original character. The western streams probably con-

EIGENMANN: FRESH WATER FISHES. 329

tributed to the Amazonian fauna when the latter was formed and received
in turn new Amazonian types.

The fauna of eastern Brazil differs much more from that of the Ama-
zon than does that of the equally old Guiana. By eastern Brazil is here
understood the northeastern part of this old land-mass, z.e., the territory
north of the Parana basin and between the eastern watershed of the To-
cantins and the coast south from Cape San Roque to: the Iguape. The
entire territory has an average height of about 2,000 feet. North of the
Parahyba it is in the form of two wide terraces, besides the coastal plain.
The western terrace is occupied by the San Francisco basin; it is sepa-
rated from the eastern narrower terrace by a series of Serras beginning
with the Serra da Tuila, near the mouth of the San Francisco, and end-
ing in the Serra do Espinhaco near Rio de Janeiro. In these Serras and
the table-land or middle. terrace to the east of them are the sources and
middle courses of a number of rivers with a general west to east course.

This middle terrace is bordered on the east by the coast ranges ex-
tending from the Parahyba north. Through it all of the rivers of the
region have broken their path and in it all of them have extensive falls.

These streams with the tributaries and localities where collections have
been made are: (1) Rio Paraguassu, emptying at Bahia; (2) Rio de Con-
tas; (3) Rios Jundiahe, Salsa, Pardo and Jequitinhonha (with its tributary
Arrasuahy) emptying at Cannavierias; (4) Rio Mucury on which are
located Porto Alegre and Santa Clara; (5) Rio de San Matheos; (6) Rio
Doce and its tributary Rio San Antonio; (7) Rio Quenda, emptying at
Santa Cruz; (8) Rio Itabapuana.

Farther south we have: (9) Rio Parahyba with its tributaries Piabanha,
Pirahy, Muriahe, Pampa and Preto (Campos; Juiz de Fora; Mendez;
Taubaté); (10) Macahe; (11) Rio de Janeiro; (12) Santos and Alto da
Serra; (13) Iguape.

The Parahyba, Santos and Iguape are opposite to the Parana basin, all the
others to the San Francisco. The trend of the Parahyba is different from
that of the coastwise streams to the north of it. It flows first southwest,
then northeast parallel with the coast range and between it and the Serra
de Mantegueira. Von Ihering has suggested that the peculiar trend of
the upper course of the Parahyba and some of its tributaries indicates that
these were formerly tributaries of the Rio Tieté and were captured from it
by the Parahyba. A partial comparison of the faunas of the Tieté and

330 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Parahyba is possible, and will be given in the lists of species of this
region.

The San Francisco, which occupies the western terrace and flows at
right angles to the coastwise streams to near the 9th degree of latitude,
where it also turns to the east, is the largest river. It has a length of
2000 km. and a basin of 652,000 square kilometers. Its origin is at an
elevation of 1200 m., and about its headwaters ‘‘are some of the highest
elevations in Brazil.’ It has an upper fall of 15 m. near Pirapora and a
lower fall of 80 m. at Paulo Alfonso.

The fish fauna of this area has become known, first, through the work of
Liitken, who has published an excellent account of the fishes of the Rio
das Velhas, an upper tributary of the Rio San Francisco ; second, through
Steindachner, who has published accounts of the fishes of the coastwise
streams, Steindachner basing his account on the very extensive collections
made in the lower and upper courses of all,the streams by Agassiz and
his assistants, and through Eigenmann who has written on the fishes at
Taubaté and the Rio Tiete.

The fauna of this area as a whole is distinguished by the absence of
certain Amazonian forms. It is of increased interest because many genera
absent from this area are found in the La Plata basin as well as in the
Amazon basin, having evidently reached the La Plata through the interior
waterway rather than from stream to stream along the coast.

8. The San Franciscan Province.

The Rio San Francisco being nearer the Amazonian Province has a
larger proportion of Amazonian genera than the streams south of Bahia.
Notably the Serrasalmoninee and the Gymnotidz, present in the San
Francisco, are not recorded from the coast streams. It forms a faunal prov-
ince distinct from the rest of the area to the east.

The following species have been recorded from the Rio San Francisco
and its tributaries, the Rio Preto and Rio das Velhas, and the Rio Cipo, a
tributary of the Velhas.

EXPLANATION OF MARKINGS USED.
L Also found in the La Plata Basin.
A Also found in the Amazons.
* Species peculiar to the San Francisco.

EIGENMANN: FRESH WATER FISHES. 331

** Genus and species peculiar to the San Francisco.
/ Species peculiar to eastern Brazil.
// Genus and species peculiar to eastern Brazil.
AL Pseudopimelodus zungaro (Humboldt).
L Rhamdia hilarid (Cuvier & Valenciennes).
* Rhamdella microcephala (Reinhardt).
| Rhamdella minuta (Liitken).
AL Pimelodella lateristriga (Miller & Troschel).
| Pimelodella vittata (Kroyer).
AL Pimelodus clarias (Bloch).
L Pimelodus valenciennis Kroyer.
AL Pimelodus fur (Reinhardt).
** Bergiaria westermannt (Reinhardt). Genus not in Guiana.

* Conorhynchus controstris (Cuvier & Valenciennes).

Genus not in
Guiana.

** Bagropsis venhardi Litken. Genus not in Guiana.
** Duopalatinus emarginatus (Cuvier & Valenciennes).
Guiana.
L Pseudoplatystoma coruscans (Agassiz).
* Doras marmoratus Reinhardt.
AL Doras costatus (Linnzus).
A Oxidoras niger (Valenciennes).
A Glanidum albescens Reinhardt. Genus not in Guiana.
AL Trachycorystes galeatus (Linnzus).
* Pseudauchenipterus flavescens (Eigenmann & Eigenmann).
L Pygidium brasiliensts (Reinhardt).
* Stegophilus insidiosus Reinhardt. Genus not in Guiana.
L Plecostomus commersont (Valenciennes).
* Plecostomus garmant Regan.
| Plecostomus wucherert Ginther.
* Plecostomus macrops Eigenmann & Eigenmann.
* Plecostomus alatus Castlenau.
AL Plecostomus vaillant Steindachner.
| Plecostomus auroguttatus (Knert.).
* Pterygoplichthys etentaculatum (Spix.).
| Pterygoplichthys lituratus (Kner.).
L Rhinelepis aspera Spix. Genus not in Guiana.

Genus not in

332 PATAGONIAN EXPEDITIONS: ZOOLOGY.

| Loricaria steindachnert Regan.
* Loricaria nudiventris Cuvier & Valenciennes.
AL Hopthas malabaricus (Bloch).
A Floplerythinus salvus (Agassiz).
* Curimatella lepidurus Eigenmann & Eigenmann. Below the falls.
L Curimatus gilberti Quoy & Gaimard.
* Prochilodus affints Reinhardt.
AL Parodon suborbitalis Cuvier & Valenciennes. Genus not in Guiana.
A Hlemiodus gracilis Ginther.
L Anostomus tsognathus Kner.
AL Characidium fasciatum Reinhardt.
" * Leforinus tentatus Liitken.
* Teporinus reinhardtii Liitken.
AL Leforinus frederict Bloch.
A Leporellus vittatus Cuvier & Valenciennes. Genus not in Guiana.
* Cheivodon piaba Liitken. Genus not in Guiana.
* Hemigrammus sante Eigenmann.
AL Hemigrammus gracilis (Reinhardt).
* Hemigrammus nanus (Reinhardt).
* Menkhausia coste Steindachner.
AL Astyanax bimaculatus lacustris (Liitken).
AL Astyanax fasciatus (Cuvier).
* Astyanax scabripinnis rivularts (Liitken).
| Brycon carpaphagus Cuvier & Valenciennes.
* Brycon lund Liitken.
AL Brycon hilarit (Cuvier & Valenciennes).
/ Brycon reinhardtii Liitken.
* Creagrutus argenteus Reinhardt. Genus not in Guiana.
A Chalcinus giinthert Garman.
AL Pygocentrus piraya Cuvier.
AL Servasalmo marginatus Valenciennes.
* Myleus micans Liitken.
* Myleus altipinnes Cuvier & Valenciennes.
A Reboides xenodon Liitken.
AL Salminus hilarit Cuvier & Valenciennes. Genus not in Guiana.
A Salminus brevidens Cuvier.
* A cestrorhynchus lacustris (Liitken).

———————

ee

EIGENMANN: FRESH WATER FISHES. 333

L Sternarchus brastlensts Reinhardt.
AL Figenmannia virescens (Valenciennes).
AL Sternopygus macrurus Miller & Troschel.
AL Gymnotus carapus Linnzus.

? Synbranchus.

? Arapaima.

A Pachyurus squamipinnis Agassiz.

* Pachyurus francisce Cuvier & Valenciennes.

L Crtharichthys spilopterus Giinther. Marine.

While this list shows many species and several peculiar genera
as a whole, except for the absentees of many Amazonian genera and
species, the fauna of this river does not differ more from that of the
Amazon at Para than that of Para differs from the fauna of the Amazon at
Tabatinga. Its fauna is much more like that of the Amazon than the fauna
of the Colorado is like that of either the Rio Grande or the Lerma. A cer-
‘tain number of genera and species may always be expected to be peculiar.

Has this area a primitive fauna and did it serve as a center of distribu-
tion? If the theory of the population of the lowland streams from these
ancient highlands is correct, we may reasonably infer that the genera
now occupying this region are part of the ancient fauna. The San Fran-
cisco, occupying an isolated position in the very center of the old land
mass, may serve as the basis of the discussion. Only the fresh-water
families are of consequence in this discussion. The Bunocephalde,
Arapainide, Lepidosirentde, Osteoglossida, and Argiude are absent.
Some of these, as the 4vgzde, might have developed from other forms
that may have emigrated from this place, but not so the others. For
these others the Brazilian plateau cannot have been the center of
divergence. This is also the case with the Sevrasalmonine, Myline
and Gymmnotide. These are found in the San Francisco, but-not in the
coastwise streams east of the San Francisco basin. Have they in the
past migrated from this place as a center? Possibly, but probably not,
for while these are found to the west, north, and south of the San Fran-
cisco basin, they are not found to the east in the coastwise streams. It
is, therefore, very probable that the San Francisco obtained these forms
from the west, where they swarm and from where the south certainly ob-
tained them, and that they have not yet succeeded in crossing the eastern
boundary of the San Francisco. It is less probable that they migrated

334 PATAGONIAN EXPEDITIONS: ZOOLOGY.

from the San Francisco westward and not eastward. This applies with
equal force to all other San Franciscan genera found in the Amazon and
not in the coastwise streams.'

How is it with the genera Pseudoplatystoma, Rhamdia, Rhamdella,
Pimelodella, Pimelodus, Astyanax and all the other genera of the San
Francisco that have a wide distribution?! These probably belong to the
original fauna, or had their origin in the San Francisco, or rather Brazilian
plateau and have become distributed from it. Guiana has of course an
equal claim’ on them; and there is no evidence that they have become
distributed from the San Francisco rather than from Guiana.

There remain only the genera peculiar to the region, and, inasmuch as
they have not become distributed they do not enter the question.

There is some evidence that part of the San Francisco fauna has been
derived from the Amazon basin. This evidence consists of the fact that
the San Francisco, which is nearer the Amazon basin, contains a larger
number of Amazonian genera than the eastern streams, and in part, of the -
fact that the peculiarly lowland forms of the Amazon have not reached this
province, and in part, of the presumption in favor of the Amazon because
it has unquestionably acted as the center of distribution of many lowland
forms that have reached the La Plata basin without entering the province
under consideration.”

It is very probable that the San Francisco has contributed to the fauna
to the south of it (see under the La Plata, etc.).

9. The Coastal Province.

The fauna of the second terrace and coastal plain from Bahia to Iguapé,
whose rivers were described in a preceding section, shows a proportion
of species peculiar to the region about equal to the proportion of peculiar
species in the Pacific province, not considering marine derivatives and
Poeciliidee. There are eight peculiar genera, as compared with three in the
Pacific province. This area contains a very much larger percentage of
species that are also found in the Amazon than does the Magdalena.

Fifty genera present in the Amazon and La Plata basins are absent from
the large streams between Bahia and Rio de Janeiro; over thirty of these
are also absent from the San Francisco. They demonstrate their migra-

"See in the Guiana list those marked with an S.
* Mr. Haseman reports an open road for the passage of fishes from the Tocantins.

EIGENMANN: FRESH WATER FISHES. 335

tion into the La Plata basin from the Amazon by some inland route, and
that for the thirty and more genera at least, that are absent from this area
and from the San Francisco, this territory could not have been the center
of dispersal.

The fifty genera thus present in the Amazon and La Plata and absent
from the coastal rivers exceed the total number of genera recorded from
them.

The following genera found in the La Plata and Amazon have not
been found in the coastwise streams. Those not recorded from the
Magdalena are marked+; those not in the San Francisco X.

x + Bunocephalus, + Curimatella, + Serrasalmo,
xX + Dysichthys, xX + Anodus, x + Praractus,
Pimelodus, +FParodon, xX + Metynuis,
X + Luciopimelodus, + Hlemtodus, + Myleus,
Pseudoplatystoma, xX + Aunzsetsia, x + Colossoma,
xX + Paulesca, xX + Nannostomus, X + Mylossoma,
x Sorubim, Anostomus, xX + Charax,
2 Doras, + Charax, xX + Roestes,
+ Oxydoras, Leporellus, Reboides,
+ Centromochtlus, xX + Holoshesthes, Cynopotamus,
+ Stegophilus, X + Odontostilbe, Salminus,
xX Cetopsts, Chetrodon, + Acestrorhynchus,
xX <Ageneiosus, X + Aphyocharax, Eugenmannia,
x + Hoplosternum, + Hemigrammus, Gymnotis,
xX + Hemiodontichthys, X-+ Gasteropelecus, + Giton,
x + Sturisoma, Chalcinius, xX Aguidens,
xX + Hypoptopoma, xX + Piabucus, x + Brotodoma,
xX + Prrrhutina, X + Pyzoprists, X + Cichlosoma.
x -+ Psectrogaster, + Pygocentrus,

This area is also characterized by positive features, for we have the
genera Stecndachneria, Pogonopoma, Werthemeria, Hemipsilichthys, Del-
turus, Pavrotocinclus, Deuterodon, Henochilus, Harttia, Hollandichthys,
Typhlobagrus, Calurichthys, Kronichthys, peculiar to this province.

In the number of peculiar genera this province differs more from the
Amazon than does that across the Andes in Eucador and Peru (see p.
312) and if the degree of difference is a criterion, it is an older fauna.

336 PATAGONIAN EXPEDITIONS: ZOOLOGY.

List of Species With Their Known Distribution.

In the following list :

A indicates that the species has been recorded from the Amazon.

L indicates that the species has been recorded from the La Plata, in-
cluding Rio Grande do Sul.

** that the species is peculiar to this province.

** that the genus is peculiar to this province.

All the observations on the San Francisco basin as a possible center
of distribution apply with equal force on the coastal province.

There are two facts that deserve emphasis: First at Taubaté, in the State
of Sao Paulo on the Rio Parahyba, fossil fishes have been found in bitu-
minous shales. The definite geological horizon of these rocks has not
been determined and they contain no other fossils. These fossils belong,
according to Smith Woodward, to the genera 47s, a genus now very
abundant along the coasts of tropical America and entering fresh
water ; Zetragonopterus, or its ally Astyanax, still abundant in the
Parahyba; Percechthys, now confined to the Patagonian region, and 4cara,
a Cichlid. Von Ihering contends that the Percichihys is not a member
of the Serranide to which Percichthys belongs, but is a Crchfd, and
Jordan makes a new genus Zoérycon of the fossil Zetragonopterus.

Von Thering is inclined to the opinion that the shales were deposited
during the Tertiary in a lake more than 120 km. long, occupying the
valley of the Parahyba from the Serra de Mantaqueira to that of Bocaina.

Second, while there are many genera in the Amazon and the La Plata
not found in eastern Brazil, there are also a number of species in the La
Plata basin and eastern Brazil that have not been found in the Amazon.
These are enumerated in the list of species of the coastwise streams and
indicate that there has been a certain amount of migration between the
La Plata and this region. The possible means of communication is indi-
cated in the description of the Parahyba.

EIGENMANN: FRESH WATER FISHES.

337

LA Genidens genidens (Cuv. & Val.).
* Pseudopimelodus parahybe Steindachner.
A P. raninus (Cuv. & Val.).
A Rhamdia sebe (Cuv. & Val.).
LA R. guelen (Quoy & Gaim.).
* Rhamdella exsudans (Jenyns).
L R. jenynsit (Giinther).
* R. minuta (Liitken).
* Acentronichthys leptos E. & E.
A Pimelodella cristatys (Miller & Trosch. ).
* P. pectinifer KH. & E.
LA P. lateristriga (Miiller & Troschel).
* P. harttii (Steindachner).
* P. eigenmanni (Boulenger).
* Conorhynchos glaber (Steind. ).
A Brachyplatystoma vaillanti (Cuv. & Val.).

* Steindachneria amblyura Eigenmann & Eigenmann.

* S. doceana E. & E.
* S. parahybe (Steindachner).

*k Werthemeria maculata Steindachner.

A Glanidium albescens Reinhardt.

LA TZrachycorystes striatulus (Steindachner).
* Pseudauchenipterus jequitinhonhe (Steindachner).
A P. affinis (Steindachner).

A P. nodosus (Bloch).

L Pygidium brasiliensis (Reinhardt).

** Scleromystax barbatus (Quoy & Gaimard).

LA Callichthys callichthys Linneus.

* C. arcifer Hensel.

* Corydoras nattereri Steindachner.
* Plecostomus punctatus Cuv. & Val.
* P. un@ Steindachner.

L P. wuchereri Giinther.

* P. liitkent Steindachner.

*k Pogonopoma wertheimert (Steindachner).
* Rhinelepis parahybe Steindachner.

*k Hemipsilichthys gobio (Liitken).

*k Delturus angulicauda (Steindachner).

* D. parahybe E. & E.

A Ancistrus stigmaticus E. & E.
* Oftocinclus affinis Steindachner.
* O. notatus E. & E.

**k Parotocinclus maculicauda (Steindachner ).
* Microlepidogaster lophophanes (KE. & E.).
* Loricaria nigricauda Regan.

A L. konopickyé Steindachner.

L. steindachneri Regan (San Francisco. )

A L. lima Kner.

PA L. acuta Castelnau.

* LZ. spixit Steindachner.
* Harttia loricariformis Steind.
A WNeoplecostomus granosus (Cuv. & Val.).

Tieté.
Rio Grande

do Sul.

Ribeira.!

Santos.

Joao.

Macahe.

Parahyba.
Itabapuana.

Quenda.

pee:
Ss. Matheos.

Mucury,
Jequitinhonha
Pardo.

Bahia and
Paraguassu.

x

XXXxXXX ~~ X {Rio de Janeiro

x

x XK KX

x

x X

x

x Se Sek

GIS

Kee

x X XK X

x

x x

X_ | Porto Seguro.

x X

x X

x Xx

'For a list of the species of this river see page 372.

338 PATAGONIAN EXPEDITIONS:

ZOOLOGY.

Tieté.
Rio Grande

do Sul.
Ribeira.

Santos.

Joao.

Macahe.

Parahyba.
| Itabapuana.

Quenda.

Doce.
S. Matheos.

Mucury.

Porto Seguro.

Pardo.

Bahia and
Paraguassu.

LA Hopiias malabaricus (Bloch).

_ LA Aoplerythrinus unitentatus (Spix).

A £Erythrinus erythrinus (Bloch).

* Curimatus elegans Steindachner.

* C. elegans bahiensis E. & E.

L C. gilberti Quoy & Gaimard.

* C. nagelii Steindachner.

* Prochilodus vimboides Heckel.

* P. brevis Steindachner.

* P. hartit Steindachner.

L P. scrofa Steindachner.

L Characidium fasciatum Reinhardt.

* Leporinus macrolepidotus Peters.

A L. melanopleura Giinther.

. megalepis Giinther.

. dbimaculatus Castelnau.

. steindachnert Eigenmann.

. frederici (Bloch).

. bahiensis Steindachner.

. copelandi Steindachner.

. mormyrops Steindachner.

. conirostris Steindachner.

. macrolepidotus Peters.
* Tetragonopterus gibbosus Steindachner.
A Fowlerina orbicularis Cuv. & Val.
* Maenkhausia doceanus (Steindachner).

LA Astyanax bimaculatus (Linnzus).
* A. janetroensis Eigenmann.

LA 4. fasciatus (Cuvier).
* 4. fasciatus parahybe Eigenmann.
* A, fasciatus jequitinhonhe Steindachner.
* 4. brevirhinus Eigenmann.
* 4. giton Kigenmann.
* 4. bahiensis (Steindachner).
* A. scabripinnis (Jenyns).
* 4A. scabripinnis laticeps (Cope).

L

XR XR KE XP e
NWNNWNANWNWNW

* A. scabripinnts longirostris Steindachner.

* A. scabripinnis intermedius Eigenmann.
* A. elgenmanniorum (Cope).

L A. teniatus (Jenyns).
** Henochilus wheatlandi Garman.

Hollandichthys multifasciatus (Eigenmann & Norris).

* Deuterodon rose (Steindachner).

* D. pedri Figenmann.

* D. parahybe Eigenmann.

* D. iguape Eigenmann & Norris.

* Brycon ferox Steindachner.

* B. devillet Castelnau.

* B, reinhardti Liitken.

* Paragoniates microlepis Steindachner.

* Acestrorhamphus macrolepis (Steindachner).

x

x X

xX XX |Rio de Janeiro

xX -*XX

xX X

x X

KOCK eer

OC eNO

x X

x
x X

Xx X

Xx X

x

x |

X_ | Jequitinhonha

x

x X

xXx XX

x X x

Xx X

1 Cubatao.

EIGENMANN: FRESH WATER FISHES. 339

2 Slelelel al (2 ./ 8a] lve
= {So |] 3) 919/9] 2/2) salsi4] lsidlae
Bl os mal Ves! allo | 1s) 21S)" | &
| od Z na n | 2 | ei Ay
L A. jenynsit (Giinther). x
L A. hepsetus (Cuvier). x x x
LA Synbranchus marmoratus Bloch. | x
A Arapaima gigas Cuvier. x
* Pachypops adspersus Steindachner. | | x
L Cichlasoma facetum Jenyns. X|X | | |
L C. autochthon Giinther. x x |X |X|}
A C. severus Heckel. |X
L Crenicichla lacustris Castelnau. X|xX x x x
L Geophagus brasiliensis Quoy & Gaimard. x X|X|X| xX] x |X| x1 X| XX] |X x
G. gymnogenys Hensel. x | |

10. The La Plata Province.

Of the streams draining toward the south nothing is known ichthyo-
logically of the Uruguay and very little of the upper Parana, while the
Paraguay, owing to the work of Natterer, Ternetz and Anisits, is one of
the best known regions of South America.

The Paraguay is a lowland stream, the elevation at its sources at
Cuyaba being but 200 m. At Corumba it has an elevation of 140 m., at
Puerto Pachero 100 m., at Asuncion it has an elevation of about 70 m.
Its sources are in contact with the sources of the Guaporé, a tributary of
the Madeira, and those of the Tapajos.

The upper Parana is, in distinct contrast to the Paraguay, a mountain
stream arising in the Serra da Mar and the southward extension of the San
Francisco plateau. The Rio Grande de Minas, one of its largest tribu-
taries, has at Porta da Rifano an elevation of 590m. The upper Parana
is separated from the lower Parana by the Gran Salto de Guaira, where it
leaps over falls of 18 m. to the lowland. _ Its upper tributaries arise mostly
in the coast range near the Atlantic coast and reach the main stream after
numerous falls.

A list of the fishes of the Paraguay basin is sufficiently startling on
account of the few things peculiar to it, and the many species and genera
it has in common with the Amazon, to warrant giving it in detail.

A separate list of species and genera for the Parana, La Plata and Rio
Grande do Sul gives us an idea of the relationship of the fauna to that of
the Amazon and to that of southeastern Brazil.

There is no room for doubt on the source of the fish fauna of this

340 PATAGONIAN EXPEDITIONS: ZOOLOGY.

region. It came from the north, in small part from the eastern plateau,
in greater part from the Amazon valley. The question of the time of
the origin of the fauna is of interest. The only observations, as far as I
know, directly bearing on this question are those of A. Smith Woodward,
1900, who recorded from,the Parana formation fragments of ‘47zzs,
Pimelodus, Platystomus and other genera which still live in the freshwaters
of South America” and also fragments of characinoids. The formation
is a marine deposit, near the city of Parana, which he thinks ‘is truly of
late Tertiary age, and may probably be correlated with the Pliocene of the
northern hemisphere.”” As these deposits contain remains of land-mam-
mals, sharks’ teeth and fresh-water fishes it is probably an estuary deposit
of the former mouth of the Parana river. The present fauna of this
stream developed with the stream itself.

List OF FISHES RECORDED FROM THE PARAGUAY BASIN.

In the following list, ‘‘*”’ indicates that the species is peculiar to the
Paraguay basin; ‘**” that both the genus and species are peculiar ;
“A” indicates that the species is also found in the Amazon basin; ‘‘a” the
genus is found in the Amazon basin; and ‘/” that the species is peculiar
tothe La Plata. ‘C”’ that it is also found in the coast streams of Eastern
Brazil and ‘‘S”’ that it is also found in the San Francisco.

A Potamotrygon hystrix Miller & Troschel.

A Potamotrygon dumeritit Castelnau.

a®* Dysichthys australe Eigenmann & Ward.

a®* Bunocephalus rugosus Eigenmann & Kennedy.

a/l Bunocephalus theringw Boulenger.

a* Bunocephalus dorie Boulenger.

A Pinarampus pirinampu (Spix).

A Luciopimelodus platanus Giinther.

AS Pseudopimelodus zungaro (Humboldt).
* Pseudopimelodus cottoides Boulenger.
| Heptapterus mustelinus Valenciennes.
AC Rhamdia quelen (Quoy & Gaimard).

A Rhamdia sebe knert Steindachner.

A Pimelodus ornata Kner.

* Pimelodus albicans (Cuvier & Valenciennes).
AS Pimelodus clarias (Bloch).

EIGENMANN: FRESH WATER FISHES. 341

S Pimelodus valenciennis (Kroyer).
AS Pimelodus fur Reinhardt.
** Theringichthys labrosus (Liitken).
** Theringichthys megalops Eigenmann & Ward.
A Pimelodella gracilis (Valenciennes).
* Pimelodella tentophorus Regan.
ACS Pimelodella lateristrviga (Miller & Troschel).
* Pimelodela mucosa Eigenmann & Ward.
A Sciades pictus (Miller & Troschel).
A Flemtsorubim platyrhynchus (Cuvier & Valenciennes).
aS Pseudoplatystomus coruscans (Agassiz).
A Sorubim tima (Bloch & Schneider).
A Doras granulosus Valenciennes.
AS Doras costatus (Linnzus).
* Doras maculatus Valenciennes.
* Doras nebulosus Eigenmann & Kennedy.
A Doras weddelit Castelnau.
/* Oxydoras knerit (Bloch).
* Oxydoras eigenmannt (Boulenger).
* Hemidoras paraguayensts Eigenmann & Ward.
aA Trachelopterus cortaceus (Cuvier & Valenciennes).
*® Auchenipterus nigripinnis (Boulenger).
AS Trachycorystes galeatus (Linnzus).
AC Trachycorystes striatulus Steindachner.
A Agenetosus valenciennest Bleeker.
A Agenewosus brevifilis (Cuvier & Valenciennes).
A HHypophthalmus edentatus Spix.
* Pyowdium borelli Boulenger.
IS Pygidium brastliensts (Reinhardt).
| Pygidium cordovensis WWeyenberg.
** Flomodeatus antsitst Eigenman & W ard.
C Plecostomus gohni Steindachner.
A FPlecostomus plecostomus (Linnzus).
S Plecostomus commersoni (Valenciennes).
S Plecostomus vaillantt Steindachner.
* Plecostomus borelii Boulenger.
Plecostomus robin Cuvier & Valenciennes.

342 PATAGONIAN EXPEDITIONS: ZOOLOGY.

CS Plecostomus wucherert Giinther.
A Hemiancistrus vittatus (Steindachner).
A Cochhodon cochhodon Kner.
* Pterygoplichthys multivadiatus Bloch.
* Pterygoplichthys anisitst Eigenmann & Kennedy.
* Prerygoplichthys guvens Eigenmann & Kennedy.
* Pterygoplichthys gigas Boulenger.
A Pseudancistrus barbatus Cuvier & Valenciennes.
A Xenocara gymnorhynchus Kner.
A Aucistrus cirrhosus Valenciennes.
A Ancistrus cirrhosus dubius Figenmann & Eigenmann.
A Auncistrus hoplogenys (Giinther).
* Oxyropsis inexpectatum Holmberg.
* Ofocinclus vittatus Regan.
A Hemiodontichthys acipenserinus (Kner).
a®* Sturisoma.vrobusta Regan.
* Sturisoma barbata Kner.
* LToricaria parva Boulenger.
%@ Lovricaria catamarcensis Berg.
A Loricaria phoxocephala Eigenmann & Eigenmann.
* Loricarta maculata Bloch.
A Loricaria typus Bleeker.
* Loricaria labialis Boulenger.
| Loricaria anus Valenciennes.
A Loricaria cataphracta Linneus.
A Loricaria carinata Castlenau.
* Toricaria apeltogaster Boulenger.
* Loricartia macrodon Kner.
* Lortwaria laticeps Regan.
* Loricaria platycephala Kner.
a®* Corydoras microps Eigenmann & Kennedy.
* Corydoras paleatus Jenyns.
* Corydoras aurofrenatus Eigenmann & Kennedy.
* Corydoras australe Eigenmann & Ward.
AC Callichthys caliichthys Linnzus.
* Caliichthys callichthys asper Quoy & Gaimard.
* Callichthys callichthys hemiphractus Hensel.

EIGENMANN: FRESH WATER FISHES. 343

* Hoplosternum pectoralts (Boulenger).
A Hoplosternum littorale (Hancock).
ACS Hofplias malabaricus (Bloch).
AC Hofplerythrinus unitentatus (Spix).
/A Pyrrhulina australe Eigenmann & Kennedy.
[A Pyrrhulina brevis Steindachner.
[A Psectrogaster curviventris Eigenmann & Kennedy.
A Curimatella alburnus (Miiller & Troschel).
* Curimatella alburnus australe Eigenmann & Kennedy.
A Curimatus spilurus Giinther.
* Curimatus gilli Figenmann & Kennedy.
A Curimatus nasus Steindachner.
* Curimatus nigrotenia Boulenger.
* Curimatus elegans nitens Holmberg.
A Curimatus bimaculatus Steindachner.
A Curimatus rutiloides Kner.
C Curimatus gilbert? Quoy & Gaimard.
A Anodus latioy Spix.
aS Prochilodus argenteus Agassiz.
C Prochilodus scrofa Steindachner.
| Prochilodus tineatus (Valenciennes).
* Anisitsta othonops Eigenmann & Kennedy.
A Hemiodus unimaculatus Bloch.
* Flemtodus semttentatus Kner.
A Hemiodus microlepis Kner.
A Parodon suborbital’s Cuvier & Valenciennes.
* Parodon gestri Boulenger.
* Parodon tortuosus Eigenmann & Norris.
* Parodon paraguayensis Eigenmann.
a®* Nannostomus lateralis (Boulenger).
S Schizodon tsognathus Kner.
* Schizodon borellit (Boulenger).
A Schizodon fasciatus Spix.
A Lahithella nasutus Kner.
A Leforinus striatus Kner.
AC Leforinus friderict Bloch.
A Leporinus obtusidens Valenciennes.

344 PATAGONIAN EXPEDITIONS: ZOOLOGY.

A Leporinus trifasciatus Steindachner.
A Leporinus eques Steindachner.
A Leporinus affints Giinther.
A Leportinus hypselonotus Giinther.
AC Leporinus conirostris Steindachner.
A Leporinus fasciatus Agassiz.
AC Characidium fasciatum Reinhardt.
Characidium laterals (Boulenger).
* Characidium borelli (Boulenger).
a®* Odontostilbe paraguayensis Eigenmann & Kennedy.
* Odontostilbe trementine Eigenmann & Kennedy.
* Chetrodon ribetrvot Eigenmann.
a® Chetvodon interruptus (Jenyns).
* Chetrodon calhurus Boulenger.
A Chetrodon insignis Steindachner.
A Chetrodon natterert Steindachner.
A Holoshesthes pegeira (Steindachner).
a* Aphyocharax dentatus Eigenmann & Kennedy.
* Aphyocharax stramimeus Eigenmann.
A Aphyocharax alburnus Ginther.
* Aphiocarax rathbuni Eigenmann
A Afphyocharax anisitst Higenmann & Kennedy.
AS Hemigrammus gracilis (Reinhardt).
2? Hemigrammus callstus (Boulenger).
* Hlemigrammus antsitst Eigenmann.
Hemigrammus liitkent (Boulenger).*
* Hemigrammus tridens Eigenmann.
* Hlemigrammus ulrey? (Boulenger).
* Hemigrammus kennedy: Eigenmann.
A Tetragonopterus argenteus Cuvier.
A Ctenobrycon hauxwellianus Cope.
* Gymunocorymbus ternetzt Boulenger.
* 4styanax allent Eigenmann & McAtee.
ACS Astyanax fasciatus (Cuvier).
* Astyanax pellegrint Eigenmann.
A Astyanax abramis (Jenyns).

' Paraguay and southeastern Brazil.

EIGENMANN: FRESH WATER FISHES. 345

ACS Astyanax bimaculatus Linneus.
* Astyanax lineatus Holmberg.
** styanacinus moor (Boulenger).
A Menkhausia agassizt Steindachner.
A Menkhausia dichrourus Kner.
A Menkhausia lepidurus (Kner).
* Bryconamericus mankhausi Eigenmann & Kennedy.
A®* Bryconamericus exodon Eigenmann.
| Bryconamericus theringw Boulenger.
AC Fowlerina paraguayensis Eigenmann.
a Brachychalcinus retrospina (Boulenger).'
* Brycon hilari (Cuvier & Valenciennes).
* Brycon microlepis Perugia.
| Brycon orbignianus (Cuvier & Valenciennes).
A Creatochanes melanurus (Bloch).
A Thoracocharax stellatus Kner.
* Chalcinus paranensis Giinther.
A Chalcinus angulatus Spix.
A Chalcinus angulatus curtus Garman.
AS Pygocentrus piraya Cuvier.
A Pygocentrus natterert (Kner).
A Pygoprists serrulatus Cuvier & Valenciennes.
AS Serrasalmo marginatus Valenciennes.
A Serrasalmo spilopleura Kner.
A Serrasalmo gymnogenys Giinther.
A Serrasalmo humerais Cuvier & Valenciennes.
A Serrasalmo rhombeus Linneus.
* Metynnuts mola Eigenmann & Kennedy.
A Metynnis hypsauchen Miller & Troschel.
A Myleus asterias Miller & Troschel.
A Praractus brachypomus (Cuvier).
A Mylossoma aureus (Agassiz).
A Mylossoma albiscopus Cope.
* Myleus levis Eigenmann & McAtee.
A Charax gibbosus Bloch.
* Chavax squamosus Eigenmann & Kennedy.

"This species at Santa Cruz and Paraguay.

346 PATAGONIAN EXPEDITIONS: ZOOLOGY

* Characinus calliurus Eigenmann.

A Restes molossus Kner.

A Reboides microlepis Reinhardt.

| Reboides bonariensts Steindachner.
* Reeboides prognathus Boulenger.

A Cynopotamus humeratis (Valenciennes).
Cynopotamus magdalene Steindachner."
Cynopotamus knert Steindachner.

A Salminus brevidens Cuvier.

A Acestrorhynchus ferox (Bloch).

C Acestrorhamphus hepsetus (Cuvier).

A Raphiodon vulpinus Spix.

A Sternarchus albifrons Linnzus.

A Rhamphichthys reinhardtii Kaup.

A Rhamphichthys marmoratus Castelnau.

A Hypopomus brevirostris Steindachner.

AS Sternopygus macrurus Bloch & Schneider.
AS Eigenmannia virescens (Valenciennes).
AS Gymnotus carapo Linneus.

AC Synbranchus marmoratus Bloch.

A Stolephorus olidus Giinther.

A Rivulus punctatus Boulenger.

A Girardinus caudomaculatus Hensel.
Cnesterodon decemmaculatus Jenyns.

* Fundulus paraguayensis Eigenmann & Kennedy.
Fundulus balzantt Perugia.

** T1yodon paraguayense Eigenmann.

A Potamorrhaphis guianens?s (Schomburgk).

A Tylosurus amazonicus (Steindachner).

| Pachyurus bonariensis Steindachner.
A Pachyurus schomburgki Giinther.
a* Plagioscion ternetzt Boulenger.
a* Chetobranchopsis australe Eigenmann & Ward.
A Astronotus ocellatus (Agassiz).
A 4&quidens tetramerus (Heckel).
A quidens portalagrensis (Heckel).

1 Paraguay and Magdalena. This species is recorded on the doubtful authority of Perugia.

EIGENMANN: FRESH WATER FISHES. 347

A Azquidens dorsigera (Heckel).

* iquidens paraguayensts Eigenmann & Kennedy.

A Asquidens vittata (Heckel).

A Cichlasoma bimaculata (Linnzus).

* Fleterogramma corumbe Eigenmann & Ward.
* Fleterogramma borelli Regan.

* Heterogramma trifasciatum Eigenmann & Kennedy.

A Mesonauta festivus (Heckel).

A Crenicichla lepidota Heckel.

A Crenicichla adspersa Heckel.

A Crenicichla vittata Heckel.

A Crenicichla saxatilis (Linnzus).

| Batrachops semifasciata Heckel.

* Batrachops ocellata Perugia.

* Satanoperca pappaterra Heckel.

* Geophagus balzanit Perugia.
Geophagus ejurupari Heckel

| Achirus jenynsid (Giinther).

It is seen that over 53 per cent. of the total are species that are also
found in the Amazon basin; 36 per cent. are peculiar, and over 15 per
cent. are common to the Paraguay and eastern Brazil. The proportions
common to the Amazon and the Paraguay are not the same for all genera.
All of the species of Sternopygide (7), Schizodon (4), Leporinus (7)
Serrasalmonine (8), and five out of seven AZy/nve@ found in the Paraguay
basin are also found in the Amazon. A comparison of the number of
species common to the Paraguay and to the Amazon and to the Paraguay
and eastern Brazil leaves no doubt as to the origin of the fauna of the
Paraguay. It came from the Amazon.

THE PARANA AND LA PLATA.

The fauna of the Parana and La Plata, as far as known, is quite
different from that of the Paraguay. Much of this difference is probably
apparent rather than real. The Parana is divisible into two regions —
that above the falls with its tributaries belonging to the Brazilian plateau
has been mentioned before. Its fauna is very imperfectly known from
collections made by Natterer at Irisanga and incidentally described by
Kner, and from a collection made by Von Ihering in Sao Paulo and
described by Eigenmann & Norris.

348 PATAGONIAN EXPEDITIONS:

ZOOLOGY.

SPECIES OF THE UppER PARANA, LOWER PARANA, LA PLATA AND Rio GRANDE DO SUL.

S:-E.
Brazil.

Upper
Parana.

Exomegas macrostomus (Burmeister).
Galeus vulgaris Miller & Henle.

Raja microps Giinther.

R. platana Giinther.

Myliobatis aquila (Linnzus).
Potamotrygon brachyurus Ginther.

P. hystrix (Miiller & Henle).
Bunocephalus theringit Boulenger.
Gentidens genidens (Cuvier & Valenciennes).
Netuma upsulonophorus (Eigenmann & Eigenmann).
LV. barbus (Lacépéde. )

Arius agassiztt Kigenmann & Eigenmann.
Luctopimelodus pati (Valenciennes).

L. platanus (Giinther).

Pseudopimelodus cottoides Boulenger.

Ps. zungaro (Humboldt).

Rhamdia quelen (Quoy & Gaimard).

&. sapo (Valenciennes).

R. hilarit (Cuvier & Valenciennes).
Rhamdella ertarcha Eigenmann & Eigenmann.
R. jenynstt (Giinther).

R. straminea Cope.

LHeptapterus mustelinus (Valenciennes).
Acentronichthys collettit (Steindachner).
Imparfinis piperatus Eigenmann & Norris.
Nannoglanis bifasciatus Eigenmann & Norris.
Pimelodus clarias (Bloch).

P. valenctennis Kroyer.

P. fur Reinhardt.

Bergiaria platana (Steindachner).
Lheringichthys labrosus (Kroyer).
Pimelodella lateristriga (Miiller & Troschel).
P. gracilis (Valenciennes).

P. eigenmanni Meek.

Paulicea jahu von Thering.
Pseudoplatystoma coruscans (Agassiz).
Sorubim lima (Bloch & Schneider).
Doras granulosus Valenciennes.

Oxydoras a’ orbigny Kroyer.

Ageneiosus valenciennest Bleeker.
Pygidium brasiliensis (Liitken).

P. borelli (Boulenger).

LP. schmidti (Berg.).

P. minutus (Boulenger).

P. spegazzinit Berg.

Cetopsis goboides Kner.

Henonemus maculatus (Steindachner).
Callichthys callichthys (Linnzus).
Hoplosternum littorale (Hancock).
Corydoras paleatus (Jenyns).

Plecostomus plecostomus (Linnzus).

P. regani von thering.

x XX

xX X

Ke OK Xx

x X

Lower

Parana Rio Grande
La Plata. Sub
x
x
x
x
x
x
x
x
x x
x
xX x
x
x
x
x
x x
x x
x x
x
x
x
x x
x
x x
x x
x
x
x
x ?
x
xX
x
x
x :
x
x
x
x
x
x
x
x
x
x

EIGENMANN : FRESH WATER FISHES.

349

SPECIES OF THE Upper PARANA, LOWER PARANA, LA PLATA AND Rio GRANDE DO SUL.

SHE:
Brazil.

Upper
Parana.

Lower
Parana,
La Plata.

Plecostomus commersoni (Cuvier & Valenciennes).
P. tretensis yon Thering.

P. garmant Regan.

P. hermanni von Shering.

P. robini Cuvier & Valenciennes.

P. paulinus von thering.

Ancistris cirrhosus Valenciennes.
Otocinclus flexilis Cope.
Microlepidogaster nigricauda (Boulenger. )
M. perforatus Kigenmann & Eigenmann.
Loricaria cadee Hensel.

. strigilata Hensel.

. microlepidogaster Regan.

. latirostris Boulenger.

. spixit Steindachner.

. anus Cuvier & Valenciennes.

. macrops Regan.

. vetula Cuvier & Valenciennes.
Hoplias malabaricus (Bloch).
Lloplerythrinus uniteniatus (Spix).
Curirmatus spiluropsis Eigenmann & Eigenmann,
C. gilberti Quoy & Gaimard.

C. gilbert’ brevipinnis Kigenmann & Eigenmann.
C. platana Giinther.

Prochilodus reticulatus Valenciennes.

P. platensis Holmberg.

P. scrofa Steindachner.

P. lineatus (Valenciennes).

P. nigricans Agassiz.

Parodon piracicabe Eigenmann.

P. affinis Steindachner.

P. tortuosus Kigenmann & Norris.
Nannostomus lateralis Boulenger.
Anostomus vittatus Cuvier & Valenciennes.
A. tsognathus Kner.

A. plate Garman.

A. nasutus Kner.

Characidium fasciatum Reinhardt.
Characidium tenuis Cope.

Leporinus friderict Bloch.

L. obtustdens Valenciennes.

L. copelandi Steindachner.

L. afinis Giinther.

L. solarit Holmberg.

Chetrodon interruptus (Jenyns).
Cheirodon monodon (Cope).
Aphyocharax alburnus Giinther.
Tetragonopterus chalceus (Agassiz).
Hemigrammus litkent Boulenger.
Astyanax erythropterus (Holmberg).
Astyanax correntinus (Holmberg).
Astyanax abramis (Jenyns).

NNNNANWW

x
x

x X

x xX

SS De Oh eS ONeS

x

x

DSO ROROR ON OSS OR NTS xX XX KX

ee OS ONS = PSOSON eeNeS

Rio Grande
do Sul.

x
x

xX XK KKK XK

x

A
350 PATAGONIAN EXPEDITIONS: ZOOLOGY.

SPECIES OF THE Upper PARANA, LowER Parana, LA PLATA AND RIO GRANDE DO SUL.

Astyanax bimaculatus (Linneus).

A. alburnus (Hensel).

A. laticeps (Cope).

A. eigenmanniorum (Cope).

A. cordove (Giinther).

A. fasciatus (Cuvier).

A. rubripictus (Berg).

Bryconamericus theringtt (Boulenger).
Markiana nigripinnis (Perugia).
Brycon lineatus Steindachner.

B. natterert Giinther.

B. orbignianus Cuvier & Valenciennes.
Chalcinus paranensis Giinther.
Pseudocorynopoma dorie Perugia.
Piabucus melanostoma Holmberg.
Diapoma speculiferus Cope.
Pygocentrus natterert (Kner).
Serrasalmo marginatus (Valenciennes).
Piaractus brachypomus (Cuvier).
Metynnis maculatus (Kner).

Myleus tiete (Eigenmann & Norris).
Colossoma mitret (Berg).

C. orbignyanus (Cuvier & Valenciennes).
Mylossoma aureus (Spix).
Asiphonichthys stenopterus Cope.
Reboides bonartensis Steindachner.
Cynopotamus argenteus Valenciennes.
C. humeralis Valenciennes.

C. knerit Steindachner.

Salminus maxillosus Cuvier & Valenciennes.
S. brevidens Cuvier.

S. hilarit (Cuvier & Valenciennes).

Catabasts acumtnatus (Eigenmann & Norris).

Acestrorhamphus brachycephalus Cope.
A. jenynsit Giinther.

A. oligolepis (Steindachner).

A. hepsetus (Cuvier).

Raphiodon vulpinus Spix.

Sternarchus albifrons (Linnzus).
Rhamphichthys marmoratus Castelnau.
Eigenmannia virescens (Valenciennes).
Gymnotus carapus Linneus.
Synbranchus marmoratus Bloch.
Stolephorus clupeoides (Swainson),

S. olidus Giinther.

Llisha flavipinnis (Valenciennes).
Cynolebias bellottii Steindachner.

C. maculatus Steindachner.

C. elongatus Steindachner.

C. robustus Giinther.

Jenynsta lineata (Jenyns).

Girardinus januarius Hensel.

Lower

Sak. Upper Rio Grande
Brazil. Pecan. Dern do Sul.
x x
x
x
x
x
x Ke x x
x
x
x
x
x
x
x ?
x
x
x
x x
x
x
x
x
x
x
x
x
x
x
x x
x
x x x
x x
xX
x
x x
x
x
x
x
x
x x Xx
x x x
x x x
x
x
x
x
x
x
x
x x
x x

EIGENMANN: FRESH WATER FISHES. 351

SPECIES OF THE UPPER PARANA, LOWER PARANA, LA PLaTA AND R10 GRANDE DO SUL.

Lower =
Be He pepe pacar Re ees
razil. arana. TAPinice o Sul.

Gtrardinus caudimaculatus Hensel. x
Cnesterodon decemmaculatus (Jenyns). x
Mugil platanus Giinther. :
Atherinopsts regia (Humboldt).

Pachyurus bonariensis Steindachner.

Cichla niederleinit (Holmberg).

C. chacoensis (Holmberg).

Equidens tetramerus (Heckel).

4. centralis (Holmberg).

4. portalegrensis (Hensel).

“@, minuta (Hensel).

Cichlasoma facetum Jenyns.

C. jenynsti Steindachner.

C. autochthon Giinther.

C. lepidota Heckel.

Batrachops semifasciata Heckel.

B. lacustris Castelnau,

Geophagus brasiliensis (Quoy & Gaimard).
G. brachyurus Cope.

G. gymnogenys Hensel.

Guavina guavina Cuvier & Valenciennes.
Achirus lineatus (Linneus).

A. jenynsit (Giinther).

Symphurus plagusia Bloch & Schneider.

xX KKK Y*KX
x

xX
RPS PRON IR OSPR OR PREK PON

x X
xXXX KKK KKXKXKXM

The Parana below the falls to Buenos Aires has been made known by
Valenciennes, Giinther, Perugia, Berg and La Hille.

At this place may also be considered the fauna of that part of Rio
Grande do Sul draining eastward. The fauna of this area is more distinct
from that of the La Plata than the Paraguayan is from the Amazonian,
only about 46 per cent. of its fishes being identical with species occurring
in the La Plata. Its affinities are, however, with the La Plata rather than
with the coastwise streams north of Rio de Janeiro, only about 30 per
cent. of its fauna being identical with that of the northern rivers.

There are known in the region under consideration 170 species which
may safely be considered to be less than one-half of the total number of
species inhabiting it. Of these, 64 species are also found in the Paraguay,
50, or a little over 29 per cent. are of wide distribution, having been
taken in the Amazon. Thirty-four species are common to this region and
southeast Brazil. Of these, 15 are included in the 49 recorded from the
Amazon, leaving 19 common to the two regions. This number gives us
a hint as to the extent to which southeast Brazil has contributed to the
Ia Plata, or the La Plata/to; the latter.

352 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Of the 31 species recorded from above the falls of the Parana 16 are
also found below it.

There is no indication that the fauna of the Rio San Francisco, or
the Alta Parana, which is geographically nearer to southeast Brazil, is
more like the fauna of the latter region than is the La Plata fauna, for we
have :

Number of species in the Upper Parana and Southeastern Brazil ................ 3
Number of species in the Lower Parana and Southeastern Brazil ................ 4
Number of species in the Alta and Baja Parana and Southeastern Brazil.......... 4
Number of species in the Rio Grande do Sul and Southeastern Brazil............ 8
Number of species in the Rio Grande do Sul, La Plata and Southeastern Brazil... .10
Number Ole speciesningalstoltalOCAli eS reroncte/eleloir cs oieyalelelelelererelel-melestetets: tere renak 5

33

VII. THE ORIGIN OF THE PACIFIC SLOPE FAUNA.

There are four distinct faunas on the Pacific slope of America between
Cape Horn and the tropic of Cancer. One of these is of common origin
with that on the Atlantic slope, one is autochthonus and the other two
are derivative from the Atlantic slope faunas opposed to them.

1. The fauna of southern Chili is essentially like that of Patagonia, and
inasmuch as it is largely made up of marine forms entering fresh water, and
fresh-water forms entering the ocean, it seems very probable that the species
migrated from river to river along the coast from Patagonia to Chili or
from Chili to Patagonia.

2. At the other extreme in the Rio Mezquital of the Transition Region
and the Yaqui just to the north of it there is a fauna essentially like that of
the Rio Grande east of them. As Meek has pointed out, the Yaqui and
Mezquital have captured tributaries of the Rio Grande together with the
fishes in them, and the migration of Atlantic slope northern forms to the
Pacific slope has been a passive one.

Thus, types which in America north of Mexico have not succeeded in
reaching the Pacific slope, have, within the Tropics, crossed the divide.
Etheostoma pottst is the only representative out of over 50 North American
species of Percide that has crossed the divide and has done it in this way.
Ameiurus dugest and priced and /stlarius balsanus are the only North
American catfishes which have crossed the divide. The Rio Mezquital
within the tropics contains Pantosteus plebetus, Hybognathus episcopus,
Leuciscus nigrescens, Notropis ornatus, all species extending northward

EIGENMANN : FRESH WATER FISHES. 353

and found in the north only on the Atlantic slope. It also contains
Ameturus pricet,one of the three Pacific slope catfishes and E¢heostoma
pottst, the only Pacific slope Percoid.

3. The third fauna is the Mexican of the Rio de Santiago. This is un-
doubtedly the relict of an old fauna reénforced by a few immigrants from
the north. It is here not a question of the origin of the fauna from an
eastern one, but of an autochthonous development that has, on its part,
contributed elements to the surrounding rivers. It passively contributed
to the Atlantic slope fauna by having one of its small rivers captured by
the Rio Panuco.

4. Of more particular interest is the origin of the fauna of western
Peru and Ecuador and that of western Central America. Not enough is
known of the fauna of the western part of Central America to attempt
an explanation of its origin.

To quote the words of Gilbert and Sparks, who considered the
marine fishes on opposite sides of Panama (The Fishes of Panama Bay,
205, 1904): ‘The ichthyological evidence is overwhelmingly in favor of
the existence of a former open communication between the two oceans,
which must have been closed at a period sufficiently remote from the
present to have permitted the specific differentiation of a very large ma-
jority of the forms involved.” They found that “Of the 82 families of
fishes represented at Panama all but 3 (Cerdalide, Cirhitide and Nema-
testitde) occur also on the Atlantic side of Central America; ‘‘while of the
218 genera of our Panama list, no fewer than 170 are common to both
oceans.” * ‘Fifty-four out of a total of 374 or 14.4 per cent. of the Pacific
coast species are identical with Atlantic coast species.”

There seems to me to be just as conclusive evidence that the present
fresh-water fauna of the Pacific slope developed affer the obliteration of
the waterway connecting the two oceans, if this waterway occupied the
Amazon valley, and that the Atlantic slope streams of the Cordilleras
became colonized by their present fauna ‘after the obliteration of this
waterway.

The similarity between the Amazonian and Pacific slope faunas of South
America is much greater than the similarity between the Atlantic and
Pacific slope faunas of North America. This similarity leaves the com-
munity of origin of the two faunas unquestioned. Is a continuous coast
line between the two slopes necessary to account for the similarity? Such

354 PATAGONIAN EXPEDITIONS: ZOOLOGY.

BRYCON

Range of the most widely distributed genus

of characins, Astyanax. (It occurs in Trini-
dad.)

Range of the characid genus Brycon.

Range of the most widely distributed cat-
fish genus Rhamdia. (In Trinidad.) Asia.

Range of Synbranchid eel in America and

One of the two South American
genera that have reached Cuba. (In Trini-
dad.) One species.

To ILLUSTRATE THE ORIGIN OF THE PAciFIC SLOPE FAUNA.

EIGENMANN: FRESH WATER FISHES. 355

EIGENMANNIA HOPLIAS

Range of the gymnotid eel, Eigenmannia. Range of the characid genus, Hoplias.
Five species. Three species. (In Trinidad.)

LORICARIA AZQUIDENS

Range of the slender mailed-catfish, Lori- Range of the Cichlid, A2quidens. (In Trini-
Caria. dad.) Probably also in the San Francisco
(white).

To ILLUSTRATE THE ORIGIN OF THE PaciFic SLOPE FAUNA.

356 PATAGONIAN EXPEDITIONS: ZOOLOGY.

PROCHILODUS

Range of the characid genus, Prochilodus. Range of the characid genus Curimatus.
(In Trinidad.)

PLECOSTOMUS

PIMELODELLA

Range of the minute catfishes of the genus Range of the mailed-catfish genus, Pleco-
Pimelodella. stomus. (In Trinidad.)

To ILLUSTRATE THE ORIGIN OF THE PACIFIC SLOPE FAUNA.

EIGENMANN : FRESH WATER FISHES. 357

STERNOPYGUS

GASTEROPELECUS
Known distribution of the Gymnotid genus Known distribution of the genera Gastero-
Sternopygus (three species). pelecus and Thoracocharax, two aberrant

characins.

Known range of the mailed-catfish genus Known distribution of the mailed-catfish
Hemiancistrus. genus Sturisoma.

To ILLUSTRATE THE ORIGIN OF THE PaciFic SLOPE Fauna.

358 PATAGONIAN EXPEDITIONS: ZOOLOGY.

CETOPSINAe
Known range of the Cetopsine. Known range of the characid genus Hemi-
brycon.’ (In Trinidad.)
v
LUCIOGHARAX’ CHEIRODON
Range of the characid genus Luciocharax. Known distribution of the characid genus

Cheirodon. The species are small, many of
them probably unknown. The only tropical
genus represented in Chili.

To ILLUSTRATE THE ORIGIN OF THE PaciFic SLOPE FAUNA.

‘It is probable that this isa nascent genus arising in different places from different species of
the genus Astyanax or Bryconamericus.

EIGENMANN: FRESH WATER FISHES. 359

ee aie
Lee
if

PYGIDIUM CYCLOPIUM

Distribution of the presumably old moun- Distribution of the catfish Cyclopium of
tain catfish, Pygidium. the high Andes.

CHAETOSTOMUS
Distribution of the mailed-catfish Chzeto- Distribution of the Andean characin, Crea-
stomus of the Andes. Pacific side of Panama. grutus.

To ILLUSTRATE THE ORIGIN OF THE PAcIFIC SLOPE FAUNA.

360 PATAGONIAN EXPEDITIONS: ZOOLOGY.

a coast line is not needed, because, first, some of the species of the Pacific
slope are mountain forms, some of them living on both slopes. If such
a genus as Pygzdium has been able to establish itself from stream to
stream along the Andes from Guiana to Patagonia, it needs no continuous
waterway to explain its presence on both slopes. The same for more
limited reasons applies to Ch@efostomus and Arges. Second, a watershed
of moderate height is not necessarily a barrier to the migration of fresh-
water fishes and several species have evidently succeeded in recent times
in crossing the divide in the northwest corner of the continent in one direc-
tion or the other, inasmuch as over 27 per cent. of the entire Pacific fauna
is also found on the Atlantic, a consensus which is significant and conclusive.

Nearly all the genera found on both slopes, the exceptions to be con-
sidered presently, are genera of very wide distribution on the Atlantic
slope, reaching the Magdalena or the Isthmus of Panama.

Only four genera are confined to the Pacific slope: Pavracetopsis
Lebtasina, Pseudochalceus and Saccodon. The other genera represented
on the Pacific slope are distributed as follows :

1. Rhamdia. Magdalena to the Rio de La Plata and Mexico; Lake
Titicaca, San Francisco and Minas Geraes, p. 354.

2. Pimelodella. Rio Chagres at Panama to La Plata; Para to Canelos,
Pp. 356.

3. Paracetopsis. Canelos to Para; Irisanga, p. 358.

4. Pygidium. Andes from Venezuela to Patagonia; French Guiana
and southeastern Brazil, central Argentina, p. 359.

5. Plecostomus. Magdalena to La Plata and Cordova; San Francisco
and southeastern Brazil ; Para to Ambyiacu, p. 356.

6. Hemiancistvus. Eastern Ecuador to Paraguay, Para and Guiana,
P. 357:

7. Chetostomus. Eastern and western slopes from Peru to Panama
and Venezuela, p. 359.

8. Loricaria. Rio Chagres and Magdalena to Para and Buenos Aires ;
Calderon and Canelos, Xeberos, p. 355.

g. Sturtsoma.! Magdalena to Paraguay; Rio Jurua, Nene oe
Xeberos, p: 357.

10. Arges. Andes of Peru, Ecuador, Colombia and Venezuela.

11. Cyclofium. Andes, north of Canelos, p. 359.

' This is probably a genus of diverse origin.

EIGENMANN : FRESH WATER FISHES. 361

12. Hoplias. Chagres and Magdalena to Trinidad, La Plata, Para and
Huallaga, p. 355.

13. Curimatus. Magdalena to Trinidad and La Plata; Para to Hual-
laga and Canelos, p. 356. i

14. Prochilodus. Magdalena to the La Plata and Peru, p. 356.

15. Luciochavax. Magdalena, p. 358.

16. Astyanax. Magdalena to the United States, Trinidad and La
Plata and Peru, p. 354.

17. Hemibrycon. Peru, Ecuador and Trinidad, p. 358.

18. Cheirodon. Magdalena south to San Francisco, La Plata and Chili,
Pp. 358.

19. Gasteropelecus.' Pacific slope of Panama, Essequibo, Amazons and
southeast Brazil, p. 357.

20. Brycon. Magdalena to Guatemala and La Plata, San Francisco,
Ambyiacu, p. 354.

21. Creagrutus. Magdalena to Rio das Velhas. Few species widely
separated, an old genus?, p. 359.

22. Eigenmannia. Magdalena to Rio San Francisco and La Plata,
Para to Pebas, Parana, p. 355.

23. Sternopygus. Magdalena to Paraguay and Rio Canelos, San
Francisco, p. 357.

24. Synbranchus. Magdalena to Rio Motagua, Trinidad, La Plata, p. 354.

25. 4tguidens. Magdalena to La Plata, p. 355.

All of these show their ability for adaptation and colonization by their
wide distribution. The probable route of transit is up the Atrato and
down the San Juan, or up the Chagres and down the Pacific side of Panama.

Conspicuous exceptions are the genera Chezvodon, Hemiancistrus and
the Ce/opsine@, a group which, until the present, has been considered a
single genus. I am entirely at a loss to account for the presence of
Cheirodon in Chili and no interoceanic channel in the region of the
Amazon would help us to account for its presence in Chili or in Patagonia.’

‘Inclusive of the genus Thoracocharax.

2 The Cetopsing are found on the Atlantic slope throughout the length of the Amazons to the
Huallaga in Peru and up the Napo to the mountain streams about Canelos, Ecuador, and
are also found directly across the Andes from Canelos in the Guayaquil. The inference is
obvious. Paracetopsis has succeeded in getting across the Andes of Ecuador at a sufficiently

long time ago to enable the Guayaquil form to become generically and specifically distinct.
Hemiancistrus has a similar distribution, being also found on the western slope of Panama.

362 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Of the genera peculiar to the western slope Proftéstus and Gastropterus
are of marine origin not related to Atlantic slope forms. They are found
at high altitudes and are remnants of early colonists from the sea. Ledz-
asina is a genus of Characide very closely related to Prabucina of wide
distribution on the Atlantic slope. Saccodon is related to Paradon, dis-
tributed from the La Plata to Colombia and to HYemzodus from the Orinoco
to the San Francisco, the Peruvian Amazon and the upper courses of the
Paraguay. From one or the other it is probably an offshoot, sufficiently
remote in time to be generically different. All of the genera are very
probably among the oldest on the continent.

It is seen that the present distribution of the fresh-water fishes does not
require an interoceanic waterway for its explanation. Does the present
distribution of fresh-water fishes offer any objection to such an inter-
oceanic connection in recent times? Two or three things are opposed to
this supposition. One is the paucity of the Pacific fauna in general, and
the second the absence of many Atlantic slope families, especially the
Lepidosivemda, Aspredinide, Hypophthalmida, Callichthyide, Electro-
phoride, Osteoglossida, Avrapamide and many dominant Atlantic slope
genera, all of them lowland forms, and the third the absence of all
tropical American forms, except Pygzadtum from Lake Titicaca.

If the reigning fauna had been present on the eastern slopes of the Andes
even in outline at a time of the interoceanic canal, a very much larger per
cent. of the Atlantic slope forms ought to be found on the Pacific slope,
unless the Pacific slope streams were uninhabitable, or, unless the fauna be-
came extinct, after the separation of the oceans and the streams were
repopulated by new emigrants from the east.

The meagreness of the Pacific slope fauna is probably in great part due
to the small rivers and precarious water supply. It can scarcely be sup-
posed that a fauna became extinct to be replaced by a similar one. It is
quite certain that its present fauna has been derived from forms whose
wide distribution shows them to be especially fit to cross existing barriers
and which were probably the very first to reach the Atlantic slope of the
Andes. All of them are old genera, among the oldest on the continent.

If the present fauna had been in existence along the Andes at the time
Lake Titicaca was much lower, it ought to have received a fair sprinkling
at least of the fresh-water fauna, but it contains little but marine types.

All of these considerations lead me to conclude that the present fauna

EIGENMANN : FRESH WATER FISHES. 363

migrated to the Pacific slope subsequent to the formation of the Andes and
after the obliteration of the interoceanic sea, if such existed in the place sup-
posed, and long after the colonization of the coast streams of Minas, etc.

If the supposition be established, that few, if any, ancient Atlantic slope
forms reached the Andes in time to cross the divide by going from stream
to stream along oceanic channels, it indicates very strongly that the
fresh-water species were unable to bridge the gap between the ancient land
masses, Archiplata and Archiguiana and the Andes. Such inability to
cross even moderate seas would be strong evidence that they could not
have crossed the Atlantic, and that there must have been continuous land,
or land at frequent intervals or better still, a land wave carrying the fami-
lies common to Africa and South America from a center to both these
continents.

The fact that the Pacific slope fauna is generically all but identical with
the Atlantic slope fauna, and that the latter is generically different from the
African fauna, makes it certain that the Pacific slope fauna was derived
from the Atlantic slope fauna long after Guiana was separated from Africa.

VIII. THE NECESSITY AND EVIDENCE OF A FORMER LAND
CONNECTION BETWEEN AFRICA AND SOUTH AMERICA.

Not a single species of South American fresh-water fishes is found in
North America, nor do any North American species reach South America.
Two prominent South American families, the Chavacide and Cichhde
have representatives as far north as the Rio Grande basin, and one of these
has succeeded in crossing over into Cuba, evidently from Yucatan; on the
other hand, several members of the North American fauna have repre-
sentatives as far south as the Isthmus of Tehuantepec. The North
American fauna is entirely distinct from the tropical American fauna.

But four genera of fresh-water fishes of South America north of Pata-
gonia are found in any other continent than North America. These are
Synbranchus, Agonostomus, Cotylopus and Fundulus. The first is found
also in brackish water, the second belongs to the marine family Mugziide
and the other to the Pectlide. Synbranchus (map, p. 354) is found
in India, Agonostomus in Middle America, West Indies, northern South
America and New Zealand, Australia, Celebes, Mauritius and Comoro
Islands. There is no reason why 4gonostomus may not have been inde-
pendently evolved in the South Sea and in America from marine mugilids.

364 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Cotylopus is found in Central America and Reunion Island, Africa,
Fundulus in America and Europe.

It is possible that Prmelodus is found in Africa, and Pseadauchenipterus
in Madagascar. Both are found in South America.

Tropical Africa and Tropical South America have two groups of families

GALAXIIDAe

Range of the Galaxiide.

To ILLUSTRATE THE ORIGIN OF THE SOUTH AMERICAN FRESH-WATER FISH-FAUNA.

EIGENMANN: FRESH WATER FISHES. 365

in common. The first group comprises the Servanida, Scienida, Mugil-
zde and Tetraodontde.

These are all marine families, some of which have also developed fresh-
water forms in Europe and North America as well as in South America.
The fresh-water forms of South America and Africa are local adaptations

Range of the Synbranchide.

To ILLUSTRATE THE ORIGIN OF THE SouTH AMERICAN FRESH-WATER FISH-FAUNA.

366 PATAGONIAN EXPEDITIONS: ZOOLOGY.

of marine families that require no change in the present conditions to
account for their origin.

The second group comprises the Lepzdosirenide, Osteoglosside, Silur-
tde, Charvacide, Pecilide and Cichhde.

Of these, the Lepidosirenide are relicts of a formerly widely distributed
group, and it requires no land connection satisfactorily to account for

OSTEOGLOSSIDA.

Range of the Osteoglosside.

To ILLUSTRATE THE ORIGIN OF THE SOUTH AMERICAN FRESH-WATER FISH-FAUNA.

EIGENMANN: FRESH WATER FISHES. 367

their presence in Africa and South America. The Pecilide live indif-
ferently in marine, brackish water or fresh water. They reach their
maximum development in the fresh waters of Mexico, the West Indies
and Central America. The marine species are found along the shores —
not at sea—and there is, therefore, at present, no known means of getting

CICHLID-Ac

Range of the Cichlide.

To ILLUSTRATE THE ORIGIN OF THE SouTH AMERICAN FRESH-WATER FISH-FAUNA.

368 PATAGONIAN EXPEDITIONS: ZOOLOGY.

PG:GiEIIDF

Range of the Peeciliide.

To ILLUSTRATE THE ORIGIN OF THE SOUTH AMERICAN FRESH-WATER FISH-FAUNA.

them from the American to the African shore. Nevertheless, Fundu-
/us is found on both sides of the Atlantic, and there must have been an
intermigration much more recent than the youngest possible land con-
nection between Africa and South America, or else there has been a very
long persistence of this genus. A land connection, while not absolutely
required for this family, would be very convenient.

The Sz/uvid@ are in part marine. All of the South American forms of
Stluride can be derived from the marine 7achisuvine and the same is
probably true of the African members of the family. Furthermore, the
catfishes are found in North America, Europe and Asia, and have been
recorded in North America also from the Tertiary. A land connection
between Africa and South America is, therefore, not absolutely required
to account for their presence in both continent$, though, as in the case of
the Peciliide such a connection would be very convenient.

The Cechiide and Characide are abundant in tropical America and in
Africa, a few species of Czch@ide being also found in India. There is no
known means by which these two forms could have crossed the existing
gap between Africa and South America. There has been no exchange
of species in recent times, for there is no species or genus common to the

Sit

EIGENMANN: FRESH WATER FISHES. 369

two continents. The South American and African elements of these two
families must have been derived from some intermediate land-mass or
must have gone from one continent to the other over a land bridge.
That this connection, whatever it was, must have been obliterated before
the Tertiary, is evidenced by the facts that the Tertiary deposits of
Taubaté and Parana show existing genera and that there are many South
American types, as the Gymmnotide, Electrophoride, Bunocephatde,
Loricaride, Argude, Pygidide, Calichthyide, Hypophthalmide, et al.,
not found in Africa, that have all arisen in South America from the
Characide and Stluride since the separation of the two continents.

Similarly, other families found in Africa and not in South America
have either arisen in Africa since that time or have immigrated from the
Fast.

A land connection, whether a land bridge, intermediate continent or
land wave between the two continents, is:imperative. This land connec-
tion must have existed before the origin of existing genera and before
many of the existing families.

On the other hand, Boulenger (Les Poissons au Bassin du Congo,
VIII) explains the similarity of the South American and African faunas
as ‘without doubt the result of the persistence in these two parts of the
world of types more generally distributed at a very remote epoch which
have disappeared in other regions, as palaeontology shows us, among others,
to be the case with the Dipneustes, which belong to this category.

‘Nothing is gained by explaining these similarities by interpolating a
hypothetical continental continuity which could not have existed except
at a time previous to the development of the groups of teleostean fishes
which Africa and South America possess in common.”

IX. THe PoINnts OF ORIGIN’AND LINES OF DISPERSION OF TROPICAL
AMERICAN FRESH-WATER FISHES.

The study of the distribution of fresh-water fishes has led to these .
general conclusions.'

The origin and distribution of the fresh-water fishes of tropical South
America has come about as follows :

‘It has been sufficiently emphasized in the first section of this report that the fauna of “ Archi-

plata” or Patagonia is not related to or contains but a few intrusives from the Tropical American
Fauna.

370 PATAGONIAN EXPEDITIONS: ZOOLOGY.

In the earliest Tertiary tropical America consisted of two land areas,
Archiguiana and Archamazona, separated by the lower valley of the
Amazon which was still submerged.

There was a land mass between Africa and South America— possibly
in contact with Guiana in South America and some point in tropical
Africa. This land mass was inhabited, among others, by Lefpzdo-
sivenide, Peciliide, Characide, Cichld@e and Stluride. This land mass
sank beneath the surface of the ocean, forcing the fauna in two directions,
towards Africa and towards South America, exterminating all types not
moved to the east or the west. From these two rudiments have developed
the present diverse faunas of Africa and South America, each reénforced
by intrusives from the ocean and by autochthonous development, by immi-
grants from Africa, also neighboring land areas. The one fauna cannot
be said to have been derived directly from the other.

The connection between Africa and South America existed before the
origin of present genera and even before the origin of some of the present
subfamilies and families, some time before the earlier Tertiary. There
has never been any exchange between Africa and South America since
that time. There must have been an intimate connection between these
two continents, for there is no evidence, such as identical species or genera
on the two coasts, to indicate an occasional or accidental exchange of
types across the Atlantic since the formation of existing genera, there-
fore, such an interchange across the ocean probably never took place.
The East Brazilian land mass south of the Amazon (Archamazona) must
have become stocked from the western end of Helenis (Archiguiana)
very early, for it contains many genera peculiar to the region, indicating
a long separation, and Tertiary fresh-water deposits in this area contain
existing genera of fresh-water fishes. The Pacific slope fauna is a more
recent acquisition from the east.

When, later, the Cordilleras arose out of the ocean at a distance from
Archiguiana and Archamazona too great to be traversed by colonists from
them, their developing streams and arms of the sea, connected with
brackish, and later fresh-water lakes, all became populated with marine
types from the surrounding sea. In the north where they came in com-
petition with immigrants from Archiguiana, most of them were exter-
minated with the continued elevation of the land. On the south which
was not or was not so early reached by immigrants, Ovestias, Gastropterus,

Apion teri

OOkxAaw
NasSHLNnoS

Vi
nae

60/442:
Be Ilg/eA

eopey eT

yoyeved

yu
pobu:

lo gusinbo>
Duss 07

Lf eaves
A osreey 3p aso

ws Opje>

fOMEL
4 ocodey
Sieet sade
wn eebofopy

wfi9o9

h
senewew's Fr 2erg|

peperp tn wor ro
ey oo

6

“uy vghod
pues

osu
(sogisso>r

HM) 2Y0 + 89111 SoI*s[>C~

VOlANY
STICdING av HINOS
10 “av Seva

2979 9p Seiyuegy

on

ng

BT, thtinaett
MS abe

ode

ON ht ately
ey

’
EIGENMANN: FRESH WATER FISHES. 371

and Protistius remain in the high Cordilleras of southern Peru as relicts
of these marine species. Later, these mountain streams, especially those
of Ecuador and Colombia, became populated by stragglers or accidental
visitors from the land areas to the east. These in their turn, with the ele-
vation of the Andes, became modified and gave rise to the genera now
peculiar to both slopes of the high Andes, Pygediam, Evremophilus, Che-
tostomus, Arges, Cyclopium, Astroblepus, etc.

With the further elevation of the Cordilleras into a continuous barrier
and the formation of the Orinoco, Amazon and La Plata valleys through
elevation and the debris brought from the land masses, and the development
of the enormous fresh-water system occupying these valleys, this system,
particularly the Amazon, became colonized from the older land areas and
became the center of unparalleled adaptive radiation and a new center
for distribution, which it has remained to the present time.t The com-
paratively few types inhabiting the old eastern land masses found them-
selves in possession of a continent and diverged along every conceivable
direction, the characins alone giving rise to over 500 species and over
100 genera.

From the Amazon species moved in all directions till they met barriers
of one sort or another. The Pacific slope fauna is derived to a very large
extent from this latter divergent migration over the isthmus of Panama
and through the Atrato valley between the western and coast Cordilleras
of Coiombia. Others possibly crossed over the Andes east of Guayaquil
before the Andes reached their present height. The Pacific slope fauna is
less different from the Amazon fauna than that of the coastwise streams of
Minas, if the number of peculiar genera is used as a measure of difference.
Amazonian types moved south till climate and barriers checked them
south of Buenos Aires. They migrated northward till they came in com-
petition with emigrants from the north in the lowlands of Mexico.

The origin of the fauna of the plateau of Mexico is a separate subject.
This fauna is in part of marine origin and antedates, in a large measure,
the mongrel fauna of the Mexican lowlands.

* Amazonian species form six or seven per cent. of the fauna of the Motagua, five of the Pacific
slope fauna, over 30 per cent. of the Magdalena fauna, over 42 per cent. of the Trinidad fauna, over
50 per cent. of the fauna of the Guianas, over 40 per cent. of that of the San Francisco, about 30
per cent. of the fauna of the coastwise streams east of the San Francisco, and over 50 per cent. of
that of the Paraguay.

372 PATAGONIAN EXPEDITIONS: ZOOLOGY.

Evidence for these conclusions has been given in detail in the foregoing
pages. The basis for the entire discussion is a list of the species giving
their geographical distribution. Such a list has been prepared and follows
this discussion.

The points of strategic importance for ichthyic chorology in South
America are, therefore, (@) western Colombia and Panama; (6) Guayaquil
and Peru to the Amazon, across the Andes; (c) the tableland of Guiana,
Archiguiana ; (@) the Rio San Francisco, with the Rio Parahyba and the
headwaters of the Tieté and Rio Grande, in Archamazona, and (e) the
area between the Rio Negro and the La Plata.

Note ON LAKE TITICACA AND ITS FAUNA.

In conversation with Prof. G. Steinmann the latter called attention to
his views concerning Lake Titicaca, which he had the kindness to write
out for me.

“The region at present covered by Lake Titicaca was formerly, as late
as late Tertiary times, a normally drained area. Its drainage was south-
east toward the Amazon. It was not till the Glacial epoch that the glaciers
of the high Andes pushed their moraines into the drainage valley and
formed the lake. It is, therefore, a glacial-dam lake which did not retain
its Amazonian drainage, but flowed over the low watershed southeast to-
ward the undrained high plateau of Bolivia. There is no evidence that
the lake was formerly connected with the Pacific. Not only are there no
marine formations in the inter-Andean high plateaus of northern Bolivia
and southern Peru belonging to the Diluvial time, but also those of Ter-
tiary times are lacking. For this reason also the fauna of Titicaca cannot
be explained as a relict, but must have arisen from the ocean by migration
in a roundabout way through former rivers and lakes.”

In my account of Lake Titicaca it was assumed that the origin of the
lake from an arm of the sea was without question, and an attempt
was made to explain its fauna on that basis. It was assumed that the
genus Orestias gave rise to numerous species, some of which succeeded
in crossing the divide into neighboring rivers. The explanation of Stein-
mann would obviate the difficulty of originating numerous species from
one type in a restricted unit environment. General observation every-
where gives evidence that segregated individuals of a given species tend
to diverge from the central type, not that a species mutates into a large or

EIGENMANN: FRESH WATER FISHES. 373

small number of species in a restricted locality. It seems as reasonable
to assume that Orestias migrated from the ocean into all streams when
the Andes were low, became different in these segregated environments,
and filtered from them into Lake Titicaca, as to assume that it became
landlocked in an arm of the sea when that was cut off by elevation, gave
rise to numerous species in the one locality, and migrated from the lake to
the headwaters of neighboring streams.

THE FAUNA OF THE RIO REBEIRO.

During the years 1907 and 1908 a series of papers on the fishes of the
Rebeiro by Rebeira and the Cubatao by Steindachner have shown that
these isolated rivers have a fauna, nearly half of which is peculiar. They
are streams in the northern part of the state Sta Catharina emptying into
the Atlantic. The material came too late to incorporate in the proper
table without entirely resetting it. A list of the fishes from these rivers
is, therefore, given here. This peculiar fauna, situated between Rio
Grande do Sul of the La Plata province and the Parahyba of eastern
Brazil, demonstrates very well that the streams emptying into the Atlantic
do not form a highway for ready migration from the North or South.

Peculiar species are marked*, peculiar genera f.

* Hemipsilichthys calmont Steindachner. Cubatao.
+ Kronichthys subteres Rebeira.
Parotocinclus maculicauda Steindachner.
+ Otocinclus leucofrenatus Rebeira.
+ Ofocinclus gibbosus Rebeira.

Loricaria lattrostris Boulenger.

* Loricaria henseli Steindachner. Cubatao.

Loricaria ima Kner.

{+ Loricaria cubataonis Steindachner. Cubatao.

Loricaria anus Valenciennes.

{+ Xenocara brevispinnis Regan. Cubatao.

Ancistrus sigmaticus Eigenmann & Eigenmann.
| Hartha kronet Rebeira.

Plecostomus commersoni Valenciennes.

+ Plecostomus obtusirostvis Steindachner. Cubatao.

Trichomycterus dispar Tschudi.

Trichomycterus brasthensis Liitken.

374 PATAGONIAN EXPEDITIONS: ZOOLOGY.

* Trichomycterus proops Rebeira.
Glanidium albescens Reinhardt.
Pseudopimelodus zungaro (Humboldt).
Pimelodus clarias Linnzus.
* Rhamdella ignobilis Steindachner. Cubatao.
Rhamdadia sebe Cuvier & Valenciennes.
* Rhamdioglanis transfasciatus Rebeira.
+ Zyphlobagrus kronet Rebeira.
Fleptapterus mustelinus (Valenciennes).
* Hoplias lacerde Rebeira.
Curimatus gilberti Quoy & Gaimard.
Characidium fasciatum Reinhardt.

_ Astyanax scabripinnis (Jenyns).

? A. microcephalus Rebeira.

* Astyanax scabripinnts longirostris Steindachner. Cubatao.

A styanax fasciatus (Cuvier).
* Astyanax eigenmanniorum depressivostris Rebeira.
| Deuterodon iguape Eigenmann & Norris.
{+ Deuterodon rose (Steindachner).
| Hollandichthys multifasciatus (Eigenmann & Norris).

Pseudochalceus perstriatus Rebeira. Pseudochalceus affinis Steindachner.

{+ Celurichthys tporange Rebeira.
Acestrorhynchus hepsetus (Cuvier).
jJanuarius Hensel. Cubatao.
Pectla vivipara Bloch & Schneider.

Geophagus brasiliensis Quoy & Gaimard. Rebeira & Cubatao.

Cichlasoma facetus Jenyns.
Cichlasoma autocthon (Giinther).
Crenicichla lacustris Castelnau.

7 . . -
s : — q
> : :

ae
alt
ie
t

f ‘

PATAGONIAN EXPEDITIONS: ZOOLOGY.

EXPLANATION OF PLATE XXX:

Fig. 1. GEOTRIA CHILENSIS Gray. From a specimen in the British Museum,
natural size.

Fig. 1a. Mouth of the same, X 2.

Fig. 2. Greorria sTENosToMuS Ogilby. After Plate.

Fig. 3 and 3a. GEoTRIA AUSTRALIS Gray. After Plate.

Fig. 4 and 4a. EXOMEGAS MACROSTOMUS GALLEGENSIS Smith. Figure by Oester-
berg after Smith.

Fig. 5. CARAGOLA MorDAX Richardson. After Plate.

Fig. 6. Caracoia LapicipA Gray. After Plate.

Fig. 7. CARAGOLA ACUTIDENS Philippi. After Plate.

(VOL. 111)

D fan an T aor \) TARKTAa VI CVT [= ATE avavs
PATAGGNIAN EXPEDITIONS VOLAJII FLATE XXX

Cc

———<

Ta

| PY arty
| | j

(f. {{ Lf fei sil
bebedobedabiebdtideldobobehad
1 Metab

GEOTRIA, EXOMEGAS & CARAGOLA

PATAGONIAN EXPEDITIONS: ZOOLOGY.

EXPLANATION OF PLATE XXXI.

Figs. 1, 1a, 16. DipLomystTE papittosus C. & V., No. 8290 in the Mus. Comp.
Zool. Cambridge, Mass., from Santiago, Chili.

Fig. 2. NEMATOGENYS INERMIS (Guichenot). No. 9839 Mus. Comp. Zool.
Curico, Santiago, Chili.

(VoL. 111)

Werner & Winter, Frankfort °M., lith

DIPLOMYSTE & NEMATOGENYS

PATAGONIAN EXPEDITIONS: ZOOLOGY.

EXPLANATION OF PLATE XXXII.

Figs. 1, 1a, 16. HatcHerta MACR&I (Girard). No. 8298 Mus. Comp. Zool.
Uspullatuo, Chili.
Fig. 2, NEMATOGENYS INERMIS (Guichenot). See plate XXXI, fig. 2

(VoL. 111)

Yn

ATCHERIA & NEMATOGENY

ag;

»
*
Ey.
7
.
‘
6 -
e
—
t
cttice
vu ' Le,

ue A Pie }
aes , 2 \ }
Pie] i, +)

«

|

<74)™ \ . } ‘
: vi Nae fee |

A ‘i —
+ Ce dy fal “eo

nt }
ay ieP Oy

ON 1 Ge aa.)
net et Xf 2 ‘a
fu Ret ie, ary? 7

ahs A

Tide, Cen aA oo sigh a ene
RNS. WRT AN Abia i

Dy pe

y

=

’ cy a te YN dad

2

PATAGONIAN EXPEDITIONS: ZOOLOGY.

EXPLANATION.OFP PLATE XXXII.

Fig. 1, 1a, 16. HatcHERIA MacuLaTA (Cuvier & Valenciennes). No. 7736
Mus. Comp. Zool., Mapocho, Chili. Total length 92 mm.
Fig. 2. HATCHERIA PATAGONIENSIS Eigenmann. Rio Blanco.

(VoL. U1)

ees.

2

PATAGONIAN EXPEDITIONS: ZOOLOGY.

EXPLANATION OF PLATE XXXIV.

Fig. 1. HATCHERIA PATAGONIENSIS Eigenmann, See plate XXXIII, fig. 2.
Fig. 2. HarcHEerta AREOLATA (C. & V.). Mus. Comp. Zool.
Fig. 3. GYMNOCHARACINUS BERGII Steindachner. After Steindachner, X 3.

Fig. 3a. Dentition of the same.
Fig. 4. AsTYANAX RUTILUS (Jenyns). Drawing by J. Green from a specimen from
the Rio Negro in the British Museum.

(VOL. 111)

Il.

VOL

BA

AN EXPEDITIONS

ATAGONI

=
=
iE

i
:
:
ot
'
ie
ie
{

\

I
|
a
W
4

- ond id
Le neo? A Sak a
: . * —
ins Lif 7 =
Ty oe | * ¥
3 — 2
: i Wei
ot a -
. we un hee
oe: zs i —, &
aa i aa
vee ek. ' , -,
i a bd
. Uy ¥ <s
: a i - : : =
at, |e a ap aay)

i 9 (Rowan

ne ie ated A
JOUR PRRLE RNIN EN
; b od ae :

ie) 3
2 A 4

me : / ~~

fh

PATAGONIAN EXPEDITIONS: ZOOLOGY

EXPLANATION OF PLATE XXXV.

Fig. 1. Side view of a young specimen of Ga/axias alpinus (Jenyns).

Figs. 2, 2a. A larger specimen of the same.

Figs. 3, 3a. Another specimen with different color pattern.

Fig. 4. GALAxtas MACULATUS (Jenyns). From a specimen in the British Museum.
Fig. 5, 5@. APLOCHITON zEBRA Jenyns. After Smith.

Fig. 6, 6a. ApLocuiron T&NraTus Jenyns. After Smith.

(VoL. 111)

ee

PATAGONIAN EXPEDITIONS VOL.ML. PLATE XXXV

Ga
SOSSS

’ i ;

ale 1
¥ oe LaF
a abe Slaw!

Lega hiye ts -s
Ey Ce :
es

; os
ess
a ld us is i

oa
=<

EXPLANATION OF PLATE

Fig. 1. Perciita Gituisst Girard. From a specimen in
Fig. 2, 2a, 26. PeRcICHTHYS MELANOpsS Girard. No. 4835
Curico, Chili.
Fig. 3. Percicutuys Trucna (C. & V.).

(VoL. 111)

4. ‘ s
PATAGONIAN EXPEDITIONS VOL.1II
= id

A

PLATE XXXVI.

Ky
eae
Oe

Werner & Winter, Frankfort ®M. {ith

PERCILIA & PERCICHTHYS

ae
4405

>

4 “*

| PATAGONIAN EXPEDITIONS :

EXPLANATION OF PLATE XXXVII.

Pig; X. PERCICHTHYS ALTIPENNIS (Regan). After Regan.

Fig. 3. PErct
Fig. 4. Meni
ital

omy,
®
%
; a
i som S -
.s +e a
m Ss
¢ bsg
“ < <2
* @ om Tea
5 oe, a ‘cite $4.
“ se ; '
| ae HN bled ae
as : oe a yes +. “48
" ape
=r Me $ RS mb 4 aye ;

Fig. 2. PercicuTHys TRucHA C, & V.
YS VINCIGNERR& (Perugia).
TCHERI Eigenmann.

(VoL. 111)

*

—

) AMD TT T 7 ur _ TTIARATA WE, B A \ :
PATAGONIAN EXPEDITIONS VOLII PRAT EOORV iI

JK mK

Oh
OL

"Poa AEM

CAC ea

bli AoW MY s CRU ee
\ ‘YS

Mp MOCO
Met ein (aS

Oe

RINK eT
¥ YN) ‘ a) ee =)

\/
PERN KORE ROY ISO ARRAS OY