. 10

MEMOIRS

CABNEGIE MUSEUM.

VOL. II. NO. 10.

THE CRAWFISHES OF THE STATE OF PENNSYLVANIA.
By Arnold E. Ortmann, Ph.D.

I. Introduction.

The present .Memoir is a continuation of. and an enlargement upon, the prelimi-
nary paper published some time ago in the Annals of the Carnegie Museum (Vol.
Ill, 1905, p. 387 et seq.) under the title "The Crawfishes of Western Pennsyl-
vania," The object of these publications is to furnish the student with an account
of the crawfish-fauna of the state of Pennsylvania as complete as possible, not only
from the morphological and zoogeographies!, but also from the biological, ecolog-
ical, and economic standpoint. It is now believed that it is possible to present an
approximately complete report upon this important branch of the fresh-water fauna
of the state, and in the prosecution of the studies of the author a number of ques-
tions were raised, the solution of which proved to lie highly interesting.

It may be well at the beginning to give an outline of the work done. At the
outset the writer resolved to go over the wbole state, and to collect specimens in as
many different localities as possible. Very soon, however, it was discovered that
the different parts of the state are of unequal interest, Large tracts, located chiefly
in the central, northern, and northeastern parts of the state, proved to lie rather
uninteresting, only one species of crawfish being present in them, while the western,
and chiefly the southwestern, and again the southeastern sections offered more
variety. Thus it became necessary to pay more attention to the latter areas. The
uninteresting regions were entered only in a few cases, but a good deal of work was
done around their edges, in order to trace their limits as accurately as possible.

The location of the writer in Pittsburgh was advantageous, being central within
that section of the state which offered the greatesl number of problems. Mosl of

343

344 MEMOIRS OF THE CARNEGIE MUSEUM

the collecting excursions were undertaken with Pittsburgh as a base. However, on
three occasions the base was shifted. Visits were twice made to the eastern part of
the state, where the writer spent several weeks in September of the years 1904 and
1905 in Philadelphia and its environs, and once to the eastern central part, where
several days were spent in Harrisburg in June, 1905. The latter visit was marred
by rainy weather.

The work of collecting was done for the Carnegie Museum by the writer in con-
nection with his duties as Curator of Invertebrate Zoology, and all the necessar}^
expenses were paid by the Museum. In order to give an idea of the amount of
field-work done, a few statistics may be interesting.

Altogether one hundred and thirty-eight days were spent in the field, counting
only those days on which actual collecting was done : four days in 1903 ; sixty in
1904 ; and seventy-four in 1905. A few additional records were obtained in 1906.

The distances covered in travelling were as follows :

Total.
By rail, in 1904 3238 miles.

" " " 1905 7579 " 10,817 miles

By team, in 1904 12 miles.

" " " 1905 26 " 38 miles.

On foot, in 1903 3 miles.

" " " 1904 173 "

" " " 1905 209 " 385 miles.

Grand Total 11,240 miles.

Collections were made at about one hundred and fifty-six different localities,
most of them in the state of Pennsylvania. Of the sixty-seven counties of the state
thirty-nine were visited.1 Besides, visits were made to a number of localities situ-
ated in neighboring states, namely : in Camden County, New Jersey ; in Allegheny
and Garrett Counties, Maryland; in Morgan, Mineral, Tucker, Preston, Monon-
galia, Pleasants, Wetzel, Marshall, Ohio, Brooke, and Hancock Counties, West Vir-
ginia ; and in Harrison, Carroll, and Stark Counties, Ohio.

The material secured on these excursions belongs to and has been deposited in
the collections of the Carnegie Museum, and comprises 303 entries in the Cata-
log, including 1869 specimens. But this does not represent the entire number of
specimens collected, since large sets, which have not been cataloged, have been set
aside as material for exchange, study, etc.

1 Material was secured, seen, or was previously known from fifteen additional counties, so that only thirteen are
not explored, namely : Carbon, Juniata, Lackawanna, Lebanon, Mifflin, Monroe, Montour, Pike, Schuylkill, Snyder,
Susquehanna, Union, and Wyoming. All these belong to the central and northeastern section of the state, where only
one species of Cambarus ( C. bartoni) is to be expected, with the exception of those localities which are in the immediate
vicinity of the main branches of the Susquehanna River, where also C. Umosus may be present (Juniata, Montour,
Snyder, and Union Counties).

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 345

Practical experience gradually revealed to the! writer the best method of col-
lecting crawfishes. At first the writer was rather inexperienced, and did not know
where to look for certain species. But the necessary knowledge and skill were soon

acquired.

To collect the species living in streams, rivers, and ponds requires no special
effort; it is only necessary to wade into the water-course to be investigated, or to
walk along its edge, and to discover the hiding-places of the crawfishes, which is
done by turning over stones. A pair of rubber-boots, or wading-stockings, pro-
tected by ordinary bathing-shoes, is very convenient; and also a small landing-net,
the hag made of minnow-netting. Frequently specimens may be caught with the
hand. In certain places, and in the case of certain species ( Cambarus limosus) water-
weeds often furnish hiding-places, and here it is advisable to use a larger landing-
net with a long handle, which is pulled or pushed through the weeds.

More work is required in collecting the burrowing species. The first thing is to
locate them, which is generally done by searching for the mud-chimneys built over
their burrows. But it is not always easy to find these, particularly in late summer
and autumn, the chimneys being then rather inconspicuous. I generally first ascer-
tain favorable localities, such as swampy places in the bottom-lands, and springy
places on the hillsides. It is a very good plan to closely watch ditches by the road-
sides. Here the chimneys generally are easily detected, and in the neighborhood
of such places large colonies often may be found. After a burrowing form has been
located, the most difficult work begins, for the inhabitant of the burrow must he
dug out of its hiding-place. Care must he taken while digging not to injure
the crawfish. Sometimes the work can be done with the bare hand, hut only in
those rare instances in which the burrows are in very soft ground. Generally the
work must be done with pick and shovel, but, as it is inconvenient to carry these
cumbersome tools along on an excursion and a gardener's trowel is a little too weak,
I have found a pioneer's bayonet, such as is used in the United States Army, to be
a tool which beautifully serves the purpose. These bayonets may be had in gun-
and ammunition-stores in the larger cities, and are practically indestructible.

After locating a crawfish-hole, I begin to dig down around it, loosening the
dirt with the bayonet, and removing it with my hands, always taking care not to
lose trace of the hole. Generally it is necessary to'go down upon the knees (rubber
boots are useful here), and even the belly, in order to reach the bottom of the hole,
to which the crawfish usually retreats when disturbed. Often, however, it retires
to a side branch, in which case it is not necessary to dig so deep. As soon as it is
felt beginning to pinch with its claws, it is a sure sign that the crawfish has been

346 MEMOIRS OF THE CARNEGIE MUSEUM

cornered, and cannot retreat further. It is then readily secured, but care must be
taken not to pull it out by the claws, which may be easily broken off, thus damag-
ing the specimen. The creature should be always seized by the carapace.

Sometimes this work is very difficult and tedious, and I have often been com-
pelled, chiefly in the case of Cambarus earolinus, to dig as deep as three feet before
succeeding in capturing the crawfish. In order to avoid unnecessary labor as far
as possible, I select burrows in which the water stands near the surface, refusing
those which evidently go for a long distance through dry soil. Generally there is
ample opportunity to choose between the numerous burrows of one and the same
colony of chimney-builders. Now and then it happens that the work is rendered
easier by the action of the crawfish itself. It occasionally comes to pass that, after
having disturbed the entrance of the hole by digging down far enough to reach the
water, the crawfish may be seen coming forward, apparently trying to ascertain the
the cause of the disturbance. This is a good chance to seize it, but one must be
quick, since it generally is the only chance to get it easily, although I remember
cases when the crawfish came out again and again, just so far as to be plainly seen,
but darted back at every attempt to seize it. Males are more frequently caught
in this wTay than females, and such captures are made most frequently in cloudy or
rainy weather. It is very rarely that there is a chance to capture the crawfish at
the mouth of the undisturbed hole, when it is sitting at or near the top of the
chimney, or on the ground away from the hole. This happened only once or twice
in my experience.

I have tried to discover means of compelling the crawfish to come out of its
burrow, but without much success. Bisulfide of carbon will not avail since it floats
upon water. I tried chloroform, which sinks in water, but without success. Only
once had I the satisfaction of driving a specimen of Cambarus monongalmsis out of
its hole by using unslacked lime. In this case I had dug a hole nearly three feet
deep without being able to reach the bottom. I happened to have with me, espec-
ially for this purpose, a small tin box with unslacked lime, and dropped the con-
tents into the hole, where it apparently sank to the bottom. Within three or four
minutes the crawfish was discovered hurriedly working its way upward in the hole,
and was easily taken. This method, however, can be used only in a limited way.
since the holes generally are not straight enough to afford a chance to drop the lime
to the bottom, and, if the lime becomes lodged somewhere above the point where
the crawfish is staying, it drives it away from the mouth of the hole, and eventually
kills it before it can be reached. (As to the use of lime for destroying crawfishes,
see infra, VI, 4.A

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 347

The tools and the outfit needed for collecting crawfishes consist of rubber-boots
(for work in swamps), or wading-stockings with low shoes (for work in streams),
bayonet (for digging up burrowing species), and a landing-net with minnow-netting.
These are indispensahle. Further, a number of collecting jars are needed, with 75
per cent, to 80 per cent, alcohol, and I prefer to carry them in an ordinary fishing-
basket, but any other portable receptacle will do. I do not recommend the use of
formaldehyde, since it makes the specimens too brittle. In order to take specimens
home alive, a so-called "bait-box" is most convenient.

I have no experience in baiting crawfish, and never attempted it, since the
methods described above proved satisfactory. Furthermore, I have never (in
Pennsylvania) used the seine, and I do not think that this would be necessary or
advisable in this part of the country, although it may be tried to advantage else-
where.

Besides the material secured by me in the manner above related, 1 made use of
the older material preserved in the Carnegie Museum, which was collected by the
following gentlemen: D. A. Atkinson, G. H. Clapp, E. Frost, B. Graf, J. L. Graf,
E. Hays, S. N. Rhoads, A. T. Shafer, Q. T. Shafer, H. H. Smith, R Taylor, W. E.
C. Todd, M. A. Wertheimer, and E. B. Williamson.

1 was aided in my researches by material kindly collected by various persons for
the Museum, while my work was in progress, and to all of them I wish to here
express my best thanks. They are the following members of the Museum staff:
Mrs. Elizabeth Courtney, Mr. C. V. Hartman, Mr. D. C. Hughes, Mr. 0. E. Jen-
nings, and Mrs. 0. E. Jennings.

The following gentlemen living in or near Pittsburgh furnished;material : Dr. D.
A. Atkinson, Dr. 0. T. Cruikshank, Mr. R. Dornberger, Mr. D. Friel, Mr. F. E.
Kelly, Dr. A. Koenig, Mr. A. Settlemoyer, and Mr. R. Settlemo}'er.

Material from other parts of Pennsylvania was received from Professor A. E.
Davison, Lafayette College, Easton, and Mr. H. Gera, Manaynnk.

Specimens of our Pennsylvanian species were received from localities outside of
the State from the following sources: Academy of Natural Sciences, Philadel-
phia, (material from Delaware, Maryland, and North Carolina, in exchange); Mi-.
H. Gera, Manaynnk, (material from New Jersey); the late Mr. J. B. Hatcher, Pitts-
burgh, (material from Iowa); Mr. 0. E. Jennings, Pittsburgh, (material from Ohio);
Mr. S. Prentice, Pittsburgh (material from Kansas); Dr. R. Ruedemann, Albany,
New York, (material from New York); Mr, F. Silvester, Princeton, New -Jersey,
(material from Maryland); Mr. E. B. Williamson, Bluffton, Indiana, (material from
Kentucky, Indiana, and Michigan).

348 MEMOIKS OF THE CARNEGIE MUSEUM

Very considerable help was received from the Department of Agriculture in
Harrisburg. The State Zoologist, Professor H. A. Surface, not only sent to me for
inspection all the crawfishes in the collection under his charge, but also submitted
to me material collected during the summer of 1905 by Mr. W. R. McConnell, who
was in charge of a survey conducted by the State Zoologist in cooperation with the
Commissioner of Fisheries, Mr. W. E. Meehan. To Mr. Meeban and Professor Sur-
face I am under special obligation for giving instructions to Mr. McConnell regard-
ing the collecting of crawfishes, and to the latter gentleman for carrying these out
in the most thorough way in parts of the state not visited by myself.

Finally, I was granted the privilege of examining the collections of the Academy
of Natural Sciences in Philadelphia, where I found, aside from older specimens
already used by Hagen and Faxon, valuable additional material, collected by
Messrs. H. A. Pilsbry, E. G. Vanatta, H. W. Fowler, and B. W. Griffiths. I also
received specimens for examination from Oberlin College, through the late Pro-
fessor A. A. Wright and Mr. R. L. Baird ; from the New York State Museum
through Mr. F. C. Paulmier; and from Dr. P. R. Uhler in Baltimore, and Professor
T. D. A. Cockerell, in Boulder, Colorado.

Last, but not least, my thanks are due to the Director of the Carnegie Museum,
Dr. W. J. Holland, who not only granted the means for carrying on my work suc-
cessfully, but has devoted much time to the editorial revision of the manuscript,
and helped me in the preparation of the colored plates accompanying this memoir,
which were made under his direction.

II. Historical Review of our Systematic Knowledge of the Crawfishes

of Pennsylvania.

The first species of the genus Cambarus ever described very likely came from
our state. Astacus barton i of Fabricius (1798, p. 407) was sent to its author by Pro-
fessor B. Smith Barton, who lived in Philadelphia, (see Faxon, 1885a, p. 65) and
presumably was collected in the neighborhood of that city.

The next record of a Pennsylvanian crawfish is given by Rafinesque (Nov., 1817),
Astacus limosus, from the muddy banks of the Delaware near Philadelphia. An-
other species mentioned by Rafinesque from this state, Astacus fossor, is not recog-
nizable. Astacus limosus from the Delaware River was described a month later
(Dec, 1817) by Say under the name of Astacus aff/nis.

Harlan (1835) mentions A. barton! from the vicinity of Philadelphia, and this
record makes Philadelphia the type-locality of this species.

Girard (1852) gives the following new localities in Pennsylvania: Cambarus

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA

349

affims = limosus, Schuylkill River at Reading, Bucks County; G. bartoni, Foxburg,
Clarion County; Carlisle, Cumberland County: Berwick, Columbia County.

In Hagen's Monograph (1870) no new species are added, although he doubtfully
records (p. 100) G. obscwrus from the state, hut this was done under the erroneous
assumption that Astacus fossor of Bafinesque is the same as C. obscurus. Thus
Hagen's Monograph adds only a few new localities, namely: for G. limosus (affinis)
the Schuylkill River, Philadelphia, and Carlisle, Cumberland County. The new
locality "Pittsburgh" for the same species is wrong.

The great revision of the genus published by Faxon (1885a) adds two species:
G. diogenes from Deny, Westmoreland County, and C. rustims from Pittsburg.
The latter record is incorrect, and should be dropped. Besides Faxon gives the fol-
lowing new localities: G. limosus, Brandy wine Creek, Chester County; Delaware
River, Bristol, Bucks County; Bainhridge, Lancaster County; C. bartoni, Bedford
and Pattonville, (see infra, foot-note 16), Bedford County; Windham. Bradford
County; Hummelstown, Dauphin County; Chester County ; Bainhridge, Lancaster
County; McKean County.

Thus only three species were known up to this date : C. limosus, C. bartoni, and

C. diogenes.

In 1898 Faxon added a fourth species for the state, C. obscurus from Westmore-
land County, and also gave a new locality for G. bartoni, Westmoreland County.

Williamson (1901) enumerated five species, and one variety from Allegheny
County, hut, as has been demonstrated by the writer (1905a), this is to be reduced
to four species and one variety, of which the species recorded as C. dubius by
Williamson (C. monongalensis Ortmann) is new for the state, as is also the variety C.
bartoni robustus. Allegheny County is for the first time cited as a locality for the
other species, C. obscurus (recorded as C. propinquus and rusUcus), C. bartoni, and G.
diogenes. Williamson's discoveries brought up the number of species known from
the state to five, and one variety, namely : C. limosus, C. obscurus, G. bartoni, < . bar-
toni robustus, C. monongalensis (as dubius), and G. diogenes.

In the preliminary report of the writer for western Pennsylvania, two other
species were added : I ',. propinquus from Erie and Crawford Counties, and G. mm/inns
from Fayette, Westmoreland, and Somerset Counties. Numerous new localities
were added to those already known. C. dubius of Williamson was recognized as a
species new to science, and described as C. monongalensis. The number of species
present in the state has not been increased by subsequent investigations, and stands
now as seven, with one variety, namely: Cambarus limosus (Rafinesque) ; C. pn>i>-
inquus Girard; C. obscurus Hagen ; C. bartoni (Fabricius) ; C. bartoni robustus (Girard);
C. carolinus Erichson ; C. monongah nsis Ortmann ; G. diogt nes Girard.

350 MEMOIRS OF THE CARNEGIE MUSEUM

These eight different forms are treated in the following pages, with the addition
of one extralimital form, a variety of G. propinquus (C. propinquus sanborni). The
recent and past observations made by the writer, together with those collated from
other sources, are presented as completely as possible in the succeeding pages, thus
giving a natural history of this group of animals, so far as found in the state of Penn-
sylvania. It has always been the aim of the writer to support his conclusions by
evidence secured within the state, but observations made outside of the state are
sometimes introduced, where a gap was to be filled, or where they were of special
interest.

III. Morphology and Chorologv of the Pf.nnsylvanian Species.
A. General Remarks.

The crawfishes (as to the use of " crawfish " in preference to " crayfish" see infra,
VI) of the state of Pennsylvania belong to the genus Cambarus2 of Erichson (1846),
of the family Potamobiidx Huxley, including the freshwater crawfishes of the north-
ern hemisphere. Faxon (1898) regards this as a subfamily, Astacime, of the family
Astaciche, which also includes the subfamily, Paraslacinse of the southern hemisphere,
now regarded as a family, ParastacicLx Huxley. There is some discussion as to the
proper name of the family, depending on the use of the generic name Astacus Fab-
ricius, or of Potamobius Samouelle, for the European crawfishes. The position of
the writer was defined in 1902, (Proc. Artier. Philos. Soc. XLI, p. 276, footnote).
The question, however, has recently been finally settled by a discovery made by
Miss M. J. Rathbun (Proc. Biol. Soc. Washington, XVII, 1901, p. 170), but not in the
manner suggested by Miss Rathbun. The fact that the name Homarus was used
first by Weber (Nomenclator Entomologicus, 1795), in place of Astacus Fabricius, 1775,
makes Homarus a pure and simple synonym of Astacus, and according to the rule
" once a synonym, always a synonym," it remains a synonym. There is no reason
to make it " desirable," as Miss Rathbun expresses it, to set aside the rule in this case,

The genus Cambarus, containing now about 70 species, has been variously sub-
divided : by Girard (1852) into three groups; by Hagen (1870) likewise into three
groups, which, however, do not exactly correspond to those of ( rirard ; and by Faxon
(1885'/) into five groups. Recently the present writer has divided the genus into
six subgenera, namely : Pardcambarus, Procambarus, Cambarus, Cambarellus, Faxo-
nius, Bartonius. (See Proc. Am. Phil. Soc, XLIV, 1905, p. 91, et seq.; Ann. Cam.
Mus. Ill, 1905. p. 437 ; and Proc. Washington Acad. Science, VIII, p. 1, 1906.)

2Erichsou made this provisionally a subgenus, but expressed the opinion that it perhaps would better rank as a
genus. Girard (18.V2) was the first to use Cambarus as a generic name.

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 351

(July the two last named subgenera, Faxonius and Bartonius, are found within the
boundaries of the state of Pennsylvania, and they arc distinguished from the rest by
the shape of the male copulatory organs. The latter consist of two parts, which are
completely separated at the tips for a shorter or longer distance, and never ]>ossiss
any accessory spines. The outer part in the male of the first form, (sexually ripe),
is almost completely transformed into a horny spine, while the inner remains soft.
In the Pennsylvanian species of these two subgenera only the third pereiopods possess
a hook (used to take hold of the female) on the ischiopodite.

The two subgenera are distinguished as follows :
Subgenus Faxonius Ortmann.

Sexual organs of first pair in the male with two shorter or longer completely
separated tips. Tips straight or gently curved, divergent, parallel, or convergent,
generally rather slender.
Subgenus Bartonius ( trtmann.

Sexual organs of the first pair in the male with two rather short, completely
separated tips. Both tips are strongly recurved, forming with the basal part about
a right angle.

B. Kc;i to the Pennsylvanian Species oftihe Genus Cambarus.

a'. Sexual organs of male of the Fnsuniiut-type. Rostrum always with a marginal spine on each side, and carapace
with one or more lateral spines. (River species. )
V '. Tips of sexual organs short, straight, and divergent. Sides of carapace with several lateral spines anterior to,
and behind the cervical groove. (Delaware, Susquehanna, and Potomac drainages.)

C. (Faxoniua) limoaua i Bafinesqne).
b" '. Tips of sexual organs long, almost straight, slightly convergent, or parallel. Sides of carapace with only one
spine behind the cervical groove.
c'. Rostrum with median keel. Sexual organs of male of first form at anterior margin without prominent
angle (shoulder). Annulus ventralis of female flat. (Lake Erie and its drainage. )

C. [Faxonius) propinquua Girard.

c". Rostrum without median keel. Sexual organs of male of first form at anterior margin with a prominent

angle (shoulder). Annulus ventralis of female with two tubercles in anterior part. (Ohio drainage).

C. {Faxonius) obscurut Hagen-
«' . Sexual organs of male of the Bartonius-type. Rostrum always without marginal spines. Carapace generally
without lateral spines.
Ii'. Areola wide. Form of carapace depressed. Color brownish or greenish. (Species of the small streams. )

C. (Bartonius) barton* i Fabricins).
/' '. Areola narrow or obliterated in the middle. Form of carapace rather compressed. ( Burrowing species. )

</. Areola narrow. Inner margin of hand generally with only one row of tubercles. Color very bright, of

tints unusual among crawfishes.

<<'. Color red. Rostrum short and very broad. Outer margin of hand serrated. ( Mountains of sonthern

Pennsylvania). * C. {Bartonius) carolinus Erichson.

A". Color blue. Rostrum short and narrower. Outer margin of hand rounded, not serrated. (Hills of

southwestern Pennsylvania.) C. {Bartonius) monongalensi* Ortmann.

352 MEMOIRS OF THE CARNEGIE MUSEUM

c". Areola generally obliterated in the middle, or extremely narrow. Inner margin of hand with several
rows of tubercles. Color greenish or brownish. Rostrum rather long and narrow. (Swamps on hills
and in lowlands, southwestern and eastern Pennsylvania.) C. (Barlonivs) diogenes Girard.

C. Description and Distribution of the Species.

1. Cambarus (Faxonius) limosus (Rafinesque).

(Plate B, Fig. 3; Plate XXXIX, Figs. 5a and 5b.)

Aslacus limosus Rafinesque, 1817, p. 4 .

Asiacus affinis Say, 1817, p. 168 ; Harlan, 1835, p. 230, fig. 2 ; DeKay, 1844, p. 23 ; Gibbes, 1850, p. 195 ipartim).

Astacus barloni Milne-Edwards, 1837, p. 331 (nan Fabrioius).

Cambarus affinis Girard, 1852, p. 37 ; Hagen, 1870, p. 60 ; PI. 1, figs. 19, 22, 84, 85, PI. 3, fig. 152, PI. 5 ; Abbott, 1873,

p. 80 ; Smith, 1874, p. 638 ; Faxon, 18846, p. 146 ; Faxon, 1885o, p. 86 ; Faxon, 18856, p. 360 ; Underwood, 1886,

p. 366 ; Faxon, 1890, p. 628 ; Hay, 1899, p. 960, 964 ; AndrewB, 1894, p. 165.
Cambarus peolei Girard, 1852, p. 87.
Cambarus (Faxonius) limosus Ortmann, 19056, p. 107, 112, 131.

Body robust, pubescent all over, but cbiefly so on carapace and chela?; but the
pubescence wears off easily, and in old individuals, especially in early spring, the
body is more or less hairless. The hairs are most persistent on the fingers of the
large chelae.

Carapace subovate, depressed, the depression being brought about by a bulging
out of the branchial regions. The vertical height of the carapace, at a point about in
the middle of the gastric region (measured from this point to a point on the sternum
just in front of the first pereiopods), the vertical height of the carapace at a point of
the areola directly above the sternum between the second pereiopods, and the great-
est width of the carapace at the hepatic regions are about the same; while the great-
est width of the carapace at the branchial regions is distinctly greater. Relation
G : H: B = 1 : 1 : 1.2 to 1.4.3 The greatest width of carapace (at branchial regions)
is well behind, at about the middle of the branchial regions. The whole carapace
appears rather flattened dorsal ly.

Cervical groove deep, continuous on the sides. Areola about half as long as an-
terior section of carapace (including rostrum). Relation of a:p = 1 :0.40 to 0.55. 4
Areola rather broad, relation of w : I = 1 : 4.0 to 5.4," with about 5 irregular rows of
punctures.

Rostrum long and broad, reaching to the middle of the fifth joint of the peduncle
of the antenna, and to the end of the peduncle of the antennula, rarely slightly

3G = vertical diameter at gastric region ; //= transverse diameter at hepatic region; B = transverse diameter at
branchial region.

•a = anterior, p = posterior section of carapace.
6 «i = width, I = length of areola.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 353

longer. Surface deeply concave, margins elevated and thickened, almost straight,
very little convergent toward the marginal spines. Marginal spines well developed.
Acumen long, triangular, acutely pointed, about one third as long as the whole ros-
trum, sometimes slightly shorter or longer. Post-orbital ridges parallel, ending in a
sharp spine anteriorly.

Surface of carapace finely punctate, and very finely granulate on the sides in old
specimens. Sides spinose. There are a number of larger and smaller spines on each
side on the hepatic region, and a few spines are found on the branchial region
immediately behind the cervical groove, of which one is generally much larger than
the others. (In most cases there are two distinct spines, one above the other, the
lower one the larger.) All these spines are well developed only in larger individ-
uals ; in young ones only two spines behind the cervical groove and one spine on
the hepatic region are present, but these are visible even in the smallest specimens
at hand (25 mm. long). External orbital angle not marked, rounded off. Branchio-
stegal spine sharp and distinct.

Abdomen longer than the carapace, slightly narrower than the carapace in the
male, slightly wider in the female. Anterior section of telson on the outer posterior
corners generally with two spines, but there may be from one to three; the num-
ber of spines may differ on either side. Posterior section of telson semicircular,
slightly wider than long, and slightly shorter than anterior section.

Epistoma with posterior part short and broad, almost three times as wide as long,
not plane, with a transverse groove posterior to the middle, and an anterior median
depression ; these are often united into a triangular or arrow-shaped depression.
Anterior section constricted at base, its anterior margin almost semicircular, with a
small median point, slightly varying in shape (sometimes it is subtriangular, some-
times the anterior point is obscure), but its transverse diameter is always slightly
greater than the longitudinal.

Antennula with a sharp spine on the lower margin of the basal joint.

Anfcimal peduncle with a sharp spine on the outer side of each of the two basal
joints.

Antermal scale long, as long as the rostrum or even slightly longer, reaching to the
middle or the end of the terminal joint of the antennal peduncle. Outer margin
with a strong spine. Laminar part rather broad ; its margin more or less regularly
curved ; the broadest part is in the middle or slightly anterior to it.

Flagellum, when laid backward, reaching to the fourth or even to the middle of
the fifth abdominal segment in the male ; in the female it generally does not reach
beyond the posterior margin of the third segment.

354 MEMOIRS OF THE CARNEGIE MUSEUM

First pereiopods comparatively short, and not very stout, considering the size of
the species. Hand short and not very broad, depressed, elongate-ovate, stronger
and more elongate in. the male than in the female. Surface punctate. Inner mar-
gin almost straight, with a double row of tubercles, which are more or less spiniform.
Outer margin smooth, bluntly angular, more distinctly so distally. Fingers dis-
tinctly longer than the palm (measured from articular tubercle on upper side of
carpopodite to articular tubercle on upper side of palm at base of movable finger),
straight, cutting edges straight, in contact all along their length, with a few very
small tubercles in the proximal part, for the rest without teeth or tubercles, but
Avith a short and dense pubescence, becoming slightly barbate proximally on lower
side. Upper surface of each finger with a low longitudinal rib, most distinct dis-
tally. Lower surface of hand almost smooth, sparsely punctate.

Carpopodite slightly longer than wide, shorter than palm, punctate. Upper
surface with a distinct longitudinal sulcus. Inner margin with a strong procurved
spine in the middle, and a small spine anterior to it. Lower surface with two
strong spines, one in the middle of the anterior margin, the other at articulation
with hand. Sometimes there are additional small spines or spiniform tubercles,
proximal to, or above, the large spine of the inner margin.

Meropodite smooth, upper margin with two (rarely more, up to four) strong
spines at a short distance from the distal end. Lower margin with two rows of
strong spines, the inner one consisting of four to ten spines, largest distally, the
outer one of two to tbree spines. A spine at the outer articulation with the
carpopodite.

Ischiopodite of third pereiopods hooked in the male ; hook in the male of the first
form strong, subcorneal.

Coxopodites of posterior pereiopods without prominent crests or tubercles in the
male.

First pleoptods of male of the first form (Plate I, Fig. 5a and 56) rather strong and
short, not reaching beyond the anterior margin of the coxopodites of the third
pereiopods. They are not articulated at the base, straight, and the two parts are
separated at the tip's only for a short distance. Tips crossed (twisted), divergent;
that of the inner part is soft, gradually tapering to a point, and is directed obliquely
outward ; that of the outer part is horny, gradually tapering to a point, and directed
obliquely forward and slightly inward.

In the male of the second form the first pleopods are articulated at the base
when young, but not articulated when old, and both tips are soft ; that of the outer
part is rather bluntly pointed.

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 355

Annulns ventralis of female transversely rhombiform, with a short transverse
groove slightly posterior to the middle, and an S-shaped longitudinal fissure. An-
terior to the central groove there is on each side of the fissure a strong, tuberculi-
form elevation, so that the fissure is situated in a rather deep depression. Posterior
to the central groove, there is a slight elevation, over which the fissure passes. The
annulus, consequently, appears trituberculate, the two anterior tubercles heing
stronger than the posterior. In young females, the tubercles are only slightly de-
veloped, and generally the posterior tubercle is almost obsolete.

Size. — Rafinesque gives as total length 3-9 inches. The maximum, 9 in., = 229
mm., seems rather strange, since no such specimens have ever been subsequently
seen, even if we infer that Rafinesque intended the whole length, including
the claws. Hagen (1*70, pi. 5) figures a very large female, which, including the
outstretched claws, would not he longer than about 7 inches (178 mm.). Its body
from the tip of the rostrum to the end of the telson, is 132 mm. long. The maxi-
mum length given by Hagen in the text (p. 01) is 4.7 in., = 119 mm; thus this
figure, although said to be of natural size, is apparently somewhat enlarged.

The largest specimens ever seen by the writer are in the museum of Oberlin
College from the Potomac River, a female measuring 120 mm., and a male of the
first form measuring 105 mm. 'The largest specimen from Pennsylvania I poss< 38,
is a female from the Delaware River at Torresdale, and measures 93 mm. in length :
the largest male (first form) is from the Delaware at Penhs .Manor, measuring 75
mm. in length. A male of the second form from Holmesburg is 85 mm. long.
Specimens over 100 mm. long, mentioned by the writer, (1899, p. 1210), as from
Philadelphia, are from the New Jersey side of the river, near Camden.

Colors. — (Plate B, Fig. 3.) An account of the color of this species has been given
by Faxon (1885o, p. 88). It runs thus: " Upper surface greenish, mottled with
darker green, especiall}' on the cheke ; tips of fingers orange, preceded by a dark
green ring, which runs along the outer border of the hand to the wrists; abdominal
somites ornamented with interrupted transverse chestnut-colored double bands.
Under surface of a lighter hue."

I have repeatedly made notes from live specimens, and have found that the
shades of color vary greatly, although the general pattern has been correctly
described by Faxon. The general color of the body may be described as olive-gret n
(Ridgway, 1^86, X, 18)*, but it varies toward tawny-olive (III, 17), and olive-yellow
(VI, 16). The sides of the carapace are generally lighter, of a whitish green.

6 In the description of colors, I have used here (and in the following species) the nomenclature of Kidgway (1886),
and the Roman and Arabic numerals refer to his plates and ligures.

356 MEMOIRS OF THE CARNEGIE MUSEUM

There is a brown [chestnut, IV, 9) spot on the anterior margin of the carapace on
each side below the eyes, not noticed by Faxon. The brown bands of the abdomen
are burnt sienna (IV, 6). In the middle of the abdomen the epimera are hazel (IV,
12). The color of the finger tips is ferrugineous (IV, 10), often paler, the preceding
band is dark olive-green, often almost black. The articular tubercles on the lower
side of the hand are tawny (V, 1), on the upper side they are dark-green. The
articular membranes of the chela? are wine-purple (VIII, 15). The darker green of
the carapace is generally confined to distinct large blotches, symmetrically disposed ;
one pair on the gastric region, and one each on the anterior and the posterior part
of the branchial regions. Often the two blotches of the gastric region run together,
which may also be the case with those of the branchial regions. They often appear
spotted or mottled with the lighter, or rather more brownish (tawny olive), ground
color. The brown spot on the anterior margin of the carapace is sometimes indis-
tinct, and in young specimens with fresh shells, it may have a trace of yellow
below. All these colors are bright and distinct only in fresh shells. On old shells,
a coat of mud is generally deposited, giving to the whole body a dirty blackish
color, and besides, the colors themselves fade considerably, so that only a dirty olive-
green remains, with some brown on the abdomen.

The color of newly laid eggs under the abdomen of the female is olive-green (X,
18).

The above description is founded upon the examination of one hundred and
twenty-one specimens, now preserved in the collections of the Carnegie Museum.
Fifty-six of these specimens are from the state of Pennsylvania, fifty-four from
New Jersey, eight from Maryland, and three from West Virginia (Potomac River
at Cherry Run, Morgan County). This, however, does not represent the total num-
ber of specimens seen by the writer, since many others were collected by him, as
well as seen in the collections of the Department of Agriculture of Pennsylvania, of
the Academy of Natural Sciences in Philadelphia, and of Oberlin College.

DISTRIBUTION.7

LOCALITIES REPRESENTED IN THE COLLECTIONS OF THE CARNEGIE

MUSEUM.

Pennsylvania : Burks Co., Delaware River, New Hope ; Delaware River, Penns
Manor; Little Neshaminy Creek, Grenoble ; Common Creek, Tullytown ; Philn-
delpliia Go., Delaware River, Torresdale Fish Hatchery, Torresdale ; Delaware

7 All localities without further record have been ascertained by the writer in person. In other cases the authority
(when published), or the collector and institution, where the specimens are preserved, is given.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 857

River, Holmesburg, (H. W. Fowler coll., exch. Acad. Phil. Nat. Sci.); Schuylkill
Canal, Manayunk; Montgomery Co., Schuylkill River, West Manayunk, (H. Gera
coll.) ; Delaware Co., Marcus Hook Creek, Marcus Hook ; Franklin Co., Back ( 'reek,
Williamson; Bedford Co.t Raystown Branoh of Juniata River, Bedford (A. Kcenig
coll,).

New Jersey : Camden Co,, Delaware River, Camden ; Delaware River, North
Cramer Hill ; Mercer Co., Stony Brook, Princeton ; Delaware-Raritan Canal,
Aqueduct near Princeton.

Maryland : Alleghany Co., Potomac River, Wiley's Ford, South Cumberland.

West Virginia : Morgan Co., Potomac River, Cherry Run.

PREVIOUS RECORDS.8

Type locality : Delaware River, Philadelphia, (Rafinesque).

Pennsylvania: Schuylkill River, Philadelphia, (Hagen) ; Bucks Co., Bristol,
(Faxon); Berks Co., Schuylkill River, Reading, (Girard) ; Chester Co., Brandy-
wine Creek, (Faxon); Tributary of Brandy wine Creek, Chadds Ford Junction, (Ort-
mann) ; Lancaster Co., Bainbridge, (Faxon); Cumberland Co., Carlisle, (Hagen);
Adams Co., Gettysburg, (Ortmann).

New Jersey: Camden Co., Camden, (Faxon); Burlington Co., Burlington,
(Faxon) ; Mercer Co., Trenton, (Abbott) ; Monmouth Co., Red Bank, (Faxon) ; Morris
Co., Schooley's Mountain, (Faxon).

Maryland : Cecil Co., (Faxon) ; Hartford Co., Havre de Grace, (Hagen) ; Balti-
more Co., Baltimore, (Andrews); Guynn's Falls, (Faxon); Druid Hill, (Faxon); Anne
Arundel Co., (Faxon); Charles Co., (Faxon); Montgomery Co., (Faxon); Washington
Co., Williamsport, (Faxon) ; Alleghany Co., Canal four miles south of Cumberland,
(Faxon).

District of Columbia : Washington, (Girard).

Virginia : Fairfax Co., Gunston, (Faxon) ; Augusta Co., Shenandoah River,
Waynesboro, (Faxon); Isle of Wight Co., Blackwater River, Zuni, (Faxon).

NEW LOCALITIES NOT REPRESENTED IN CARNEGIE MUSEUM.

Pennsylvania: Montgomery Co., Roberts Run, Abrams, (B. \Y. Criftitbs it 11.
W. Fowler, coll., Acad. Nat. Sci., Philadelphia).

The following records are from the collections of Mr. W. R. MeConnell, belong-
ing to the Department of Agriculture, Harrisburg : Yellow Breeches Creek, New
Cumberland, Cumberland Co.; Conedogwinet Creek, West Fairview, Cumberland
Co. ; Sherman's Creek, L&ndisburg, Perry < !o. ; Montour Run, Greenpark, Perry

8 A number of doubtful records have been dropped (see Ortmann, 1905/-, p. 131). Too general record? are also
omitted, for instance : " Susquehanna Kiver, I'a., ( Faxon)," since just-in this pase more exact information is desired.

358 MEMOIRS OF THE CARNEGIE MUSEUM

Co.; Susquehanna River, Northumberland, Northumberland Co. ; Fishing Creek,
Bloomsburg, Columbia Co. ; Bald Eagle Creek, Milesburg, Center Co. ; Conoco-
cheague Creek, Cbambersburg, Marion, and Williamson, Franklin Co. ; Maiden
Creek, Maiden Creek, Berks Co.

REMARKS.

Cambarus limosus is the common river species of eastern Pennsylvania. Its
morphological characters are very constant, and give it a rather isolated position
within the genus, which is also expressed by, and very likely due to, its geograph-
ical isolation, the most closely allied species being found far to the west, in Indiana
and Kentucky, (see Ortmann, 19056, p. 114, 127). The most prominent specific
characters are furnished by the male sexual organs, and the spinosity of the sides
of the carapace. The description, as given above, does not indicate any important
variations, and the specimens are generally very uniform. The spinosity of the
carapace, however, changes with age, young specimens being much less spinose than
old ones. In the spines of the chelipeds and of the anterior section of the telson,
there is some variation, but this is only slight and not subject to any rule. The
shape of the carapace and rostrum is very constant, the only differences of age no-
ticed are found in the acumen of the rostrum, which in young specimens is slen-
derer than in those which are older, and in the bulging out of the branchial regions
of the carapace, which is most marked in old individuals. The changes in the
pubescence of the whole body are apparently due to wear. The short hairs gener-
ally present in newly moulted individuals slowly wear off, and specimens with a
distinct coat of dirt upon them, indicating age, generally have the pubescence more
or less, sometimes entirely, worn off. Only on the hands and fingers are traces of
it left.

I myself have never found any freaks in this species. But Mr. W. R. McConnell
found a male (first form), 66 mm. long, at Bloomsburg, Columbia County, (the only
specimen taken at this locality), in which the rostrum had two pairs of marginal
spines. The additional pair in this specimen is smaller, and stands about midway
between the normal pair and the base of the rostrum.

2. Cambarus (Faxonius) propinquus Girard.
(Plate XXXIX, Fig. 6o and 66.)

Cambarus propinquus, Girard, 1852, p. 88 ; Hagen, 1870, p. 67, PI. 1, f. 34-38, PI. 3, f. lo3 ; Smith, 1874, p. 638 ; Forbes,
1876, p. 4, 19; Bundy, 1877, p. 171 ; Bundy, 1882, p. 181 ; Bandy, 1883, p. 402; Faxon, 1884, p. 147 ; Faxon, 1885™,
p. 91 ; Faxon, 18856, p 360; Underwood, 1886, p. 371 ; Faxon, 1890, p. 628 ; Hay, 1896, p. 497, Fig. 11 ; Ward,
1896, p. 15 ; Faxon, 1898, p. 651 ; Hay, 1899, p. 960, 962 ; Ortmann, 1905«, p. 400.

Cambarus {Faxonius) propinquus Ortmann, 1505fc, p. 112, 132.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 359

Body not very robust, not pubescent, with only a few, scattered, short hairs,
chiefly on the chelte, but the hair wears off very soon, and the body becomes smooth,
with exception of a slight pubescence at the base of the cutting edges of the fingers.

( 'urapace subovate, depressed. Relation G: H:B = 1 : 1 : 1.2 or 1.3. (The width
of the branchial regions appears slightly less than in G. limosus, but this is probably
due to the fact that the specimens of C. propinqwus at hand are rather small).
Greatest width of branchial regions well behind, at about the middle of the bran-
chial regions. Carapace flattened dorsally.

Cervical groove deep, more or less distinctly interrupted on the sides just above
the lateral spine. Posterior section of carapace about half as long as anterior (rela-
tion a :p = 1:0.42 to 0.62.

Areola rather broad, w.l = 1:4.7 to 6.0, with about four irregular rows of
punctures.

Rostrum long and broad, reaching to the middle of the fifth joint of the peduncle
of the antenna, and to the end of the peduncle of the antennula, sometimes slightly
shorter. Surface concave, with a more or less distinct, low, longitudinal median
keel toward the tip. Margins elevated, but not much thickened, straight, more or
less convergent toward the marginal spines. Marginal spines generally well devel-
oped in young specimens, less so in older ones, sometimes quite small. Acumen
long, triangular, comparatively longer in young specimens, pointed, about one-third
as long as the whole rostrum, or shorter. Postorbital ridges slightly divergent pos-
teriorly, ending anteriorly in a more or less distinct small spine.

Surfaot of carapace finely punctate, and slightly granulate on the hepatic regions
in old specimens. Sides with only one spine on the branchial regions, immediately
behind the cervical groove. This spine is always present and sharp, but generally not
very large. No other spines on the sides of the carapace. External orbital angle not
marked. Branchiostegal spine small, but sharp, or tuberculiform, or even obsolete.

Abdomen longer than carapace, slightly narrower than the carapace in the male,
about as wide as the latter in the female. Anterior section of telson on the outer
posterior corners with one to three spines (two spines is the general condition).
Posterior section of telson semielliptical, considerably wider than long, and slightly
shorter than anterior section.

KpistonKi with posterior part short and broad, almost three times as wide as long,
not plane, with a transverse groove and an anterior median depression running into
each other. Anterior section constricted at base, its anterior margin generally almost
semicircular, with a median point and an indistinct angle on each side, but shape
rather variable: sometimes it is truncate anteriorly, with or without median point.

360 MEMOIRS OF THE CARNEGIE MUSEUM

the lateral angles being more distinct. Its transverse diameter is slightly greater
than the longitudinal.

Antennula with a distinct, sharp spine on the lower margin of the basal joint.

Antennal peduncle with a distinct spine on the outer side of the first joint, and a
smaller, sometimes tuberculiform spine on the second joint. Antennal scale long, as
long as rostrum or slightly longer, reaching to the middle, or almost to the end of
the terminal joint of the antennal peduncle. Outer margin with a strong spine.
Laminar part rather broad, almost semicircular, the broadest part is slightly anterior
to the middle.

Flagellum reaching to the beginning of the fifth abdominal segment in both the
male and the female.

First pereiopods not very robust, comparatively longer in the male, shorter in
the female. Hand elongate-ovate, depressed, moderately wide. Surface punctate.
Inner margin almost straight, with a double row of tubercles. Outer margin
smooth, marginated and bluntly angular, but almost evenly rounded near the
proximal end. Fingers longer than palm, almost straight in the female, and meet-
ing all along their edges ; in the male, the fingers are slightly gaping at the base,
and the movable one is slightly curved in the shape of an " S," which curve is chiefly
noticeable along the outer margin. Outer margin of movable finger slightly tuber-
culate at base. Cutting edges with a few small tubercles near the base, for the rest
slightly pubescent. Upper surface of each finger with a low, longitudinal rib.
Lower surface of hand sparsely punctate.

Carpopodite slightly longer than wide, shorter than palm, punctate, and with a
longitudinal sulcus on upper side. Inner margin with a strong, slightly procurved
spine in the middle ; generally there is a tubercle (rarely spiniform) anterior to
this spine. Lower surface with a low and broad tubercle in the middle of the
anterior margin, which is very rarely subspiniform ; a similar tubercle with a
spiniform tip at the articulation with the hand. No other spines or tubercles on
the carpododite, except that sometimes there is a small tubercle at the proximal
end of the inner margin.

Meropodite smooth ; upper margin with two small, often indistinct, or tuberculi-
form, spines near the distal end. Lower margin with two rows of spines ; the outer
i'ow consisting of only one, rarely of two, spines ; the inner row consisting of a large
distal spine, and a number (up to seven or eight) of very small ones, which may be
entirely absent. Thus there are often only two anterior spines present, representing
the distal spine of each row. A small spine at the outer articular tubercle with
carpopodite.

0RTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 361

Ischiopodite of third pereiopod hooked in the male, the hook in the male of the
first form being strong, subcorneal.

Goxopodltes of posterior pereiopods without prominent crests or tubercles in the
male.

First pleopods of male of the first form (Plate XXXIX, Fig. 6a) slender, but
rather short, hardly reaching beyond the middle of the coxopodites of the third
pereiopods. They are not articulated at the base, and the two parts are completely
separated at the tips for a rather considerable distance (about one third of the
length from the inner basal tuKercle to the tip). Both parts are almost parallel,
only slightly convergent at the tips, which is due to a very slight curve of the outer
part. Outer part gradually tapering from base to tip, horny. Inner part soft, of
about the same shape as the outer, and of the same length, gradually tapering to
an acute tip. Both parts are slightly twisted, so that the tip of the outer is directly
anterior to that of the inner. Anterior margin of this organ without shoulder
shortly below the point of separation of the two parts ; sometimes, indeed, there is
a slight notch, but never a sharp shoulder.

In the male of the second form, the first pleopods (Plate XXXIX, Fig. 6/;) are
articulated at the base when young, (only young specimens are at hand) ; both parts
are separated only for a short distance, and are soft; the outer one is rather
blunt, while the inner one tapers to a point. No notch or shoulder on anterior
margin.

Annulus vmbralis of female transversely rhombiform or ovate, rather flat, very
slightly depressed in the middle, with an S-shaped longitudinal fissure. No tuber-
cles on anterior part. In young females, the median depression is very indistinct,
and the annulus is almost completely flat.

Size. — Hagen gives 2.6 in. = 6f5 mm. as the maximum length for this species. The
largest individuals from the state of Pennsylvania observed by the writer, are a
male (first form) from Albion, Erie County, 61.5 mm. long, and a female from the
same locality 69 mm. long. I have seen, however, two larger males (first form)
from Lake Erie, off the shore of Lorain County, Ohio (Mus. < >berlin), one measuring
77 mm., the other 81 mm. in length. Nevertheless, this seems to be one of the
smaller species, for in the streams running to Lake Erie in Pennsylvania a consider-
able number of individuals have been taken, none of which was longer than the
above mentioned specimens.

Colors. — The colors of this species agree closely with those of ( '. obsou rus i which
see for further particulars). The following notes were taken from an adult female,
collected on the shore of Lake Erie at Miles < Jrove.

362 MEMOIRS OF THE CARNEGIE MUSEUM

General color, olive-green (Ridgway, 1886, X, 18), sides of carapace cream-color
(VI, 20). A rufous (IV, 7) spot on anterior margin of carapace. Lower side
whitish. Anterior half of abdominal segments hazel (IV, 12). Chelae olive-yellow
(VI, 16), mottled with olive-green. Finger tips orange (VI, 3), followed proximally
by a citron-yellow (VI, 15) band. One rufous articular tubercle above on the hand.
Articular membranes of hand lake-red (VII, 2). Finger tips of chelse of second and
third pereiopods, and dactylopodites of fourth and fifth pereiopods orange. Pereio-
pods pale brownish-white, mottled and marbled with olive-yellow. Antennal flag-
ellum annulated olive-green and ochraceous (V, 7). Spines on sides of carapace and
rostrum buff (V, 13). Antennal scale olive^yellow, its center olive-green.

It is to be remarked that in this specimen no dark olive-green band is found
near the finger-tips. The same was the case generally in specimens from Temple
Creek, Albion, and from Elk Creek (all collected in autumn). However, specimens
from Conneautville Station, Crawford County, collected in June, generally had a
dark green, almost black band, succeeding the pale band. A similar dark band
appeared in some of the Temple Creek specimens, after they had been preserved
for some time in alcohol, but it disappeared again with the progress of the bleach-
ing action of the preserving fluid. In collecting the specimens of this species and
of C. obscuras in Erie County in October, 1904, I was generally able to distinguish
the two species, where they were found associated, by the color of the finger tips.
However, too much reliance should not be placed upon this character, since I was
not subsequently able to test this observation.

The description of this species, as given above, is drawn from sixty-one speci-
mens preserved in the collection of the Carnegie Museum. Of these, fifty-three are
from the State of Pennsylvania (forty-eight from streams flowing into Lake Erie,
five from the lake itself). One specimen is from Lake Erie, Erie County, Ohio, and
seven are from the northern parts of Michigan.

DISTRIBUTION (see Plate XLII, Fig. 3).
LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.

Pennsylvania: Erie County, Lake Erie, Presque Isle, (1). A. Atkinson coll.);
Lake Erie, Miles Grove ; Walnut Creek, Swanville ; Elk Creek, Girard ; Elk Creek,
Miles Grove ; Conneaut Creek, Albion ; Temple Creek, Albion ; Crawford County,
tributary of Conneaut Creek, Conneautville Station.

Ohio : Erie County, Lake Erie, Cedar Point, near Sandusky, (O. E. Jennings
coll.).

Michigan : Emmet County, Crooked Lake, Oden near Petoskey, (E. B. William-
son coll.).

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 363

PREVIOUS RECORDS.9

Type locality: Oswego, Oswego County, New York, (Girard).10

Canada: Montreal, Quebec, (Faxon); Toronto, Ontario, (Faxon).

New York: St. Lawren.ce County, Grass River, (Hagen); Canton, (Faxon); Black
Lake, (Faxon); Ogdensburg, (Faxon); Jefferson County, Garrison Creek, Sackett's
Harbor, (Girard); Oneida County, Oneida Lake, (Hagen); Cayuga Conn!;/, Cayuga
Lake, (Faxon); Monroe County, Rochester, (Hagen); Niagara County, Niagara-
(Hagen); Chautauqua County, Forestville, (Faxon).

Ohio: Lorain County, Lake Erie, (Ortmann); Ottawa Count//, Portage River, Oak
Harbor, (Faxon).

Michigan: Wayne County, Detroit River, (Faxon); Northville, (Faxon); Ecorse,
(Faxon); Washtenaw Count//, Ann Arbor, (Faxon); St. Clair County, St. Clair River,
(Faxon); Calhoun County, Marshall, (Faxon); Allegan County, Otsego, (Faxon); Sagi-
naw Count//, Saginaw Liver, (Faxon); Charlevoix, Count//, Lake .Michigan, Bound and
Pine Lakes, Charlevoix, (Ward)."

Indiana: Dc Kalb County, Waterloo, (Hay); Noble County, Rome City, (Bundy);
Kosciusco County, Turkey Lake, (Hay); Marshall County, Maxinkuchee Lake, (Hay);
Twin Lakes, (Hay); Laporte County, Michigan City, (Faxon); Carroll County, Delphi,
(Hagen); Tippecanoe County, Lafayette, (Faxon); Marion County, Indianapolis
(Faxon); Irvington (Hay); Millers ville (Hay); Franklin Count//, Brookville (Hay);
Brown Count//, Salt Creek (Hay); Monroe County, Clear Creek, Blooniington
(Faxon); Greene County, Switz City (Faxon); Sullivan County, Turman Creek
(Faxon).12

Illinois: Macon Count//, Decatur (Faxon); McLean Cownty, Normal (Forbes);
Tazewell Con nt /i, Pekin (Forbes); Kane County, Geneva (Faxon): Ogle County,
(Hagen); Stephenson County, Freeport (Forbes).13

Wisconsin: G reeve Count// (Faxon); Dane County, Madison (Faxon).

Iowa : Scott County, Davenport (Faxon); Wapello Count//, Ottumwa (Faxon).

New locality, not represented in Carnegie Museum: Spencerport, Monroe Co.,
New York, (Mm Oberlin).

9I have omitted "Lake Superior" (Hagen), as unconfirmed (see Ortmann, 1905fc, p. 132), and "Green River,
Edmonson County, Ky." ( Hay, 1803a, p. 235), as doubtful, being founded upon young specimens only.

'"This is the first locality given by Girard, and consequently is the type locality.

11 Faxon gives also from "Michigan": "St. Mary's Lake"; "Mouth of Battle Creek"; and "Lake Douglass."
I have not been able to locate these.

12 Faxon gives in addition : " White River, Indiana " (southeastern section, tributary to Ohio.)

13 Faxon gives also : " Aux Plains River, Illinois," which I have not been able to locate.

364 MEMOIRS OF THE CARNEGIE MUSEUM

REMARKS.

Cambarus propinquus in Pennsylvania belongs to Lake Erie and its drainage.
The range being rather restricted, the material at hand is not very rich, and its
study does not promise many results as to variation. Nevertheless there are a few
striking facts, which may be mentioned. In the first place one of the chief specific
characters, the longitudinal keel of the rostrum, is decidedly variable. All speci-
mens at hand from outside of the state (eight) show a keel plainly, but this is not
so with the Pennsylvanian specimens. The keel in these is often distinct, but shows
a tendency to disappear. This is chiefly the case in young individuals, where the
rostrum is comparatively narrower, the marginal spines are sharper, and the acumen
is slenderer than in older individuals.

The armature of the chelipeds is also rather variable. There is always a strong
spine in the middle of the inner margin of the carpopodite, and invariably a small
tubercle anterior to it, which in young specimens is often spiniform. Sometimes
there is also a small tubercle at the proximal end of the inner margin, but I have
generally found this only in larger individuals. The lower side of the carpopodite,
as a rule, has only one spine, located at the articulation with the hand, and this is
present in all Pennsylvanian specimens I have seen. The anterior margin is often
without any spine, or even tubercle; there is, however, a low tubercle developed in
many cases, and in two cases it was spiniform, viz., in a male (second form) from
Elk Creek, Miles Grove, and in a female from Presque Isle. Both of them had a
distinct keel on the rostrum, so that they undoubtedly belong to this species. The
rows of spines on the lower margin of the meropodite are generally represented by
only two spines, the distal spine of each row being alone present. But it is remark-
able that in the set from Conneautville Station, composed of twelve individuals, ten
show an increase of the spines of the inner margin, from four to eight little teeth
being present behind the large distal spine, while in eight specimens an additional
smaller spine is found behind the distal spine of the outer margin. In every case
this occurs only on one side, while the other side is normal. A similar increase of
the number of spines of the meropodite is also to be observed in a few specimens
from Temple Creek, Albion, in the two specimens at hand from Elk Creek, Miles
Grove, in the female from Presque Isle, mentioned above, and in the specimen
(female) from Sandusky, Ohio. Since the latter has also a spine on the anterior
margin of the lower side of the carpopodite, the tendency to develop additional
spines may extend simultaneously to carpopodite and meropodite.

The set of seven specimens from northern Michigan is remarkable for the fact

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 365

that in no case is there a spine on the carpopodite at the lower articulation with the
hand, but only a tubercle ; in other respects they are typical, with a spine and
anterior tubercle on the inner margin of the carpopodite, a low tubercle at the
anterior margin of the lower side of the carpopodite, and only two spines on the
lower margin of the meropodite.

In the shape of the sexual organs of the male there is much uniformity in their
length and the shape of the tips. However, there is a tendency in the Pennsyl-
vanian specimens toward the development of a slight notch on the anterior margin
in the place where C. obscurus has a shoulder. I have only twelve males of the first
form from Temple Creek, Albion, four from Walnut Creek, and two from Elk
Creek. Of these, six from Temple Creek and one from Walnut Creek show a notch,
while all the rest (eleven) have no trace of it. The notch never assumes the shape
of the "shoulder" of G obscurus, and the sexual organs differ in other respects
from the later species, chiefly in that the tip of the inner part is never blunt or
dilated.

The female annulus is rather constant, as has been said above ; only slight differ-
ences due to age are noticeable.

We may sum up the variations of G propinqnus in the state of Pennsylvania by
saying that there is a distinct inclination toward G obscurus, indicated by the tend-
ency of the rostral keel to disappear, of the chelipeds to increase in spinosity, and of
the male copulatory organs to develop a notch at the anterior margin. Nevertheless
there are numerous specimens winch represent the typical ( '. propinqwus. This fact
is to be borne in mind, and we shall learn more about it when we come to discuss
the geographical distribution of this and the related forms.

I introduce here the systematic account of a variety of this species, which is
extralimital to the state of Pennsylvania. I have, however, decided to treat of it
more fully, since its relation to the representative Pennsylvania form is highly inter-
esting, and since we shall have to refer to it repeatedly in the chapter on distribution.

2a. Cambarus (Faxonius) propinquus sanborni (Faxon).

Cambarus sanborni Faxon, 1884J, p. 128.

Cambarus propinqnus sanborni Faxon, 1885«, p. 91, PI. 5, f. 3, PI. 9, f. 10 ; Underwood, 1886, p. 372 ; Osburn & William-
son, 1898, p. 21 ; Williamson, 1899, p. 20, 48 ; Hay, 1899, p. 960, 964 ; Ortmann, 1905/;, p. 132.

According to Faxon, this variety differs from the typical G "propinquus in the
following characters: 1) The two parts of the male sexual organs are less deeply
separated, and the tips are closer together. 2) The rostrum is not carinate. 3) The
hands are finely pubescent. 4) The inferior median anterior spine of the carpopodite
is evident.

366 MEMOIRS OF THE CARNEGIE MUSEUM

I possess five specimens (obtained by exchange from Oberlin College) from one
of the two localities originally mentioned by Faxon for this form (Oberlin, Ohio),
which agree well with his account, with the exception that the pubescence of the
hands is not developed ; there are, indeed, a number of short hairs in some, chiefly
the younger, specimens, implanted in the punctures, but such are also very fre-
quently present in G propinquus (as well as in C. obscurus). These hairs are gener-
ally present in new, recently moulted specimens, but wear off with age.

The male copulatory organs are very similar in shape to Faxon's figure, although
they vary slightly with reference to the length of the separated tips. In this
respect, however, the variety is closer to the typical form than to G. obscurus. In
addition, I notice in the two males of the first form of this set that the inner part of
the male organs, although it tapers to a point on a side view, is different on a poste-
rior view. From behind it is'broadly and bluntly rounded off, a fact which is due
to a marked compression and flattening in an anteroposterior direction toward the
tip. This is also the case in the male of the second form at hand. Here both tips
of the sexual organs are blunt, that of the inner part less so than that of the outer.
I cannot see that the tips of the male organs, either in the first or in the second form,
are closer together than in G. propinquus.

The armature of the chelipeds in these specimens is slightly different from that
of the typical C. propinquus, although similar variations have been observed in the
latter. In all five specimens there are two distinct spines on the lower side of the
carpopodite, one at the articulation with the hand, the other on the anterior margin.
The inner margin of the carpopodite, besides the one strong spine, has a distal and
a proximal tubercle, and in the two largest individuals (male and female) there are
a few additional tubercles on the upper surface of the carpopodite. The inner lower
margin of the meropodite invariably has in these specimens behind the distal
spine a row of small teeth, becoming spiniform in the larger individuals. The num-
ber of teeth in this row is from six to eight. In two specimens the outer lower margin
possesses a small tubercle behind the distal spine of each cheliped. In one specimen
there is only a tubercle on the right side, and two others have only the distal spine.

In addition to the above specimens from Oberlin I have seen among the Oberlin
collections other specimens from the state of Ohio, and have myself collected in
eastern Ohio and northern West Virginia a number of specimens, which undoubt-
edly belong to the same form. The characters are practically the same, and only a
few remarks are necessary.

1. The keel of the rostrum is invariably lacking. There is not a single indi-
vidual which shows any trace of it.

OKTMANN: THE CKAWFISHES OF THE STATE OF PENNSYLVANIA 367

2. The hand of the adult male of the first form has a distinct tendency to
become broader than in the typical propinqv/us. This is well shown in the largest
male from Oberlin. However, this may be due to the fact that the specimens of
this variety at hand are larger than those of the typical form. I notice, however,
in specimens from the Tuscarawas drainage and from West Virginia, a tendency in
old specimens, chiefly males, to develop on the upper surface of the hand, near the
double row of tubercles of the inner margin, additional low tubercles. These may
be scattered over the inner half of the surface, or a few of them (3-5) may form an
indistinct row between the upper articular tubercle with the carpopodite and the
articular tubercle with the dactylopodite. This is a distinct approach toward < '.
obscurus, where similar tubercles are present in larger individuals.

3. The two spines of the lower side of the carpopodite an; almost always well
developed. There are a number of specimens where they are only 1 duntly spiniform,
or even tubercular, but this is apparently due to wear, a large number of the speci-
mens at hand having been collected in spring, and possessing old worn shells, which
had gone through the winter. In a few cases the tubercle on the anterior margin
is barely indicated, but all these are cases of regenerated chelse, as indicated by
their size. The armature of the inner margin of the carpopodite entirely corre-
sponds to the Oberlin specimens, old specimens developing additional tubercles on
the upper side. A large female from Middle Island Creek, W. Va., has on the left
carpopodite a small, but sharp, accessory spine behind the large median spine.

4. The armature of the meropodite is similar to the Oberlin specimens. There
is always a series of small teeth behind the anterior spine of the inner lower margin
(in old shells they may be indistinct, due to wear) ; the outer lower margin has an
anterior spine, and often a tubercle or a small spine behind it. The latter is very
frequent in specimens from the Tuscarawas basin, while in those from the tributaries
of the Ohio in West Virginia it is rare; nevertheless, in the large female from Mid-
dle Island Creek this second spine is very prominent on the left meropodite.

5. The male copulatory organs are of the propinquvs-tyipe, that is to say, without
a shoulder. There is, however, a distinct tendency, not noticed in the Oberlin
specimens, to develop at the anterior margin a small notch in the male of the first
form, and it seems that this tendency increases in specimens taken toward the
south. Out of ten males of the first form collected at Canton, Ohio, five have no
trace of this notch, two have a slight curve in its place, and three show it clearly.
This notch in these cases never assumes the shape of a "shoulder." In specimens
from Conotton Creek in Harrison and < 'arroll ( 'onnties, Ohio, (only a few males of
the first form are at hand), no notch was observed. But out of thirteen males of

368 MEMOIRS OF THE CARNEGIE MUSEUM

the first form collected in Fishing Creek, West Virginia, only two had no trace of
it ; six had a curve developed in its place, and in five others it was distinct, in one
or two representing a blunt angle. The length of the tips of this organ varies
slightly, but it is generally less than in G. obscurus. The tip of the inner part
always corresponds to that of the Oberlin specimens, being compressed and rounded
off.

6. The annulus of the female is always of the propinquus-type, that is to say,
flat, with no tubercles. In old females it becomes a little uneven, the anterior and
posterior parts being slightly swollen, but there are never two distinct tubercles as
is the case in C. obscurus.

We may condense the varietal characters of this form as follows :

G propinquus sanborni clearly is nearer to propinquus than to G. obscurus on
account of the lack of a distinct shoulder on the anterior margin of the copulatory
organs of the male of the first form, on account of the general shape and size of this
organ, and further, on account of the flat female annulus. It differs from C. pro-
pinquus in the flattened and rounded tip of the inner part of the male organ, in the
lack of a median keel on the rostrum, and in the shape and armature of the cheli-
peds, although the latter differences are slight and not always reliable. Just in the
latter characters, and in the tendency to develop a notch on the anterior margin of
the male organ, it inclines toward C. obscurus. Thus it is clearly a transitional
form toward G. obscurus of western Pennsylvania, and its geographical distribution,
as we shall see below, is also intermediate between G. propinquus and C. obscurus.

The colors of C. sanborni agree throughout with those of G. propinquus and
( '. nl >sc a r a k. The color of the newly laid eggs is dark olive-green, sometimes almost
black.

There are one hundred and sixteen specimens of this variety at hand ; five are
from the Lake Erie drainage in northern Ohio ; eighty-one from the Tuscarawas
drainage in eastern Ohio, and thirty from Fishing and Middle Island Creeks in West
Virginia.

DISTRIBUTION.

* (See Plate XLII, Fig. 3.)
LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.
Ohio: Lorain County, Waterworks Reservoir, Oberlin (R. L. Baird coll., exch.
Mus. Oberlin); Stark Count;/, West Branch of Nimishillen Creek, Canton; Carroll
('mint!/, Conotton Creek, New Hagerstown ; Harrison Count//, Conotton Creek,
Bowerstown; Tuscarawas Gourd)/, Dennison (V. Sterki coll.).

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 3t>t>

West Virginia: Wetzel County, Fishing Creek, New Martinsville; Pleasants
County, Middle Island Creek, St. Marys.

PREVIOUS RECORDS.

Type locality : Smoky Creek, darter County, Kentucky (Faxon).

Ohio: Lorain County (Faxon)) Vermilion River; Beaver Creek; French Creek
(Ortmann) ; Wayne County, Killbuck Creek, Creston (Ortmann) ; Tuscarawas
County, Tuscarawas River, Gnadehhutten (Ortmann); Knox County, Big Jelloway
Creek (Osburn and Williamson) ; Licking County (Williamson); Franklin County,
Alum Creek (Osburn and Williamson).

3. Cambarus (Faxonius) orscurus Hagen.
(Plate A, F^ig. 1 and 2 ; Plate XXXIX, Fig. la-le ; Plate XL, Fig. 1.)

Cambarus obseurus Hagen, 1870, p. 09, PI. 1, f. 72-75, PI. 3, f. 154 ; Smith, 1874, p. 639 ; Faxon, 1884i, p. 148 ; Faxon,

1898, p. 652 ; Ortmann, 1905a, p. 402.
Cambarus propinqu.ua obseurus Faxon, 1885n, p. 92 ; Faxon, 18856, p. 360 ; Underwood, 1886, p. 372 ; Hay, 1899, p.

960, 964.
Cambarus propinquus and C. rustieus Williamson, 190], p. 13.
Cambarus ( Faxonius) obseurus Ortmann, 19054, p. 112.

Body of the same shape as in ('. propi/nquus, but slightly more robust in old
specimens.

Carapace similar to C. propinquus, but the width of the hepatic, as also of the
branchial regions, is slightly greater; G :H :B = 1 : 1.1 : 1.3 to 1.5. These differ-
ences of dimension may, however, be due to the fact that large individuals of this
species are at hand.

Cervical groove and <tr<<>l<t identical with those of C. propinquus, but the areola
generally is slightly longer than half of the anterior section of the carapace.

Rostrum similar to that of C. propinquus, but always without cny trace of a
median keel. In young specimens the rostrum and its acumen are about identical
in shape with those of C. propinquus. In older specimens there is a tendency to a
shortening of the acumen, which often reaches only to the distal end of the second
joint of the peduncle of the antennula and to the base of the terminal joint of the
peduncle of the antenna. The marginal spines in old individuals are often very
small and indistinct, represented by mere angles. The postorbital ridges are as in
< '. propinquus.

The punctures and spines of the carapace are identical with those of C.
propinquus.

870 MEMOIRS OF THE CARNEGIE MUSEUM

The abdomen, epistoma, antennula, and antenna are also similar to those of G.
propinquus.

The first pereiopods (Plate XL, Fig. 1) are generally more robust than in C. pro-
pinquus, particularly in adult males. Hand wider and more distinctly depressed.
The fingers more widely gaping in old males, and the S-shaped curve of the movable
finger more pronounced ; in old females there is also a slight gap at the base of the
fingers. The upper surface of the hand possesses, particularly in large specimens,
a small number of scattered low tubercles near the inner margin, and very often
(but not always) there is a row of 3-5 tubercles running toward the base of the
movable finger, parallel to the inner margin. Tubercles of the outer margin of the
dactylopodite more pronounced. The sculpture of the hand is rather variable, and
most distinctly developed in old males. The shape of the hand is rather different
in the male and female ; in the female the fingers are shorter, less gaping (or not at
all), rendering the outline of the hand more regularly ovate. (See Plate A, Figs.
1 and 2.)

The earpopodite differs from that of C. propinquus in the development of a
strong tubercle on the anterior margin of the lower side. This tubercle very rarely
is indistinct (chiefly so in regenerated claws) ; generally it ends in a distinct, stout,
conical spine. On the inner margin and on the upper face additional low tubercles
are not infrequently found.

The meropodite differs from that of G. propinquus by the constant presence of a
series of 4-8 small tubercles, or teeth, behind the distal spine on the inner lower
margin. These teeth are never wanting in any of my specimens. The outer lower
margin has one or two spines. The latter number is comparatively rare. In re-
generated claws very often there is no spine at all on the inner lower margin.

The other characters of the pereiopods are similar to those of G. propinquus.

The first pleopods of male of the first form (Plate XXXIX, Figs, la and lb) are
of the general type of those of G. propinquus, but slightly longer, reaching to the
anterior margin of the coxopodites of the third pereiopods. The inner part does
not gradually taper to the tip, but is of nearly uniform thickness, with the tip
rounded off and slightly compressed in the antero-posterior direction. Sometimes
the tip is even slightly thickened. The anterior margin of this organ, at a point
somewhat below the separation of the two parts, has a rather sharp, well marked
shoulder, which is absent in none of the specimens at hand (several hundred).

In the male of the second form (Plate XXXIX, Fig. 7c) this shoulder is missing,
and the inner part is blunt, similar in shape to the male of the first form, and not
tapering to a point as in the typical C. p>ropinquus.

0RTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 371

The awnulus ventralis of the female has the general shape of that of C. propin-

quus, but the depression in the middle is well marked, and the anterior part has
two distinct, subcorneal tubercles. The posterior part is also elevated into a flat
and low tubercle. These tubercles are less distinct in young specimens, but always
well developed in females of medium and large size.

Size. — Hagen gives the length as 3.5 in. =89 mm. The largest individual at
hand is a female from Pucketta Creek, Allegheny County (Atkinson coll.), which
measures 93 mm. in length. The largest male of the first form is from the Ohio
River at Neville Island, Allegheny Co., which is 86 mm. in length. Individuals
over 80 mm. in length are not rare in the larger rivers.

Colors (Plate A, Figs. 1 and 2). — The colors of this species are identical with tbose
of C. propinquus. In fresh specimens the general ground color is light olive-green
(Ridgway, 1886, X, 18), with darker spots in young specimens ; in older individuals
it is rather tawn/y-olive (III, 17). On the branchial region there is an oblique band of
cream-color (VI, 20), edged by olive-green near the margin of the carapace, which is
again cream-color. This cream-color in very brightly colored specimens sometimes
becomes primrose-yellow (VI, 13). On the anterior margin of the carapace below the
eyes there is a spot which may be rufous (II, 7), ochraceous-rufous (V, 5), edged with
sulphur-yellow (VI, 14), ochre-yellow (V, 9), or primrose-yellow (VI, 13). The abdomen
is olive-green or tawny-olive, shading into chestnut (IV, 9) on the anterior margins of
the segments. There are one (sometimes two) median and two lateral rows of dark
olive-green patches. The chela? are light olive-green, shading distally into olive-yellow
(VI, 16). The finger-tips are orange-buff (VI, 22), orange (VI, 3), or raw sit una
(V, 2), followed by a pale, and a dark green, sometimes almost black band. The
latter is not always present. The upper surface of the hand at the base of the dac-
tylopodite has two (rarely one) rufous or ochraceous-rufous (V, 5) articular tubercles.
The tubercles of the hand are buff-yellow (VI, 19) or buff (V, 13). The articular
membranes of the hand are wine-purple (VIII, 15). The legs are olive-yellow (VI,
16) and whitish, with olive-green on upper edges.

The above colors fade in old specimens, and are often obscured by blackish or
brownish coats of dirt. A variety with the chela} and anterior parts of the carapace
of a pale dirty bluish color was repeatedly observed in the Alleghany River at
Sandy Creek and Twelve Mile Island, but only old specimens of this form were
found. Young specimens generally vary more toward green, old ones toward tawny
or brown.

The color of the newly laid eggs ranges from sage-green (X, 15) to dark olive-
green (X, 18), or often to almost black. When somewhat advanced in development,

3 72 MEMOIRS OF THE CARNEGIE MUSEUM

the egg becomes in part prv/ne-purple (VIII, 1), in part cream-color (VI, 20), or

whitish.

Of this species, seven hundred and twenty-one specimens are at hand. Most of
them (six hundred and sixty-two) are from the state of Pennsylvania; fifty-seven
are from the "Panhandle" of West Virginia, and two from Maryland (Wills
Creek, Ellerslie). Many others have been collected, but no record has been kept,
since they were used for exchange, dissection, and experiment.

DISTRIBUTION.

(See Plate XLII, Figs. 2 and 3.)

LOCALITIES REPRESENTED IN THE COLLECTIONS OF THE CARNEGIE MUSEUM.

Pennsylvania : Greene Comity, Pennsylvania Fork of Fish Creek, Deep
Valley ; Smith Creek, Waynesburg ; Bates Fork, Deer Lick ; Pumpkin Run, Rice's
Landing; Fayettt ('mint;/, Cheat River, Cheat Haven; Youghiogheny River, Con-
nelsville ; Washington County, Buffalo Creek, Taylorstown ; Harmon's Creek, Dins-
more ; Raccoon Creek, Burgettstown ; Pigeon Creek and Taylor's Run, Mononge-
hela City; Beaver County, Beaver (S. N. Rhoads coll.); Raccoon Creek (Atkinson,
Williamson and Todd coll.); Little Beaver Creek, New Galilee (A. Kcenig coll.);
Brady's Run, Fallston ; Ohio River, Baden; Ohio River, Ambridge ; Lawrence
County, Eckles Run, Wampum; Big Run, Newcastle (1). C. Hughes coll.); Mercer
dm ut y, Otter Creek, Mercer; Shenango Creek, Hadley (0. E. Jennings coll.);
Crawford County, Shenango River, Linesville ; Shermansville (O. E. Jennings coll.);
Conneaut Outlet (D. C. Hughes coll.): Oil Creek, Spartansburg ; Erie County,
Conneaut Creek, Albion; Elk Creek, Miles Grove; French Creek, Union City;
Butler County, Tributary of Slippery Rock Creek, Branchton; Thorn Creek, Ren-
frew ; Rough Run. West Winfield; Allegheny County, Ohio River, Neville Island;
Ohio River, Bellevue (E. Hays and R. Taylor coll.); Ohio River, Shoustown ;
Flaugherty Run, Moon Township (Q. T. Shafer coll.); Chartiers Creek, Carnegie
(D. A. Atkinson coll.); Chartiers Creek, Bridgeville (D. A. Atkinson coll.); Turtle
Creek, Pitcairn (D. A. Atkinson coll.); Youghiogheny River, Boston (D. A. Atkinson
coll.); Crystal Lake, Pittsburgh (D. A. Atkinson coll.); Girty's Run, Millvale; Stone
Run, Thornhill ; Pine Creek, below Bakerstown Station (D. A. Atkinson coll.);
Alleghany River, Six Mile Island, (S. N. Rhoads and E. B. Williamson coll.);
Squaw Run, Aspinwall ; Alleghany River, Sandy < !reek : Alleghany River, Verona
(D. A. Atkinson coll.); Alleghany River, Twelve Mile Island ; Deer Creek, Harmar-
ville ; Little Deer ( Ireek, Russelton ; Pucketta Creek (D. A. Atkinson coll.); Little

ORTMANN: THE JRAWF1SHES OF THE STATE OF PENNSYLVANIA 373

Bull Creek, Tarentum (A. Koenig coll.); Alleghany River, Butler Junction ; West-
moreland Count;/, Kiskiminetas River, Livermore; Conemaugh Eliver, Blairsville
Intersection; Reservoir of McGee Run, Derry ; Whitethorn Creek, Dundale; small
tributary of Loyalhanna River, New Alexandria; Loyalhanna River, Ligonier ;
Loyalhanna River, Crisp; Indiana County, Two Lick and Yellow Creeks, Homer;
Crooked Creek, Creekside ; Little Mahoning Creek, Goodville ; Armstrong ('omit;/,
Long Run, Avonmore Station; Alleghany River, Kittanning; Alleghany River
and Pine Creek, Mosgrove ; Alleghany River, Templeton ; Clarion County, Alle-
ghany River, Red Bank; Jefferson Count;/, Pond at Punxsutawney ; Clearfield
Count;/, Sandy Lick Creek, Pu Bois ; Venango County, Alleghany River, Franklin;
Oil Creek, Oil City; Forest Count;/, Alleghany River, Tionesta; Warren County,
Brokenstraw Creek and Grouse Run, Garland ; McKean Count;/, Alleghany River,
Larabee ; Bedford County, "Wills Creek, Hyndman.

West Virginia: Hancock County, Harmon's Creek, Holidays Cove; Brooke
I 'ounty, Harmon's Creek, Colliers ; Ohio ( 'ounty, Wheeling Creek, Elm Grove ; Mar-
shall Count;/, Wheeling Creek, Union Township; Grave Creek, Cameron;
Pennsylvania Fork of Fish Creek, Nuss ; Wetzel Count;/. Fishing Creek, New
Martinsville.

Maryland : AUcghan;/ Count;/, Wills Creek, Ellerslie.

PREVIOUS RECORDS.
Type locality : New York, Monroe County, Genessee River, Rochester (Hagen),
New York : Cattaraugus Count;/, Alleghany River, Salamanca (Ortmann).
Pennsylvania: Westmoreland Count;/ (Faxon); Allegheny County (Williamson);
Warren Count;/, Alleghany River, Corydon (Ortmann).

ADDITIONAL LOCALITIES.

Material in the Department of Agriculture, Harrisburg, collected by W, R. McOon-
!,<■//. — Shenango River, Jamestown, Mercer County; French ('reck, Franklin,
Venango County ; small stream and pond, below Indiana, Indiana County ; Branch
of Genessee River, Ulysses, Potter County.

Alleghany River, Montrose, Allegheny ('ounty, Pa. (collected by the writer, but
material used for study); Indian < 'reek, .Jones Mills, Westmoreland County, Pa.,
(seen by the writer);14 Harmon's Creek, Hanlan, Washington County, Pa. (seen by
the writer); Ohio River, Congo, Hancock County, W. Va., (seen by the writer).

uThis locality was discovered in the beginning of the investigation by the writer, and since its importance was
then not understood, no specimens were preserved ; but the record is absolutely trustworthy.

374 MEMOIRS OF THE CARNEGIE MUSEUM

REMARKS.

Gambarus obsmrus is the river species of the Upper Ohio drainage. It is widely
distributed in western Pennsylvania. Compared with the allied species C. propin-
quus, which occupies a much wider area, it is rather uniform in its characters all
over its known range. It nowhere reveals a tendency to vary in the direction of
G. propinquus, or of propinquus sanborni. This is the more remarkable because C.
propinquus distinctly inclines toward this species in Erie and Crawford Counties, (in
the lake drainage), and likewise because C. propinquus sanborni shows such a ten-
dency in Wetzel County, West Virginia.

The variations observed in our abundant material have been briefly indicated
above. However, it deserves special mention that the specific characters are
scarcely subject to any variation.

Very interesting conditions are offered by the spines of the outer lower margin
of the meropodite of the cheliped. One or two spines may be present, the prox-
imal one smaller and often represented only by a small tubercle. Looking over our
material, I find that only one spine is present in all individuals from the upper
Alleghany drainage, including all the tributaries from Red Bank Creek northward
(sixty-one specimens are at hand). In Armstrong, Indiana, Westmoreland, and
Allegheny Counties, in the drainage of the Alleghany River, and in the whole
drainage of the Monongahela, the Beaver, and Ohio proper, a second spine may be
present, but such cases are not frequent, and generally this spine is found only on
one of the two chelipeds. There is a tendency of this character, more frequent!}'
displayed in the southwestern extremity of the range. Two such spines on either
side (right and left) are very rare, and I have found them only in twenty speci-
mens ; fifteen of which belong to the Ohio drainage : two to that of the Monongahela,
six to that of the Beaver, and seven to that of the Ohio below Beaver. Two cases
were discovered in Wills Creek, Maryland, and three in Conneaut Creek at Albion,
Erie County, Pa.

The latter specimens are interesting inasmuch as in Erie and ( 'rawford Counties
two drainage areas come together with that of Lake Erie, namely, that of the
Shenango River, a tributary of the Beaver, and that of French Creek, a tributary
of the Alleghany. In the latter creek and its tributaries I have never seen
an individual with two spines (seventeen specimens are at hand). Among the
material from the Beaver River drainage (fifty-six specimens) there are twenty-
one with two spines. Thus the tendency to develop two spines is markedly present
in the drainage of the Beaver, while it is apparently absent in French Creek.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 375

The specimens from the Lake Erie drainage in Conneant Creek quite often have
two spines (eight specimens out of twenty-two), and thus correspond to the Beaver
River form, and to those from Elk < Ireek, in which one specimen out of six has two
spines. Thus it appears that the form in the drainage of Lake Erie more closely
approaches the form found in the Beaver River than that found in French < Ireek,
although it must be granted that the material at hand seems to be not entirely
satisfactory, being somewhat too scanty from French Creek, and decidedly insuf-
ficient from Elk Creek.

A few freaks have come under observation in the following cases :

1. As has been said, the rostrum reveals in old individuals a tendency to a
shortening of the acumen. The extreme is reached in a specimen (male of the first
form) 74 mm. long, from Conneaut Outlet, Crawford County (D. C. Hughes coll.),
where the acumen is broadly triangular and hardly longer than the short marginal
spines, reaching only to the distal end of the basal joint of the peduncle of the
antennula. The acumen is well formed (not deformed), showing no traces of injury.
But that this specimen undoubtedly has been injured at some time earlier in its life, is
revealed by the fact that both claws are comparatively small, and by the characters
of regeneration (lack of spines on the outer lower margin of the meropodite, the
absence of a tubercle on the anterior margin of the lower side of the carpopodite,
and the generally weak and slender shape).

2. A female (46 mm. long) from Brokenstraw Creek, < Jarland, Warren ( Ymnty,
has the acumen of the rostrum directed obliquely to the left side, and the right
margin of the rostrum has five marginal spines. This seems to be due to an injury
received in earlier life. The left claw is also smaller and of the regenerated type.

3. A specimen (55 mm. long) from the Alleghany River at Sandy Creek (col-
lected by the writer, Nov. 19, 1904, Cat. No. 74. 479), has the characters of a female
in the shape of the chelre and the lack of hooks on the pereiopods. The annulus
ventralis, however, is very indistinct, although its outlines and slight median
depression are visible, as is also the median fissure. But this individual has the
male genital opening in the coxopodite of the fifth pereiopod, and the first pleopod
is of the male type, although small ; it is unusually short, reaching only to the
anterior margin of the coxopodites of the fourth pereiopods ; it is of the type of the
first form, with a distinct shoulder; the outer part is horny and distinctly longer
than the inner part. The second pleopods are entirely of the male type. Accord-
ing to the sexual orifice and the eopulatory organs, we are to regard this as a male
with certain female characters.

4. A pendant to the last specimen is one (67 mm. long) from the Ohio River,

3 76 MEMOIRS OF THE CARNEGIE MUSEUM

Neville Island, (collected by 1). A. Atkinson, May 14, 1899, Cat. No. 74. 36). The
claws are intermediate between male and female, but inclining toward the male
form. The third pereiopods have strong and well developed hooks on the ischio-
podites of the type of the first form male. The first pleopods are very peculiar,
(Plate XXXIX, Figs. Id and 7e), and unlike those of C. obscuruusj they rather
resemble those of G. limosus. Their length and strength are normal, but there is no
shoulder, and the two parts are separated only for a short distance at the tips, similar
to C. limosus, but the tips are not twisted. The outer tip is horny and pointed, the
inner soft, thicker, and tapers to a blunt point. The second pleopods are of the
normal male type. In addition this individual possesses a well developed annulus
ventralis, and sexual orifices only on the third pereiopods. Thus it appears to be a
female, with the secondary sexual characters of the male well, but not specifically,
developed.

None of the two cases of apparent hermaphroditism just described (Nos. 3 and 4)
agrees with any of the four cases mentioned by Faxon, (1885a p. 13, 14), or the four
described by Hay, (1905, p. 226 and 227). Additional cases will be described below
under G. bartoni. There is in the Carnegie Museum a further individual of herma-
phroditic character, namely a specimen of Cambarus rusticus Girard, from the
Wabash River, Bluffton, Indiana, collected by Mr. E. B. Williamson, June 1, 1905,
Cat. No. 74. 578. I append a description of it.

The specimen is externally a female, possessing the female type of claws, a well-
developed annulus, female sexual openings, and no hooks on the third pereiopods.
But the first pleopods are peculiar ; they are short and stout; the bases are iden-
tical with those of the male pleopods ; the distal parts, however, reach only to
about the middle of the coxopodites of the fourth pereiopods ; their tips are soft,
blunt, and slightly curved inward, and possess the furrow which divides them into
an outer and inner part, but these parts are not separated at the tips. The second
pleopods are of the female type. This case corresponds in a certain degree to the
second, third, and fourth, mentioned by Faxon, chiefly so to the third (in G. diog-
enes). The specimen is apparently a normal female, only the first pleopods are
transformed in a peculiar way, resembling the male type generally, but differing
from the specific shape. In the present case the first pleopod is different from
Faxon's case in detail.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 377

4. Cambarus (Baktonius) bahtoni (Fabricius).
vPlate 15, Fig. 1 ; Plate XXXIX, Fig. la-1/, and Fig. 8; Plate XL, Fig. 2.)

Astanis barioni Fabricius, 1798, p. 407 ; Say, 1817, p. 107 ; Harlan, 1835, p. 230. f. 3; Gould, 1841, p. 330 ; Thompson,

1842, p. 170 ; De Kay, 1844, p. 22, PI. 8, f. 25; Gibbes, 1850, p. 195, (partim).
Axtariis riiinris Raflnesque, 1817 p. 42.
Aslacus pusillus Raflnesque, 1817, p. 42.
Astaeus offlnia Milne- Edwards, 1837, p. 332 (turn Say).
Cambarus barioni Girard, 1852, p. 88; Bell, 1H59, p. 210; Hagen, 1870, p 75, PI. 1, f. 47-50, PI. 2, f. 135-139, PI. 3,

f. 106; Abbott, 1873, p. 80; Smith, 1874, p. 039 ; Putnam, 1874, p. 191 ; Faxon, 1884fc, p. 22; Faxon, 1885a, p.

59 ; Faxon, 1885i, p. 358 ; Underwood, 18S6, p. 367 ; Ganong, 1887, p. 74 ; Faxon, 1890, p. 622 ; Hay, 1896, p.

487, f. 6; Faxon, 1898, p. 649; Osburn and Williamson, 1898, p. 21 ; Williamson, 1899, p. 47; Hay, 1899, p.

959, 966 ; Williamson, 1901, p. 11 ; Ortmann, 1905n, p. 390 ; Paulmier, 1805, p. 134, f. 6 ; Rathbun, 1905, p. 18.
Cambarus pusillus and montanus Girard, 1852, p. 88.
Cambarus (Jiiirtonius) barioni Ortmann, 1905i, p. 120, 134.

Body robust, very sparsely pubescent in fresh, but perfectly naked in old speci-
mens, with only a few hairs on the Hngers of the chela?, and sometimes a slight
pubescence on the cutting edge of the fingers.

Carapace subovate, strongly depressed. G:H:B= 1:1.3 or 1.4:1.5 or 1.6.
Greatest width of branchial regions well forward, at a short distance behind the
cervical groove. Upper surface of carapace very flat.

Cervical groove deep, not interrupted on the sides. Areola distinctly longer than
half of the anterior section of carapace; a:p = 1 :0.6 Areola rather broad (>/■:/
= 1 : 5 or 0), with about 3-5 irregular rows of punctures.

Rostrum (Plate XXXIX, Fig. la-1/) broad and short, reaching generally to the
distal end of the second joint of the peduncle of the antennula, and hardly be-
yond the middle of the fourth joint of the peduncle of the antenna. Upper surface
almost flat or only slightly concave, but margins elevated, without marginal spines.
The margins converge more or less from the base, sometimes they are almost par-
allel, and near the apex they are suddenly contracted into a short, triangular acu-
men having a sharp point. The angles at the base of the acumen are rounded, but
generally well marked, and the elevated margins are continued to the apex, al-
though slightly decreasing distally from the lateral angles. Postorbital ridges short,
almost parallel, angulated anteriorly, but without spine, except in young
specimens.

Surface of carapacejpunctate, distinctly granulated on the hepatic region in larger
specimens. There are also a few more or less distinct granulations immediately be-
hind the cervical groove, hut no spine. External orbital angle well marked by an
angulation or a small tubercle, more rarely, and only in young specimens, spiniform.
Branchiostegal spine formed by a small tubercle, which is sometimes obsolete.

378 MEMOIRS OF THE CARNEGIE MUSEUM

Abdomen as long as carapace, or slightly shorter or longer ; it is slightly wider in
the female than in the male, but hardly wider than the carapace in the former.
Anterior section of telson on the posterior lateral corners generally with two, more
rarely with three spines. Posterior section semi-elliptical, distinctly wider than
long, slightly shorter than anterior section.

Epistoma with posterior part broad and short, about two and a half times as
broad as long, with a distinct transverse groove on either side slight^ posterior to
the middle, and an anterior median depression. Anterior section constricted at
the base, semi-circular, with a median anterior point. This point may be strongly
developed, or almost entirely absent. Transverse diameter distinctly greater than
the longitudinal.

Antmnula with a small, often spiniform, tubercle on the lower margin of the
basal joint.

Antennal peduncle with a tubercle on the outer side of the first joint, which is
often spiniform, chiefly so in }^oung specimens ; second joint with or without a very
indistinct tubercle.

Antennal scale short and narrow, slightly longer than the rostrum, reaching to,
or almost to, the end of the fourth joint of the antennal peduncle. Spine of outer
margin strong. Laminar part not much broader than the marginal spine.

F/agellum reaching to the anterior margin or to the middle of the telson in the
male, slightly shorter in the female, but sometimes considerably shorter, without
apparent trace of having been injured. In some cases it reaches only the middle of
the second abdominal segment.

First pereiopods (Plate XL, Fig. 2) very strong and robust in old individuals,
particularly males. Hand elongate-ovate, broad, and strongly depressed. Surface
punctate. Inner margin of palm short, curved, with a single marginal row of more
or less distinct, low tubercles. Outer margin smooth, rounded proximally, carinate
distally. Fingers longer than palm, not gaping in young individuals, but with a
wide gap at the base, meeting only at the tips, most noticeably in old males.
Outer margin of movable finger punctate, or, in older specimens, with a few indis-
tinct tubercles. Cutting edges with tubercles, larger in the proximal part. Upper
surface of each finger with a low longitudinal rib, bordered by rows of punctures.
This rib often becomes indistinct, especially on the movable finger in old males.

Carpopoditc slightly longer than wide, shorter than palm, with a deep longitud-
inal sulcus above. Inner margin with a strong pointed or blunt spine, which is
generally distinctly hooked, going off almost at a right angle, but curving forward
in the distal part. A small spine or tubercle (sometimes double) may be added to

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 379

it proximally. Lower surface with a blunt conical tubercle in the middle of the
anterior margin (occasionally spiniform). The tubercle at the articulation with the
hand is generally obsolete. There are sometimes additional tubercles ; the one
which most frequently occurs is a small spine or tubercle between the large one on
the inner margin and that on the anterior margin of the lower side.

Meropodite smooth, with 1-3 tubercles near the distal end of the upper margin,
one of which is often spiniform in young specimens; in old specimens they are
generally very indistinct or wanting. Lower side with two rows of spiniform
tubercles. The outer rows consist of 1-6 (very rarely only one tubercle). Six were
found in only one instance, that of a regenerated cheliped. Generally there are
two or three. The inner row has 6-11 spiniform tubercles, of which the distal is
the largest. A small tubercle on the outer articulation with the carpopodite may
be present or absent.

Ischiopodite of third pereiopod hooked in the male. The hook of the first form
is strong and subcorneal.

The coxopodite of the fourth pereiopod in the male possesses a prominent rounded
and compressed tubercle.

First pleopods of the male of the first form (Plate XXXIX, Fig. 8) stout and
short, reaching to the posterior margin of the coxopodite of the third pereiopods.
They are not articulated at the base. The two parts are separated at the tips for a
short distance, and both are curved sharply backward, forming almost a right angle
with the basal part. Distally they are partly twisted, so that the outer part is
directly anterior to the inner. The outer part is horny, compressed, falciform, the
tip pointed, with a small posterior accessory point (often worn off). The inner
part is soft, swollen at the base, and suddenly tapering to a blunt point.

In the male of the second form this organ may be articulated at the base (in the
case of the young) or not articulated (in older specimens). Both parts are separated
distally for a short distance, and the outer part is soft, not horny, less distinctly
compressed, and blunt. In the young these organs are considerably shorter than in
older specimens.

Anmilus v&ntralis of the female transversely rhonihiforin, with a deep central
depression and a longitudinal S-shaped fissure. Anterior and more particularly the
posterior margins elevated. The whole anterior portion of the annulus often appears
depressed compared with the elevated posterior margin. Where the longitudinal
fissure passes over the posterior margin the latter is slightly depressed. In young
females the central depression is less marked, and the margins are consequently less
elevated, giving a rather flat appearance on the annulus.

380 MEMOIRS OF THE CARNEGIE MUSEUM

Size. — This species in western Pennsylvania reaches a considerable size, although
the maximum recorded by Hagen (3.6 = 91 mm.) has not been observed. Faxon,
(1885a, p. 64), mentions a specimen from the Mammoth Cave, Ky., measuring 108
mm., but this is not the typical form. The largest individuals in the Carnegie
Museum are two females, the one from Braeburn, the other from Deny, Westmore-
land County, both measuring 87 mm. in length. The largest male (first form) is
from North Versailles Township, Allegheny County, and measures 83.5 mm.15 In
western Penns3dvania specimens over 80 mm. are not altogether rare.

In the eastern portions of the state this species is much smaller. The largest
specimen at hand is a female from Roxboro, 67 mm. long, and a male (first form)
from Manayunk, Philadelphia County, 66.5 mm. long, (both collected by H. Cera).
Specimens over 60 mm. are not frequent in eastern Pennsylvania.

Colors. — Generally dull and not much varied, greener in young specimens,
browner in old ones. (See Plate B, Fig. 1.)

The carapace ami abdomen olive-green (Ridgway, 1886, X, 18) to tavmy-olive
(III, 17), chestnut (IV, 9), and burnt umber (III, 8), a shade darker dorsally, lighter
on the sides. Margins of rostrum, in the browner specimens, fcrrugincous (IV, 10).
Distal third of finger rufous (IV, 7), or tawny (V, 1). Tubercles of the cutting edges
of fingers ochraceous buff (V, 10). In brown individuals there is generally some
green on the chelae.

Aside from young individuals, where the normal olive-green prevails, this species
shows a distinct tendency toward the brown and chestnut shades, more so than the
river species, C. limosus, G. propinquus, and C. obscurus.

In some cases the colors are brighter. Individuals shading to a copper-color are
not rare, and I have seen a few where a dirty slate-blue was the ground-color. Of
course, as in other species, in old specimens the original colors are largely obscured
by a deposit of mud, rendering the specimens sometimes almost black.

In very young specimens (10 to 20 mm. long) the color is olive-green, semitrans-
parent, with the chelse almost entirely ferrugineous.

The color of the newly laid eggs is almost black, with, or without, a purplish hue
(indian-purple, VIII, 6). In a more advanced stage they become particolored:
prune-pnrple (VIII, 1) or dahlia-purple (VIII, 2) on one side, grayish or whitish on
the other.

The Carnegie Museum possesses seven hundred and fifty-five specimens of this
species, six hundred and fourteen of which are from the state of Pennsylvania,

15 The female from Hill, Westmoreland County, mentioned previously (Ortmann, 1905n, p. 391) is 85 mm., not
89 mm. as stated ; the male from Cheat River (ibid. ) is not 92 mm., but 82 mm. in length.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 381

nineteen from New York, ten from New Jersey, six from Ohio, seventy from West
Virginia, thirty-four from Maryland, and three from North Carolina.

DISTRIBUTION.

Localities represented in the collection of the Carnegie Must um.

Pennsylvania: Delaware County, Dicks Run, Wallingford ; Philadelphia ''mint;/,
Manayunk, (H. Gera coll.); Domino Lane Run, Roxboro. (H. Gera coll.); Wissa-
hickon ; Bucks County, Grenoble; Dark Hollow Run, New Hope; Northampton
County, Bushkill Creek, Easton, (A. E. Davison coll.) ; Lehigh County, Little Lehigh
Creek, Emaus; Montgomery County, West Manayunk, (H. Gera coll.); Bucks County,
Shoemakersville ; Chester County, Valley Forge; Lancaster County, Pequea ; York
County, Arthur Run, York Furnace; Dauphin County, Susquehanna River, Halifax ;
Northumberland County, Georgetown ; Franklin County, Dickey; Williamson; Ful-
ton County, Dogtown ; Big Cove Creek, McConnellsburg ; Blair Comity, Frankstown
Branch of Juniata River, Loop near Hollidaysburg ; Bedford County, Bedford
Springs (A. Koenig coll.) ; Cameron County, Sinnamahoning < 'reek, Driftwood ; Sin-
namahoning ; Cambria ( 'ounty, Tributary of ( ilearfield < 'reek, Ashville ; Headwaters
of Clearfield Creek, Cresson; Summit, (S. N. Rhoads coll.); Laurel Run, Lovett ;
Somerset County, Wills Creek, Mance; Flaugherty Creek and tributaries, Sandpatch :
Casselman River, Rockwood ; Windber; Laurel Hill, west of Jennerst<>wn ; Indiana
(.'on id;/, Cush-( 'ushion ( Jreek, west of Cherry Tree ; Homer ; < Ireekside ; < Joodville :
Jefferson County, Mahoning ( 'reek, Punxsutawney ; Brockwayville ; Brookville ;
Clearfield Chunty, Falls Creek; Elk County, Elk ('reek, Ridgway; Potter County,
Keating Summit ; McKean ('mint;/, Larabee; Warren County, Grouse Run. Garland ;
Forest County, Tionesta; Venango County, Sage Run. Oil city; Clarion County,
Alleghany River, Red Bank; Armstrong County, hong Run, Avonmore Station;
Weskit, opposite Kittanning ; Pine Creek, Mosgrove; Alleghany River, Templeton ;
Westmoreland County, Tub Mill Run, Ross Furnace, South of New Florence; Crisp
(H. H. Smith and M. A. Wertheimer coll.); Lynn's Run, Mechanicsburg ; Loyal-
hanna River, Ligonier ; Indian Creek, Jones Mills; Reservoir of McGee Run.
Derry; Withethorn Creek, Dundale ; Li verm ore ; Hill, opposite Leechburg ; Brae-
burn; Fayette County, Youghiogbeny River, Ohiopyle ; Jacobs Creek, Laurelville;
Dunbar; Cheat Haven ; Allegheny County, Alleghany River, Butler Junction ; Little
Bull Creek, Tarentum, (A. Koenig coll.); Deer Creek, Harmarville ; Little Deer
Creek, Russelton ; Power's Run. Montrose; Squaw Run, Aspinwall; Verona. (D. A.
Atkinson coll.) ; Quigley's Run, Verona ; Breakneck Run. Bakerstown Station ; Pine
Creek, below Bakerstown Station, (D. A. Atkinson coll.); Stone Run, Thornhill ;

382 MEMOIRS OF THE CARNEGIE MUSEUM

Girtv's Run, Millvale ; Westview (D. A. Atkinson coll.); Avalon ; Edgeworth, (G.
H. Clapp coll.) ; Schenley Park, Pittsburgh, (E. B. Williamson coll.) ; Fern Hollow,
Pittsburgh ; Edgewood Park, Swissvale ; North Versailles Township, opposite Stew-
art; Jacks Run, South Versailles Township ; Boston, (D. A. Atkinson coll.) ; Thomp-
son's Run, Kennywood, (F. E. Kelly coll.) ; Carnegie, (D. A. Atkinson coll.) ; Moon
Township, (D. A. Atkinson, B. Graf, E. B. Williamson, A. T. Shafer, Q. T.
Shafer coll.); Thorn's Run, Moon Township; Flaugherty Run, Moon Township,
(Q. T. Shafer coll.) ; Butler County, West Winfield ; Renfrew ; Slippery Rock
Creek, Branchton ; Erie Count//, Elk Creek, Girard; Walnut Creek, Swanville ;
Crawford County, Spartansburg ; Linesville ; Mercer County, Stoneboro, (D. A.
Atkinson) ; Mercer; Lawrence County, Wampum ; Big Run, Newcastle (D. C. Hughes
coll) ; Beaver County, Ambridge ; Baden ; Beaver, (S. N. Rhoads coll.) ; Brady's Run,
Fallston ; Smith's Ferry; Monaca ; Washington County, Monongahela City; West
Brownsville ; Francis Mine, near Burgettstown ; Taylorstown ; Greene County, Rice's
Landing ; Bates Fork, Deer Lick ; Waynesburg ; Deep Valley.

New York : Herkimer County, East Canada Creek, Dolgeville, (R, Ruedemann
coll.).

New Jersey : Mercer County, Princeton.

Maryland : Washington County, Home's Valley, (F. Silvester coll.) ; Alleghany
County, South Cumberland ; Corriganville ; Rawlings ; Garret Coun ty, Selbysport ;
Stoyer.

West Virginia : Morgan County, Cherry Run ; Tucker County, Blackwater
River, Davis; Shavers Fork, Parsons; Monongalia County, Cheat River, (H. H.
Smith coll.); Morgantown ; Pleasants County, St. Mary"s ; Wetzel County, New
Martinsville; Marshall County, Cameron; Ohio County, Elm Grove; Brooke
('mi nty, Colliers; Hancock County, Holidays Cove ; Congo.

Ohio: Harrison County, Bowerstown ; Carroll County, New Hagerstown.

North Corolina : Watauga County, Blowing Rock, (Wilcox coll., Exch. Acad.

Nat. Sc. Philad.).

PREVIOUS RECORDS.

Type Locality : North America, (Fabricius) ; Philadelphia, Pa., (Harlan).

( 'anada : Falls of Ouiatchouan, Lake St. John, Quebec, (Bell) ; Metis and Mata-
pediac Rivers, Quebec, (Bell); Montreal, Quebec, (Faxon); Restigouehe River,
New Brunswick. (Bell); Upsalquitch River, New Brunswick, (Ganong); Miramichi
River, New Brunswick, (Ganong) ; St. John, New Brunswick, (Faxon) ; St. John
River, Grand Falls to Fredericton, New Brunswick, (Ganong).

Maine: Houlton and Maysville, Aroostock County, (Faxon) ; Outlet of Moose-
head Lake, Piscataquis County, (Faxon) ; Madison, Somerset County, (Faxon).

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 383

Vermont : (Thompson) ; Burlington, Colchester, and Shelburne, Chittenden
County, (Faxon).

Massachusetts : (Gould) ; Williamstown, Berkshire County, (Faxon) ; North
Adams, Berkshire County, (Faxon) ; North ( 'ration, Worcester County, (Faxon).

New York: Lake Champlain, (Rafinesque); Ellenburg, Clinton ( lounty, ( Faxon);
Elizabeth town, Essex County, (Hagen) ; Westport, Essex County, (Faxon); Lake
George, Warren County, (Rafinesque) ; Saratoga County, (Rafinesque) ; Fishkill,
Dutchess County, (Rafinesque) ; Newburgh, Orange County, (Rafinesque) ; Port
Jervis, Orange County, (Faxon) ; New York City, (Paulmier) ; Fallsburg, Sullivan
County, (Faxon); Fulton Lakes, Hamilton and Herkimer Counties, (Faxon); Can-
ton, St. Lawrence County, (Faxon) ; Utica, Oneida County, (Rafinesque) ; Oswego,
Oswego County, {Rafinesque) ; Cazenovia, Madison County, (Faxon) ; Sherburne,
Chenango County, (Faxon) ; Berkshire, Tioga County, (Hagen) ; Ithaca, Tompkins
County, (Faxon) ; Rochester, Monroe County, (Faxon); Niagara, Niagara County,
(Hagen) ; Forestville, Chautauqua County, (Faxon).

New Jersey: Schooley's Mountain, Morris County, (Hagen); Orange, Essex
County, (Faxon) ; Trenton, Mercer < lounty, (Abbott) ; Princeton, Mercer County
(Ortmann).

Pennsylvania: Windham, Bradford County, (Faxon) ; Headwaters of Loyal-
sock Creek, Sullivan County, (Ortmann) ; Ganoga Lake, Sullivan County, (Ort-
mann); Berwick, Columbia County, (Girard) ; Schuylkill River, Philadelphia,
(Hagen); ( Ihester ( lounty, (Faxon) ; Bainbridge, Lancaster ( lounty, (Faxon) ; Hum-
melstown, Dauphin County, (Faxon); Carlisle, Cumberland County, (Girard);
Pinegrove, Cumberland County, (Ortmann); McKean County, (Faxon) ; Foxburg,
Clarion County, (Girard); Westmoreland County, (Faxon); Pittsburgh, Allegheny
County, (Williamson); Bedford and Loysburg,18 Bedford County, (Faxon).

Delaware: Greenville, New Castle County, (Ortmann).

Maryland : Harford ( lounty, ( Faxon) ; Howard < lounty, (Faxon) ; Montgomery
County, (Faxon); Frederick County, (Faxon); Washington County, (Faxon);
Cumberland, Alleghany County, (Girard); Garrett County, (Faxon).

District of Columbia : Georgetown, (Hagen); Washington, (Faxon).

Virginia: Alexandria County, (Faxon); Clarke County, (Faxon); Stafford
County, (Faxon); James River, (Faxon); Franklin, Southampton County, (Faxon);
Lunenburg, Lunenburg County, (Faxon); Waynesboro, Augusta County, ( Faxon );

16 Faxon cites " Pattonville, Bedford County." The name of Pattonville lias been changed to Loysburg, which
the hamlet originally bore. It is situated in the valley known as Morrison's Cove, a beautiful spot full of clear moun-
tain streams, formerly abounding in brook-trout. The locality must not be confounded with Pattonville in Delaware
County, (P. O. Fern wood).

384 MEMOIRS OF THE CARNEGIE MUSEUM

Rockbridge County, (Girard); Bath County, (Faxon); Pulaski, Pulaski County,
(Faxon); Wytheville, Wythe County, (Faxon); Smith County, (Faxon).

North Carolina : Kinston, Lenoir County, (Faxon) ; Newman's Fork, Blue
Ridge, McDowell County, (Faxon); Black Mountain, McDowell County, (Faxon);
Waynesville, Haywood County, (Faxon); Roan Mountain, 6,000 feet, (Faxon).

Tennessee: Doe River, Carter County, (Faxon); Claiborne County, (Faxon);
Monroe County, (Faxon); McMinn County, (Faxon).

Kentucky : Kentucky River, Hickman's Landing, (Hagen)17 ; Cumberland Gap,
Bell County, (Faxon); Smoky Creek, Carter County, (Faxon); Little Hickman,
Jessamine County, (Faxon); Albany, Clinton County, (Faxon); Grayson Springs,
Grayson County, (Faxon); Mammoth Cave, Edmonson County, (Hagen).18

West Virginia : Patterson Creek, (Faxon); South Branch of Potomac River,
(Faxon); Williamsport, Grant County, (Faxon); Glade Creek, Randolph County,
Faxon); Petroleum, Ritchie County, (Faxon).

Ohio : Marietta, Washington County, (Faxon); Tuscarawas County, (William-
son); Knox County (Williamson); Licking County, (Williamson); Columbus, Frank-
lin County, (Hagen); Alum Creek, and tributaries of Big Walnut and Big Darby,
Franklin County, (Osborn and Williamson); Yellow Springs, Greene County,
(Faxon); Warren County, (Faxon); Cincinnati, Hamilton County, (Hagen).

Indiana19: New Albany, Floyd County, (Faxon); Cave near Paoli, Orange
County, (Hay); Down's and Connelly's Cave, Lawrence County, (Hay); Blooming-
ton, Monroe County, (Faxon); Clear Creek, Monroe County, (Hay); May's Cave,
Monroe County, (Hay); Indianapolis, Marion County, (Faxon); Irvington, Marion
County, (Hay).

The locality " Lake Superior" (Hagen) has been dropped, since it is, no doubt,
erroneous, (see Ortmann, 1905, p. 135); the same is the case with "Osage River,
Missouri " (Hagen).

ADDITIONAL NEW LOCALITIES.

New York : Altamont, Albany County, (N. Y. State Museum) ; Mill Creek,
Wilmurt, Herkimer County, (N. Y. State Mus.) ; Spencerport, Monroe Co., (Mus.
Oberlin).

Pennsylvania :

Specimens preserved in the Academy of Natural Sciences of Philadelphia : 5717 Ger-
mantown Ave., Philadelphia ; small stream near Holmesburg, Philadelphia County;

17 Not located by writer.

18 The form from Mammoth Cave is not typical, according to Faxon and Hay.

19 Specimens from Indiana differ from the typical form, according to Faxon and Hay.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 385

Roberts Run, Abrams, Montgomery County ; Port Alleghany, McKean County;
Corydon, Warren County.

Specimens preserved in the Department of Agricultv/re, Harrisburg: Buckingham,
Bucks County; Avondale, Chester County; Highspire, Dauphin County; Harris-
burg, Dauphin County; Rockville, Dauphin County; Dauphin, Dauphin County;
Marshrun, Perry County ; ( tettysburg, Adams County; Montalto, Franklin ( lounty;
Emniaville, Fulton County; Huntingdon, Huntingdon County; Williamsport,
Lycoming County; New Albany, Bradford County; Wellsboro, Tioga County.

Mr. W. R. McGomiell collected this species at the following localities : Stream flowing
out of Beach Lake, Wayne County; small stream tributary to Delaware River,
Portland, Northampton County; stream flowing into Lehigh River, Slatington,
Lehigh County; Schuylkill River, Reading, Berks County; Toby's Creek, Kings-
ton, Luzerne County ; Fish Creek, near Stillwater " fifteen miles above Blooms-
burg," Columbia County ; Montour Run, Greenpark, Perry County ; Big Buffalo
Creek, Erly, Perry County; Conococheague Creek, Chambershurg and Marion,
Franklin ( lounty ; tributary of Conococheague Creek, Mercersburg, Franklin
County; Laurel Run and Shafer's Run (probably Shaver's Creek, both in north-
eastern part of county), Huntingdon County; Slab Cabin Creek and Thompson's
Spring, State College, Center County; Bear Meadows and branch of Spring Creek,
Boalsburg, Center County; Sinking Creek, Center Hall, Center County; Bald
Eagle Creek and Wall is Run, Milesburg, Center County; Beech Creek, Beech
Creek Station, Clinton County; Fishing Creek and tributary, Lamar, Clinton
County; Nipponose Creek, Jersey Shore, Lycoming County; branch of Genessee
River, Ulysses, Potter ( lounty.

lite writer has seen this specie* at and from the following localities:

Lafayette, Montgomery County, (H. Gera coll.); Leopard, Easttown Township.
Chester County, (J. F. Sachse, Philadelphia, coll.) ; Wills Creek, Hyndman, Bed*
ford County; Big Meadow Run, Ohiopyle, Fayette County ; West Branch of Sus-
quehanna, Cherry Tree, Cambria and Clearfield Counties.; Blairsville Intersection,
Westmoreland County; Springs on Chestnut Ridge, near Deny, Westmoreland
County, elevation 1800 feet; Coalpit Run, Millbank, Westmoreland County; Done-
gal, Westmoreland County; Jeanette, Westmoreland County; Alleghany River,
Hulton, Allegheny County; tributary of Thompson's Bun, south of North Bes-
semer, Allegheny County; Sandy Creek and Alleghany River, Sandy Creek, Alle-
gheny County ; Nine Mile Bun, Pittsburgh, Allegheny County ; Dinsmbre, Wash-
ington County ; Summit and ( lonrieautville Station, Crawford ( lounty.

Maryland: Sideling Creek, Washington County, (H. A. Pilsbry coll., Acad.

386 MEMOIRS OF THE CARNEGIE MUSEUM

Nat. Sc. Phil.); Town Creek, Alleghany County, (H. A. Pilsbry coll., Acad. Nat-
Sc. Phil.) ; Deer Park, Garrett County, (P. R. Uhler coll.). (See below under G.
diogenes, footnote 26.)

REMARKS :

Cambarus bartoni is the crawfish of the small streams in Pennsylvania, and is
exceedingly abundant all over the state.

In spite of its wide distribution over parts considerably different in physical con-
ditions, this species is in Pennsylvania very uniform with regard to its morpholog-
ical characters (disregarding the variety robuslm, to be discussed below). It is true
that in the foregoing description many characters are pointed out which vary within
certain limits ; but these variations are not restricted to certain parts of the state,
but occur everywhere. It is even hard to say of any character that itf prevails in a
certain region. In general there are indications that the species is more flourishing
and also more variable in the western part of the state than in the eastern. This
observation, however, applies chiefly to characters which appear in very old speci-
mens, as for instance, a stronger development of the tubercles on the inner margin
of the hand and on the outer margin of the movable finger. Since the eastern form
is much smaller, such characters, which are only occasionally present in very large
specimens, are not found in specimens from the east, namely, strong sculpture of the
hand and very thick margins on the rostrum.

The most variable feature of this species is the rostrum. Generally it is narrower
and more gradually tapering in very young specimens (PI. XXXIX, Fig. \d and le).
In older individuals it becomes broader, and is more suddenly constricted into a
longer or shorter acumen. Beyond this there is no rule. The most frequent shapes
are those figured on PL XXXIX, Fig. lb and lc. The one delineated in Fig. If is
exceptional. The other extreme is shown in Fig. 1«, with margins practically par-
allel, and a very sudden constriction into a comparatively short and broad acumen.
Although this last shape is more frequent in the western part of the state, it is also
found in the extreme eastern portions of the commonwealth.

There is only one character in which regional variation may be observed, and
this is the size of the body. As has been mentioned, in the eastern part of the state
this species is considerably smaller than in the western, and the largest specimens
are found west of the Chestnut Ridge. Individuals 80 mm. and more in length are
not rare in Westmoreland, Allegheny, Elk, and Lawrence Counties. Specimens
between 70 and 80 mm. long have been found, in addition to the counties just
named, in Crawford, Venango, Potter, Jefferson, Butler, Armstrong, Washington, and

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 381

Fayette. All of these are west of the Chestnut Ridge. East of the Chestnut Ridge,
but west of the Alleghany Front, specimens over 80 ram. in length are rare; only
one was found at Sand Patch. Somerset County, (86 mm. long). Specimens over

70 mm. in length are not infrequent here. In the Alleghany Mountain region,
(between the Alleghany Front and the Blue Ridge) only a few instances of speci-
mens over 70 mm. in length have been found, and none as large as 80 mm. The
largest is from McConnelsburg, Fulton County, 77 mm. East of the Blue Ridge

(South Mountain) the length 7(1 mm. is never attained. In the easternmost extrem-
ity of the state (Northampton and Bucks Counties) even the length of 60 mm. is
not represented among my material, although I possess large series of specimens
from this region. Thus it appears that the size gradually decreases from west to
east. There is hardly any appreciable decrease in size from south to north. The
smaller number of large individuals from the northwestern section of the state is
very likely due to the fact that large collections were not made in that part of the
state.

Freaks have been observed in several cases. Aberrant forms of the rostrum
have been repeatedly found, and one variation has been encountered four times in
which the rostrum has a very slightly developed acumen, so that it is almost evenly
rounded off anteriorly, with only a small and indistinct median angle or point,
(female, 25 mm., Schenley Park, Pittsburgh ; female. 50 mm.. Templeton, Arm-
strong County ; male, first form, 64 mm., Branchton, Butler County ; female, 80 mm.,
Monongahela City, Washington County). The rostrum, in these cases, is exception-
ally short, due to a reduction of the acumen. A case of an unsymmetrical rostrum,
with the left angle at the base of the acumen cut off, has also been observed ; this
is clearly a malformation due to some previous external injury.

We may class with the freaks a single specimen in which the carapace possesses
a lateral spine. The specimen is a female (42 mm. long) from Weskit, Armstrong
County, and it has a small, sharp, lateral spine, but only on the left side of the
carapace. This is the more remarkable since it demonstrates the importance of this
specific character. There is not a single other individual among the large material
at hand which possesses such a spine, although granulations in its place are not
infrequent.

Two interesting cases of abnormally developed sexual characteristics have been
noticed (compare the other cases mentioned under C. i Faxonvus) obscurus).

1. A large female, 71 mm. long, found March 31. 1905, at Hollidays Cove, Han-
cock County, W. Va., (Cat. No. 74. 491), which is normal in every respect but one,
and besides, is undoubtedly sexually normal, since it carried under the abdomen

388 MEMOIRS OF THE CARNEGIE MUSEUM

ten young ones, ready to leave the mother, (very likely a number had left already
when the mother was captured). It has on the ischiopodite of the left third perei-
opod the copulatory hook of the male ; this hook is not small or rudimentary, but
strong, and similar to the hook as found in the male of the first form. The ischi-
opodite of the corresponding right pereiopod has no trace of this hook.

2. A specimen, 48 mm. long, was found in Fern Hollow, Pittsburgh, November,
22, 1905, (Cat. No. 74. 681), which externally (in the shape of the claws) looks like
a female, but shows very indistinctly the sexual openings of the male, and no traces
of those of the female. It also has the first pleopods of the male of the second form,
but the second pleopods are built according to the female type. Further, it lacks
entirely the hooks of the third pereiopods, and has a distinct female annulus, of
juvenile type.

This case does not correspond exactly to any of those described previously. It
resembles to a certain degree one of the cases in ('. obscwrus described above (No. 3),
with the exception that here the first pleopods are of the type of the male of the
second form, and that the second pleopods are not of the male, but of the female type.

4a. Cambarus (Bartonius) bartoni robustus (Girard).
(Plate B, Fig. 2. Plate XXXIX, Fig. 2a and 2b. Plate XL, Fig. 3.)

Cambarus robustus Girard, 1852, p. 90 ; Hagen, 1870, p. 80, PI. 3, 1. 167 ; Smith 1874, p. 639 ; Faxon, 18846, p. 143.
Cambarus bartoni robustus Faxon 1885a, p. 01 ; Faxon, 18856, p. 358; Underwood, 1886, p. 367; Faxon, 1890, p. 622;

Faxon, 1898, p. 649 ; Osburn & Williamson, 1898, p. 21 ; Williamson, 1899, p. 20. 47 ; Hay, 1899, p. 959, 966 ;

Williamson, 1901, p. 11 ; Ortmann, 19(>5a, p. 391 ; Ortroann, 19056, p. 135.
Cambarus bartoni Williamson, 1905, p. 310.

The differential characters of this form are the following :

Body robust, attaining decidedly a more considerable size than the typical bar-
toni. The largest individuals at hand are a male, first form, from Spartansburg,
( 'rawford County, measuring about 98 mm. (estimated, since rostrum is damaged) ;
a female from Squaw Run, Allegheny Count)', measuring 94 mm. and a male,
second form, from Puketta Creek, Allegheny County (A. Kcenig coll.), measuring
101 mm. I have quite a number of males (of the first and second form) and of
females over 90 mm. long. It is also remarkable that specimens of this variety, of
a considerable size (60 to 70 mm.) display characters which are distinctly juvenile,
showing no tendency on the part of the chelae to attain a large size. This tendency
is also evidenced by the fact that the smallest males of the first form at hand are two
individuals measuring 72 mm. (Union City and Hulton). From Oberlin, Ohio, I
have seen a male of the first form, 71 mm. long, while the minimum size of sexually
ripe males of the typical form is 50 mm. for western, and 49 mm. for eastern
Pennsylvania.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 389

The shape of the carapace is similar to that of C. bartoni, but, in old specimens,
appears slightly more depressed on account of the wider hepatic and branchial re-
gions ; G:H:B = 1:1.3 to 1.5:1.5 to 1.7.

The rostrum (Plate XXXIX, Figs. 2a and 2b) is markedly different from the
typical form, narrower, more tapering, and with a longer acumen ; the upper sur-
face is slightly concave. The shape is rather variable. In young specimens of the
typical C. bartoni the rostrum often approaches the form of ('. ba/rtoni robustus, hut
on the other hand young specimens of the latter have a narrower rostrum than the
former. (( lompare Plate XXXIX, Fig. hand 2/;). In many cases the rostrum of the
variety is not longer than in the typical form, hut in others it surpasses it, reaching
to the middle, or even almost to the end of the third joint of the peduncle of the
antennula, or to the base or almost to the middle of the fifth joint of the peduncle
of the antenna.

Areola similar to that of the typical form, but with a larger number of rows of
punctures (4-6), which is due to the punctures being more crowded, not only on
the areola, hut on the whole carapace.

The carapace is often provided with small and sharp lateral spines ; they are
sometimes obsolete, or replaced by tuhercles, or even entirely absent.

Antenna! scale generally slightly wider than in the typical form, and slightly
longer.

Antennal flagellum not differing greatly from that of the typical form, and quite
variable in length. There are a few cases where it reaches to the end of the telson,
and even slightly beyond, thus surpassing any case known in C. bartoni.

The first pereiopods (Plate XL, Fig. 3), display remarkable and important differ-
ences from the typical form. The hand has nearly the same shape, hut the fingers
are less gaping, and meet all along their edges even in individuals of a considerable
size (retention of juvenile character); it is only in very large specimens that they
are distinctly gaping, hut less so than in much smaller individuals of ('. bartoni.
The sculpture of the hand is much more strongly developed. The inner margin of
the hand has a distinct and regular double row of tubercles. This double row is a
very important character, and is noticeable in specimens from the size of about 30
mm. upward. In very young individuals it is obsolete, and hecomes more and
more distinct with advancing age. Fvery specimen at hand, without excep-
tion, possesses this character, when the cliche are normally developed ; but it must
be mentioned that in regenerated chelre, which are always recognizable by their
shape, this double row is sometimes indistinct or irregular. Further, there is a tri-
angular depression both on the upper and lower side of the hand at the base of the

390 MEMOIRS OF THE CARNEGIE MUSEUM

immovable finger. Both impressions are always present in large individuals; in
younger ones they are indistinct, but are always marked by punctures, which are
much crowded, and consequently by the denser hairs implanted in them. Traces
of the impression on the upper surface are often seen in the typical form, but that
of the lower surface is always absent, or marked only by a slight flattening of the
surface. These two impressions give to the hand of this variety a very strongly
marked marginal keel or ridge. The outer margin of the movable finger possesses
a number of irregularly placed tubercles, indistinct, and restricted to the proximal
part in young specimens, but very distinct, and occupying about two-thirds or
three-fourths of the margin in old specimens. The longitudinal ribs of the upper
surface of the fingers are always well developed, and there is hardly any tendency
in older specimens for them .to become obscure, chiefly in the case of the immov-
able finger, where this rib is always well marked on account of the strongly devel-
oped punctures of the depression accompanying it on the outside.

The armature of the carpopodite and the meropodite is almost identical with that
of the typical form, but the carpopodite in old individuals is often provided with
accessory low tubercles on the upper face. The spines of the meropodite are more
distinct and more numerous ; those on the distal upper margin (generally two of
them) well developed, even in large individuals; those of the lower margin consist-
ing of two to six in the outer row (two are rare, found only in young ones ; in regen-
erated chelas as many as eight); and seven to twelve in the inner row (as many as
fourteen in regenerated chela*).

All the other characters, including the color (see Plate B, Fig. 2), agree with the
typical form. The color of the eggs (in the only specimen ever found with eggs, at
Spartansburg) is prune-purple (VI 11, 1 ), almost black.

There are in the collections of the Carnegie Museum one hundred and forty-
seven specimens of this form, all of which are from Pennsylvania, with the excep-
tion of four, which are from Kentuck}\

DISTRIBUTION.
LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.
Pknnsvlvania : Allegheny County, Chartiers Creek, Carnegie. (D. A. Atkinson
coll.); Pine Creek, below Bakerstown Station, (D. A. Atkinson coll.); Squaw Run,
near Aspinwall ; Alleghany River, Sandy Creek; Alleghany River, Hulton ; Little
Bull (-reek, Tarentum, (A. Koenig coll.) ; Puketta Creek, (A. Koenig coll.) ; McKeen
l 'mi niy, Alleghany River, Larabee ; Warren County, Croijse Run, Garland ; Craw-
ford County, Oil Creek and tributaries, Spartansburg ; small tributary of Conneaut

ORTMANN : THE CKAWFISHES OF THE STATE OF PENNSYLVANIA 391

Creek, Conneautville Station; Erie County, French ('reck. Union City; Sixteen
Mile Creek, Northeast, (Miss G. Kinzer coll.); Walnut < 'reck, Swanville ; Elk ( 'reek.
Girard ; ( 'onneaut and Temple ( "recks, Albion.

Kentucky: Small stream tributary to Rockcastle River, Livingston, Rockcastle

County, (E. B. Williamson coll.).

PREVIOUS RECORDS.

Type locality: Humber River, Toronto, Ontario, Canada, (Girard).

( \\nada : Don River, Toronto, Ontario, (Faxon) ; Weston, Ontario, (Faxon).

New York : Tributary of Racket River, near Tupper's Lake, St. Lawrence
County, (Faxon); Canton, St. Lawrence County, (Faxon); Natural Bridge, Jeffer-
son County, (Faxon); Fulton Lakes, Hamilton and Herkimer Counties, (Faxon);
Petersboro, Madison County, (Faxon); Sodus, Wayne County, (Faxon); Genessee
River, Rochester, Monroe County, (Hagen) ; Forestville, Chautauqua Couuty,
(Faxon).'2"

Pennsylvania : Squaw Run, near Aspinwall, Allegheny County, (Williamson);
tributary of Alleghany River, Port Alleghany, McKean County, (Ortmann).

Ohio: Big Jelloway Creek and tributaries, Knox County, (Osburn and William-
son); Oberlin, Lorain County, (Ortmann).

Illinois: Decatur, Macon County, (Faxon).21

Maryland : Montgomery County, (Faxon).

Virginia : Fredericksburg, Spottsylvania County, (Hagen) ; Wytheville, Wythe
County, (Faxon).

New Localities :

Pennsylvania: Waterford, Erie County, (Dep. Agriculture, Harrisburg).

Maryland: Deer Park, Garrett County, (P. R. Uhler coll., see below under < '.
ctiogenes. Footnote 27 ).'2'2

REMARKS :

In Pennsylvania this variety is well marked, and might safely be regarded as a
species. Its chief characters are found in the shape of the rostrum and the sculp-
ture of the hand. I never was in doubt as to this form, with the exception of
young individuals (less than 30 mm. long), in which the hand does not show its
characteristic features; but in such specimens the shape of the rostrum generally

20 Hagen also gives: "Regis Lake, N. Y.", in the Adirondaoks.

21 This locality sliould be confirmed, see Ortmann, 10(15/;, p. 135.

21 In Rough Ran, West Winlield, Butler County, Pa., I found on June 20, 1904, several young specimens appar-
ently belonging to this variety. I did not take them, expecting to get larger ones, in which hope I was disappointed.
Thus this locality is somewhat doubtful.

392 MEMOIRS OF THE OA.RMEGIE MUSEUM

gives a clue. This, however, is not the case in very young specimens, (less than 20
mm. long), and such I am unable to distinguish from the typical form.

The characters are slightly variable, as has been pointed out above, but this
variety generally is ver}7 uniform in its characters in Pennsylvania. I have not
found any variations worthy of special mention. With reference to the lateral spine
of the carapace, there are specimens which show no trace of it, (young as well as old).
In old specimens this spine is often tubereuliform, and in about half of the number
at hand there is on each side a sharp, but always small lateral spine. In this respect
there is no difference in the specimens of northwestern Pennsylvania from those
found in Allegheny ( 'ounty.

It seems to me that the southern records for this variety (Maryland, Virginia,
and also Kentucky)23, do not refer to exactly the same form which is found
in the north (Canada, New York, northwestern Pennsylvania, northern Ohio).
Hay (1899, p. 966), in the key to the species gives as one of the differential
characters of G. bartoni robustus: ''carapace cylindrical, sides nearly parallel as
far forward as cervical groove, then curving abruptly to the base of rostrum,"
while, under C. bartoni, the carapace is described as " . . . depressed, sides gently
curving toward the front and rear." This cylindrical shape of the carapace is deci-
dedly not present in our northern form ; on the contrary, the depression of the cara-
pace in our robusbus is, if anything, more pronounced than in the typical bartoni;
and our robustus agrees in this respect with < Jirard's type from Canada, preserved in
the Academy of Natural Sciences, Philadelphia, and which has been examined by
the writer.

On the other hand, our specimens from Kentucky seem to approach the form
from Virginia and Maryland. The shape of the carapace is more cylindrical, as
Hay describes it, G:H:B= 1 : 1.05 to 1.2: 1.2 to 1.3. This shows that the width
of the carapace at the branchial and hepatic regions is decidedly less, compared with
the vertical diameter at the gastric region, than in the typical bartoni. There are
other slight differences in the form from Kentucky : ( I ) the rostrum is not quite so
narrow ; (2) the lateral spine of the carapace is absent ; (3) the punctures of the areola
are not so crowded (about three rows), and are similar to those of bartoni; (4) the
impressions of the hand are indistinct ; (5) the double row of tubercles on the inner
margin of the hand is different, the outer row being distinct, but the inner consist-
ing of only a few more or less distinct irregular tubercles. All four specimens from
Kentucky are comparatively small, (the largest is 54 mm. long), and thus the two
last described characters may be due to age, although the specimens differ slightly

"Faxon, 1890; Hay, 1899 ; Williamson, 1905 ; Ortmann, 19056.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 393

from northern individuals of the same size. The other characters incline
toward the typical bartoni, while the shape of the carapace varies to the other
extreme.

A single male of the second form, ahout 60 mm. long, from Deer Park, I larrett
( lounty, Maryland, sent to me for examination by l>r. P. R, Uhlerof Baltimore, was
found under a lot of typical C. bartoni, (supposed to be C. diogenes). This male
agrees fairly well with the specimens from Kentucky. The carapace is rather cylin-
drical ; there are no lateral spines on the carapace ; the punctures of the areola are
like those of C. bartoni; the chelae, which are unequal, and apparently both regen-
erated, have rather distinct impressions on the upper surface, but the inner margin
has only one row of tubercles. The rostrum is of the robustius-type. Thus, of the
characters of robustus, only the shape of rostrum and the impressions of the chelae
were present, all other characters being those of typical bartoni.

Specimens possessing a rather elongated rostrum, but with the other characters
of typical bartoni, 1 have seen associated with individuals of the typical form taken
at Gettysburg, Adams < !o., Pa. (Dep. Agric, Harrisburg) ; but these I have recorded
with typical G. bartoni. (See above, p. 385.)

It is very desirable that the southern form in Maryland, Virginia, Kentucky, and
adjacent localities, should be investigated more closely. The records at hand, and
the few, immature specimens the writer has seen, do not permit a final conclusion
as to whether we have to deal in the south with a form differing from that in the
north, or not. The same reason forbids us to restore our G. robustus to the rank of
a species, which I surely would have done if the Pennsylvanian material alone
were to be considered.

In Pennsylvania G. bartoni robustus is not always associated with C. bartoni. I
found it thus in every case in Allegheny < lounty, in < Irawford < Jounty, and in War-
ren ( lounty. In McKean ( 'ounty I found it associated with C. obscurus in the Alle-
ghany River at Larabee, but the typical G. bartoni was not there, although occurring
not far away in small streams and springs. In Erie County G. bartoni was found
only twice, in Elk Creek and Walnut ('reek, associated with G. bartoni robustus, but
then only a single individual of the former was found in each case. At Albion and
Union ( )ity ( '. bartoni robustus alone was present, and I am sure of it, since 1 hunted
for G. bartoni, but without success. The rich material from N oil beast (forty-four
specimens are now in the Museum, but many more were originally in the lot) did
not contain a single C. bartoni. Thus it is beyond doubt that G. bartoni robustus is
not infrequently found without the typical form, and chiefly so in the most northern
and western sections of the state.

394 MEMOIRS OF THE CARNEGIE MUSEUM

5. Cambarus (Bartonius) carolinus Erichson.
(Plate A, Fig. 4 ; Plate XXXIX, Fig. 3a and 36, and 9 ; Plate XL, Fig. 4).

Aslacus (Cambarus) carolinus Erichson, 1846, p. 96.

Cambarus dubius Faxon, 18846, p. 114 ; Faxon, 1885a, p. 70, PI. 4, f. 3, PI. 8, f. 7 ; Underwood, 1886, p. 368 ; Faxon,

1890, p. C24 ; Hay, 1899, p. 959, 965.
Cambarus carolinus Hay, 19026, p. 38; Ortmann, 1905o, p. 393.
Cambarus (Bartonius) carolinus Ortmann, 19056, p 120, 135.

Bodij robust, smooth, except for short hairs, chiefly on the chela? in freshly
moulted individuals.

Carapace subovate, not depressed, but rather compressed in comparison with the
species described above. G : H : B = 1 : 1.07 to 1.1 : 1.1, that is to say, the vertical
diameter and the transverse diameters of the hepatic and branchial regions are
practically the same, the two transverse diameters being onty slightly greater than
the vertical. The greatest width of the branchial regions is well forward, immedi-
ately behind the cervical groove.

Cervical groove deep, not interrupted on the sides.

Areola distinctly longer than half of tbe anterior section of the carapace
(a:p = 1 : 1.65 to 1.74), very narrow (w:l = 1 : 10 to 15), with only one, or rarely
two, very irregular rows of punctures, which occasionally are almost entirely lacking.

Rostrum (PI. XXXIX, Fig. 3o and 36) slightly curved downward toward the
tip, broad and short, never reaching beyond the distal end of the second joint of
the peduncle of the antennula, but generally only to the middle of it, being some-
times even shorter than that. Upper surface slightly concave, with elevated mai--
gins. Margins straight, sub-parallel, or slightly converging toward the tip, sud-
denly contracted into a broad, short, triangular acumen. Basal angles of acumen
rather sharp, but without any trace of marginal spines. These angles are empha-
sized by the sudden disappearance of the slight swelling of the lateral margins,
which are not at all swollen on the acumen. Acumen pointed, but point short.
Postorbital ridges .short, almost parallel, ending bluntly anteriorly.

Surface of carapace punctate, granulated only on the hepatic region, and some-
times with a few indistinct granules immediately behind the cervical groove on the
branchial region. No lateral spine. External orbital angle rarety or not at all
marked, generally formed by a rounded or slightly angular, insignificant projection,
but never with a spine. Branchiostegal spine formed by a small, often indistinct,
tubercle.

Abdomen always distinctly shorter than the carapace, narrower than the latter
in the male, almost as wide as the carapace in the female. Anterior section of telson

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 395

on the posterior lateral corners with one or two spines. Posterior section semi-ellip-
tical, distinctly wider than long, about as long as the anterior section.

Epistoma with posterior part comparatively long, hardly one and one-half times
as broad as long ; transverse groove indistinct and close to posterior margin; an-
terior median depression also indistinct, and thus the posterior part of the epistoma
appears rather flat and smooth. Anterior section constricted at base, semi-circular
or semi-elliptical or subquadrate, rarely with a median anterior point, often more or
less truncate anteriorly, or even slightly emarginate. Transverse diameter never
greater than longitudinal, as great as the latter or shorter.

Antennnla with a small tubercle on the lower margin of the basal joint.

Antennal peduncle without distinct spines or tubercles on the two proximal,
joints.

Antenna! scale short and small, slightly longer than the rostrum, and not reach-
ing beyond the distal end of the fourth joint of the antennal peduncle. Spine of
outer margin strong. Laminar part only slightly wider than the spine.

Flagellvm short, not reaching beyond the third abdominal segment in the male,
and not beyond the second segment in the female, but often hardly longer than the
carapace.

First pereiopods (PI. XL, Fig. 4) short, not very robust, not undergoing much
change with age, and not differing much according to sex. Hand ovate, broad and
depressed. Surface punctate. Inner margin of palm convex, with a single row of
distinct tubei-cles, larger proximally; occasionally there is a second, incomplete row
inside of and parallel to this. Outer margin smooth and rounded proximally for
a short distance, but soon becoming angular and forming a distinct ridge along the
edge in the region of the base of the immovable finger. This ridge has a regular
row of deep punctures, giving the distinct appearance of serrations along the outer
margin of the hand. Fingers as long as, or slightly longer than the palm, slightly
gaping at the base in both young and old individuals, straight. Outer margin of
movable finger with punctures, but very rarely with tubercles. Cutting edges with
a few irregular tubercles in the proximal half. Upper surface of each finger with
a low rib, bordered by rows of punctures.

Carpopodite slightly longer than wide, about as long as the palm, with a deep
longitudinal sulcus above. There is always a strong, more or less pointed, spine on
the inner margin, directed obliquely forward, and a strong, often spiniform, conical
tubercle in the middle of the anterior margin of the under side, (rarely obsolete).
The tubercle at the lower articulation with the hand is low and indistinct. Rarely
there are accessory tubercles, which, however, are never spiniform. The most fre-

396 MEMOIRS OF THE CARNEGIE MUSEUM

quent are a proximal tubercle on the inner margin, and one between the two larger
spines first mentioned, and another just behind the base of the large spine of the
inner margin.

Meropodite smooth, with 1-3 indistinct tubercles near the distal end of the upper
margin, often entirely obsolete. Lower side with two rows of spiniform tubercles.
The outer row consists of 2-6 (if only 2, they are followed by a few undulations
produced by punctures); the inner consists of 6-11 tubercles. Outer articular tuber-
cle with carpopodite without spine. All spines of the first pereiopods are indistinct
in very young specimens.

Ischiopodite of third pereiopod hooked in the male. Hook of the male of the
first form strong, subcorneal.

Coxopodite of fourth pereiopod in the male with a prominent, blunt, and slightly
compressed subcorneal tubercle.

First pleopods of the male (Plate XXXIX, Fig. 9) similar to those of C. barton i.

Annulus venlralis of the female likewise of the type of C. bartoni, but less trans-
verse, and the posterior margin more swollen and elevated, while the anterior is
hardly elevated at all, but depressed.

Size. — The largest male of the first form at hand is from Dunbar, Fayette
( !ounty, and measures 67 mm. in length. The largest female is from Ohiopyle,
Fayette County, and measures 80 mm. in length.

Color (Plate A, Fig. 4). — Whole body rather uniformly orange-chrome (Ridgway,
1886, VII, 13) to chinese-orange (VII, 15), very brilliant in fresh shells, shading to
orange-rufous (VII, 12) and cream-color (VI, 20) on the sides. Color most intense
on anterior part of carapace and on the claws. Abdomen orange-chrome, shading to
salmon-color (VII, 17) or ferrugineous (IV, 10), or ochraceous-buff (V, 13). Chelae
varying from orange-chrome to saturu-red (VII, 16). Legs chinese-orange to salmon-
color and ferrugineous. Often a brown or blackish coat of mud covers a great part
of the body, obscuring the colors. The color of young specimens is semitransparent,
with more or less red prevailing, but the rostrum and chelse are always distinctly
red. Color of eggs salmon (VII, 17) to salmon-buff (IV, 19).

There are one hundred and thirty-eight specimens. in the Carnegie Museum ;
ninety-three from Pennsylvania, thirty-six from West Virginia, and nine from
Maryland.

DISTRIBUTION.

LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.

Pennsylvania: Westmoreland County, Jones Mills ; Fayette Count//, Dunbar; In-
dian Creek ; Rainier Park, Ohiopyle ; Somerset County, Windber ; Listie ; Rock-
wood ; Myersdale.

i

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 397

Maryland : Garrett Count y, Selbysport.

West Virginia : Preston Comity, Reedsville ; Tucker County, Parsons ; Mineral
County, Sch ell.

PREVIOUS RECORDS.

Type locality: Greenville, Greenville County, S. C., (Erichson and Faxon,
1885<(, pp. 9, 56).

Tennessee: Cumberland Gap, < llaiborne County, (Faxon).

Virginia: Pennington Gap, Lee County, (Faxon).

West Virginia: Southwestern West Virginia, (Hay); Terra Alta (Cranberry
Summit), Preston County, (Faxon).

Indian Territory: Cherokee Nation, (Faxon).-'

New Locality : 2' Blowing Rock, Watauga < iounty, North ( 'arolina, (J. P.
Moore coll., Acad. Nat. Sei. Philadelphia).26

REMARKS :

This species, the Red Crawfish of the mountain regions, occupies, next to C.
propinquus, the smallest area in this state. Its morphological characters are very
uniform, not only in this state, but also in the neighboring parts of West Virginia
and Maryland. The chief variations, as already mentioned in the description, are
found in the armature of the chelipeds, but they keep within comparatively narrow
limits. The rostrum, which is quite variable in C. bartoni, is here very constant;
only the degree of convergence of the lateral margins and the length of the acumen
varying to a certain extent, (See PI. NX XIX, fig. 3a and ?>!>.) Of course, we must
disregard monstrosities, under which head 1 place two cases, (from Dunbar and
Parsons), where the rostrum has almost no acumen at all, being rounded oil' apically.

The armature of the meropodite and carpopodite varies as described above; the
carpopodite in particular showing a various number of tubercles, though they never
become spiniform as in C. monongalensis. The lower outer margin of the merop-
odite has always more than one spine, when spines at all are visible, which is gen-
erally not the case in very young individuals.

"This locality needs conlirniation, sec Ortnianii, 1905b, p. 135.

"This species is abundant in certain parts of Somerset and Fayette Counties, Pa., and well known to the natives-
I have seen chimneys in abundance at Confluence, Somerset County, which undoubtedly belong to this species, but was
not able to secure specimens, my time being limited. I have seen a specimen at Flanigan, 4 miles north of Suinerlield,
Fayette Co., and one from Humbertston, Fayette Co. (O. T. Cruikshank coll. ). Reports received from farmers, always
emphasizing the red color and borrowing habits of the crawfish in question, are the following : Great Meadow Run,
west of Obiopyle, Fayette County ; Millrun, Fayette County ; Ursina, Somerset County ; Salisbury, Somerset County.

26 These specimens (collected June, 1H93, and, according to Professor Moore's recollection, dug out of holes) agree
in all essential points with our material, only the rostrum is nairovver.

398 MEMOIRS OF THE CARNEGIE MUSEUM

I think I have observed that in specimens from the south (Maryland and West
Virginia) there is a more pronounced tendency to develop a second row of tuber-
cles at the inner margin of the hand. Such specimens, with a few additional tuber-
cles, occur also in Pennsylvania, but in West Virginia they are more frequent, and
the additional row becomes more distinct and more regular. There is moreover a
specimen from Parsons, W. Va., in which traces of a third distal row are visible.
On the other hand specimens with one row only are also found in West Virginia.

No interesting freaks or monstrosities have been observed.

6. Cambaeus (Bartonius) monongalensis Ortmann.
(Plate B, Fig. 4. Plate XXXIX, Fig 4a, Ab and 10. Plate XL, Fig. 5.)

Cambarvs dubious Williamson, 1901, p. 11, (non dubius Faxon).

Cavibatw monongalensis Ortmann, 1905n, p. 395.

Cambarvs (Bartonius) monongahmis Ortmann, 19056, p. 120.

This species being closely allied to C. carolinus, the description will be given in
terms of comparison with the latter.

General shape of body, carapace, cervical groove, and areola identical with that in
C. carolinus. G:H:B=\: 0.9 to 1.1 : 1.1 to 1 .3.

Rostrum (PI. XXXIX, Figs. 4a and 4b) markedly different from that of C. caro-
linus. It is as long as that of the latter species, or slightly shorter in the average,
never reaching beyond the middle of the second joint of the antennula, and is uni-
formly narrower. The upper surface is concave. The margins are less sharply
elevated, the elevation decreasing gradually to the apex. Margins distinctly con-
verging, and contracted to form the short, triangular acumen, but the contraction
is not so sudden as in G. carolinus, so that the angles at the base of the acumen are
not so sharp, but rounded. Acumen with short point. Postorbital ridges short
and rather indistinct, distinctly divergent posteriorly.

Sculpture of carapace and other details as in C. carolinus. The abdomen and telson
are also identical, but the lateral corners of the anterior section of the telson" have
only one spine.

Epistoma similar to that of G carolinus, but the truncated (subquadrate) shape
prevails in the anterior section, which has often a small median anterior point.

Antenna and antennula similar to those of C. cam/inns, but antennal scale- shorter,
not much longer than the rostrum, and reaching only to the middle of the fourth
joint of the antennal peduncle.

First pereiopods (PI. XL, Fig. 5) in general shape similar to those of C. carolinus,
but hand not quite so broad, and there are important differences in the armature.
The inner margin of the hand invariably has only one, but a distinct, row of tuber-

OKTMANN : THE CRAWFISHES OF THE STATE OF PEFNSYLVANIA 399

cles. The outer margin is rounded, with no indication of angulation, and entirely
lacks the serrations of 0. enrol in us, the punctures producing the latter forming in
this species no regular row.

The carpopodite is much more spinous. The large spine of the inner margin is
well developed. The spine in the middle of the anterior margin of the under side
is tuberculiform, and the tubercle at the lower articulation with the hand is insig-
nificant. But there always are additional distinct spines, which are well developed,
although smaller than the large spine of the inner margin. A spine on the proxi-
mal end of the inner margin is always present, and also a spine between the large
spine of the inner margin and the anterior tubercle of the lower side. (If the latter
spine is missing, the claw has been regenerated.) Often there are other spines.
The proximal spine of the inner margin may be double, and there may be one or
several spines or tubercles near the base of the large spine of the inner margin,
anterior or posterior to it.

Meropodite with the distal tubercles of the upper margin very indistinct, gener-
ally missing. The outer lower margin is formed by a smooth keel, which has in
most cases only one small tubercle near the distal end, which may be obsolete. In
rare instances there are two tubercles. If there are more, the claw has been regen-
erated. The inner lower margin has a row of 6-12 spiniform tubercles; if less,
the claw has been regenerated.

All the other organs are similar to those of C. carolinus, more particularly the
first pleopods of the male (PI. XXXIX, fig. 10) and the annulus of the female.

Size. — The largest male of the first form, from Edgewood Park, Allegheny
County, is 68 mm. long; the largest females (same locality and Monaca, Beaver
County), are 76 mm. long.

Color (PL B, Fig. 4). — In specimens with fresh shells the middle of the cara-
pace and abdomen is of a beautiful marine-blue (Ridgway, 1886, IX, 2); the hepatic
and branchial regions are cyanine-blue (IX, 3) and china-blue (IX, 13), shading
toward the lower margin to pale-blue (IX, 16). The marine-blue of the abdomen is
restricted to the anterior parts of the segments ; the posterior parts and the epimera
are china-blue shading to pale-blue. Margins of rostrum maroon-purple (VIII, 9).
The hand is cyanine-blue above, shading toward the lower side to cobaWblue (IX, 12),
azure-blue (IX, 15) and pale-blue. At the base of the dactylopodite there is a good
deal of royal-purple (VIII, 7). The dactylopodite is cyanine-blue, the outer margin
violet (VIII, 10). The finger-tips are orangwermilion (VI I, 12), shading proximally
to salmon-color (VII, 17) and whitish. The carpododite of the chelipeds is marim-
blue, shading to cyanine-blue and French-blue (IX, 6). Tubercles and spines of

400 MEMOIRS OF THE CARNEGIE MUSEUM

hand and carpopodite pale vinaceous-pink (lighter than IV, 21). Basal joints of
antennae cyanine-blue ; flagellum annulated dahliaspurple (VIII, 2) and very pale
Mac (lighter than VIII, 19). Pereiopods pale-blue, upper edges shading to china-blue,
near the bases white. Lower side of body whitish, on the abdomen suffused with
china-blue and pale-blue.

These brilliant colors fade more or less in old specimens, but the marine-blue
always remains the ground-color. A brownish or blackish deposit of mud often
obscures the colors, and such specimens often appear blackish-blue all over.

Young individuals are semitransparent, with a distinct pale-blue hue on the
abdomen and chela?, and heliotrope-purple (VIII, 18) on the carapace.

Distinct color varieties are rare, and the few observed will be mentioned below.

The color of the newly laid eggs is hazel (IV, 12) or ochraceous (V, 7) ; later on
they are vinaceous-buff' (V , 15) on one side, pinkish-white on the other.

The Carnegie Museum possesses two hundred and seventy-nine specimens of this
species, two hundred and thirty-six of which are from Pennsylvania, and forty-three
from West Virginia.

DISTRIBUTION.

LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.

Pennsylvania : Westmoreland County, Hill, (opposite Leechburg) ; Braeburn ;
Dundale ; Jeanette ; Allegheny County, south of Logan's Ferry, (C. V. Hartman
coll.); Hulton, (R. Dornberger coll.); Wilkinsburg, (D. A. Atkinson coll.); Edge-
wood Park, (type locality) ; Nine-Mile Run, Pittsburgh ; Fern-Hollow, Pittsburgh ;
Schenley Park, Pittsburgh ; Squaw-Run, near Aspinwall, (D. A. Atkinson) ; North
Versailles Township, (opposite Stewart) ; Jack's Run and Long Run, South Ver-
sailles Township ; Boston, (D. A. Atkinson) ; Thompson's Run, Kennywood, (F. E.
Kelley coll.); Carnegie, (D. A. Atkinson); Moon Township, (A. T. Shafer coll.) ;
Lashell's Hollow, Moon Township : Beaver County, Doctor Heights, Monaca ; Wash-
ington County, Francis Mine near Burgettstown ; Taylorstown ; Monongahela City ;
West Brownsville ; Fayette County, Smithfield ; Cheat Haven ; Greene County, Deer
Lick.

West Virginia : Hancock County, Congo ; Hollidays ( 'ove ; Brooke County, Col-
liers ; Ohio County, Elm Grove; Marshall County, Cameron; Nuss; Monongalia

County, Morgantown.

PREVIOUS RECORDS.

The only previous records are those given by Williamson (as C. dubius),

Schenley Park and Fern Hollow, Pittsburgh, and Moon Township, Allegheny

County.

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 401

ADDITIONAL LOCALITY.
The writer has seen this species at Sandy ('reek, Allegheny County, Pa.

REMARKS.

The Blue Crawfish is rather abundant in the region indicated by the above
localities, and apparently represents a geographical differentiation of 0. carol in us,
the Red ( Jrawfish. The points in which it differs from the latter species, aside from
color, are slight, but are constant according to my observations.

The rostrum is markedly different from that of C. carolinus, although slightly
variable. It always is smaller than that of the latter species and narrower, with less
distinctly marked lateral angles. The lateral margins are swollen, but the swelling
is less marked, and does not suddenly decrease at the lateral angles. The degree of
convergence of the margins is variable, but generally much more pronounced than
in the Red < irawfish ; cases like that figured in Plate XXXIX, Fig. 4b, are rather
rare, in fact, this case forms the extreme in this direction.

In the shape of the hand, the rounded outer margin and the absence of serra-
tions on it are other striking characters of this species ; and the carpopodite is dis-
tinguished by the larger number of spiniform tubercles, as described above. The
outer lower margin of the meropodite generally has a blunt and smooth keel, with
only one small tubercle near the distal end (which may lie absent). Among my
material I had only ten specimens which revealed an exception, where two such
tubercles were present, and in only two of them were these tubercles present on both
chelipeds. In the others they occurred only on one side. There are instances in
which a large number are found, but always in claws which have been regenerated.

Aside from the slight variations indicated above only a few exceptional cases
have been encountered in which marked deviation from the above description
of the prevalent colors occurs. The blue ground-color is always present on the
anterior part of the carapace and the cliche. In a rare variation, which has
been observed about half a dozen times, the ground-color of the posterior part of the
carapace and parts of the abdomen are more or less purplish [av/ricv^Orpurple, VIII,
3). Furthermore a single adult female was found at Monaca in which all red tints
were absent; the blue of the body was very clear, the margins of the rostrum were
blue like the carapace, the finger tips were whitish, all spines and tubercles were
pure white, and the antennal flagellum was a pure blue. Specimens in which the
margins of the rostrum have the same shade of blue as the carapace are not
infrequent.

402 MEMOIRS OF THE CARNEGIE MUSEUM

Other freaks are occasionally found, as for instance a specimen without an acumen
on the rostrum, the latter heing evenly rounded off; a specimen with the immov-
able finger of the left hand with a double tip, the outer one the larger, and a spec-
imen with the movable fingers of both hands only half as long as usual, thick, short,
and conical. Such cases apparently are due to some previous injury, and suggest
nothing of special interest.

All the variations mentioned above are rare and are not restricted to certain
parts of the range of the species, so that we do not distinguish any regional varieties.

7. Cambarus (Bartonius) diogenes Girard.
(Plate A, Pig. 3 ; Plate XXXIX, Fig. 11 ; Plate XL, Figs. 6 and 7.)

Cambarus diogenes Girard, 1852, p. 88 ; Faxon, 1884*, p. 144 ; Abliott, 1884, p. 1157; Faxon, 1885n, p. 71 ; Faxon,

1885ft, p. 359 ; Faxon, 1885c, p. 140; Underwood, 1886, p. 368 ; Faxon, 1890, p. 624 ; Ortmann, 1891, p. 12;

Hay, 1896, p. 489, Fig. 7 ; Faxon, 1898, p. 650 ; Oshnrn & Williamson, 1898, p. 21 ; Williamson, 1899, p. 20, 48 :

Hay, 1899, p. 959, 961 ; Harris, 1900, p. 267 ; Williamson, 1901, p. 11 ; Hay, 1902<i, p. 235 ; Ortmann, 1905a, p. 398.

Cambarus ob<sus Hagen, 1870, p. 8], pi. 1, f. 39-42, pi. 3, f. 163, pi. 9 ; Smith, 1874, p. 639 ; Forbes, 1876, p. 5, 19 ;

Bnndy, 1877, p. 171 ; Bnndy, 1882, p. 183; Bundy, 1883, p. 403.
Cambarus diogenes ludoiicianus Faxon, 18846, p. 144 ; Hay, 1899, p. 959, 962.
Cambarus dubius Osburn and Williamson, 1898, p. 21, (non dubius Faxon).
Cambarus [Bartonius) diogenes Ortmann, 19056, p. 120, 135.

Body robust, smooth, except for short hairs, chiefiy on the chelae in fresh shells;
the hairs also to a certain degree persist upon the hand and fingers in older indi-
viduals.

Carapace subovate, not depressed, but rather compressed. G : H : B = 1 :0.88 to
1.0:1.06 to 1.2; that is to say, the transverse diameter of the carapace is very
slightly greater than the vertical, at the hepatic region sometimes even less. Great-
est width of branchial regions well forward, not far from the cervical groove.

Cervical groove deep, not interrupted on the sides.

Areola distinctly longer than half of the anterior section of the carapace (a :p =
1 :0.61 to 0.75), very narrow, and generally obliterated in the middle; that is to
say, the two lines bordering the branchial regions are in contact in the middle of
the carapace. In rare instances a small space is left between them, upon which
there is no room for punctures.

Rostrum more or less lanceolate, rather narrow, but not very long, reaching
hardly beyond the distal end of the second joint of the peduncle of the antennula,
being often shorter. Upper surface slightly concave, with elevated margins. Mar-
gins not much swollen, the swelling gradually disappearing toward the tip, con-
verging, straight, or slightly convex, contracted to form a short triangular acumen.
Basal angles of acumen indistinct, rounded, without any trace of marginal spines.

ORTMANN : THE CRAWFISHES OF THE STATE OK PENNSYLVANIA 403

Point of acumen short. Postorbital ridges short, terminated bluntly in front,
slightly divergent posteriorly, ending in a low, indistinct swelling.

Surface of carapace punctate, slightly granulate on the hepatic region, and with
a few granules on the branchial region, immediately behind the cervical groove.
No lateral spine. External orbital angle present, distinct, angular or rounded, but
without tubercle or spine. Branchiostegal spine formed by a small, often indistinct,
tubercle.

Abdomen about as long as the carapace, or very slightly shorter or longer, nar-
rower than the carapace in the male, markedly wider and about as wide as the
carapace in the female. Anterior section of tehon with 1-3 (generally 2) spines on
the posterior lateral corners. Posterior section semi-elliptical, slightly wider than
long, and about as long as the anterior section.

Epistoma similar to that of C. carolinus and monongalensis, comparatively long
and narrow, rather flat, and with the anterior section semi-circular, semi-elliptical,
or truncate and subquadrate, with or without median point, and about as long as
wide.

Antennula with a small tubercle on the lower margin of the basal joint.

Antennal peduncle without spines or tubercles on the proximal joints.

Antenna! scale small and short, slightly longer than the rostrum, and reaching to
the base of the fifth joint of the peduncle of the antenna. Spine of outer margin
strong ; laminar portion not much wider than the spine, its inner margin parallel to
the outer margin of the spine for a considerable distance.

Flagellum short, often only as long as the carapace or even shorter, never reach-
ing beyond the second abdominal segment.

First pereiopods (PI. XL, Fig. 6) stout and very robust in old individuals; not
much different in the male and female, except for their very large size in old males.
Hand ovate, broad, depressed. Surface punctate. Inner margin of palm convex,
with two irregular rows of tubercles, and a few scattered tubercles on the upper
surface near the marginal rows. Outer margin smooth, rounded proximally,
slightly angular distally. Fingers at least one and one-half times as long as palm
(the latter measured from articular tubercle with carpopodite to articular tubercle
with dactylopodite), gaping at the base, straight both in young and old specimens.
Cutting edges with a number of strong but irregular tubercles ; one tubercle at
about the middle of the edge of each finger is generally the largest. Outer margin
of movable finger with more or less distinct tubercles at the proximal end. Upper
surface of each finger with a low longitudinal rib, bordered by rows of punctures.

Carpopodite about as long as wide, shorter than palm, with a deep longitudinal

404 MEMOIRS OF THE CARNEGIE MUSEUM

sulcus above, and a few more or less distinct tubercles between sulcus and inner mar-
gin. A strong pointed spine in the middle of the inner margin, straight, and di-
rected obliquely forward. A tubercle or spine on anterior margin of lower side, and
a low tubercle at articulation with hand. A few additional tubercles may be pres-
ent on the inner margin and the lower side, but they are very rarely spiniform.

M< ropodite smooth, with 1-3 indistinct tubercles near the distal end of the upper
margin. Lower side with an outer row of 1-4, and an inner row of 7-1 1 spiniform
tubercles.

Ischiopodite of third pereiopods hooked in the male ; hooks in the male of the
first form strong, subcorneal.

Coxopodite of fourth pereiopods with a strong, slightly compressed tubercle.

First pleopods of male (PL XXXIX, Fig. 11) similar to those of C. bartoni, the
tip of the inner part, however, tapering gradually to the point.

Awnulus ventralis of female similar to that of C. carolinus.

Size. — The largest specimens at hand from the eastern part of the state are a
male (first form) and a female from Ridley Park, both 83 mm. long. From the
western part of the state 1 have a male of the first form from Nine-Mile Run, Pitts-
burgh, which measures 92 mm. in length, and a male of the second form from Mill-
vale, Allegheny County, which is 93 mm. long. The largest female is from Nine-
Mile Run, and measures 97 mm. in length.

In the west this species attains a much larger size. The maximum length has
been recorded by Hagen, 4.5 in. = 115 mm. However, the Carnegie Museum pos-
sesses a male of the first form from Bluffton, Wells County, Indiana (collected by E.
B. Williamson), which is now (in alcohol) 122 mm. long, but measured 124 mm.
when alive.

Color (PL A, Fig. 3). — The color is rather variable within certain limits, but the
ground-color is similar to that normally seen in crawfishes, brownish or greenish.

< Iround-color on carapace and abdomen from olive-green (Ridgway, 1886, X, 18)
to raw-umber (III, 14), mummy-brown (III, 10) and ferrugineous (IV, 10), shading on
the sides through drab (III, 18) or russet (III, 16) to fawn-color (III, 22) and whitish.
Margins of rostrum rufous (IV, 7) or ferrugineous (IV, 10). The hand is tawny-olive
(III, 17) to burnt sienna (IV, 6) and rufom, shading to olive-yellow (VI, 16) toward
the outside. At the bases of the fingers there is often a distinct shade of olive-green
(X, 21). The finger tips are rufous, the tubercles of the hand cream-buff (V, 11) or
whitish. The legs are ockraceous-buff (V, 10) with olive-buff (V, 11), or russet (III, 16)
with olive-green (X, 18) at the joints. Lower side of body pale rufous or pale
orange-buff (VI, 22), or whitish. The antennal flagellum is annulated dark olive-

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 405

green and whitish. The relative amount of green and brown changes very consider-
ably with different specimens, but in general we may say that green prevails in the
young, and brown in older specimens.

Color of newly laid eggs buff (V, 13); when more advanced prune-purple
(VIII, 1), or indian-pv/rple (VIII, 6) on one side, cream-color (V. 20), or dirty-white on
the other.

In the Carnegie Museum there are altogether thi'ee hundred and twenty-three
specimens of this species, of which twenty are from, eastern and two hundred and
sixty-six from western Pennsylvania. Two specimens are from Maryland, five
from West Virginia, twenty-four from Indiana and three from Iowa ; while one speci-
men is from each of the states of Delaware, Ohio, and Kansas.

DISTRIBUTION.

LOCALITIES REPRESENTED IN THE CARNEGIE MUSEUM.

Pennsylvania: Bucks County, Penns Manor; Philadelphia County, Kssington ;
Delaware County, Ridley Park; Marcus Hook; Greene County, Waynesburg; Rice's
Landing; Fayette County, Smithfield ; Dunbar; Pennsville ; Washington ('mud!/.
Francis Mine near Burgettstown ; Beaver County, Baden ; Racoon Township ; Alle-
gheny County, Troup's Retreat, Moon Township, (E. B. Williamson coll.); Stowe and
Neville Townships, (Atkinson, Graf, and Williamson coll.); Edgeworth, (Mrs. E.
Courtney coll.); Westview, (D. A. Atkinson' coll.) ; Millvale ; Fern Hollow and Nine-
Mile Run, Pittsburgh ; Schenley Farm, Pittsburgh, (F. E. Kelly coll.); Silver Lake,
Pittsburgh, (E. B. Williamson coll.) ; Bruce's Ice Pond, Pittsburgh, (Atkinson and
Williamson coll.); Carnegie, (D. A. Atkinson coll.); Jack's Run, North Versailles
Township; Rankin, (O. T. ( Vuikshank coll.); Squaw Run, near Aspinwall ; Mon-
trose ; Harmarville ; Russelton ; between CJibsonia and Bakerstown Station ;
Thornhill ; Westmoreland County, Donohoe ; New Alexandria ; Dundale ; Reser-
voir of McGee Run, Derry ; Blairsville Intersection ; Livermore ; Kiskiminetas
Junction; Indiana County, Homer; Creekside ; Jefferson County, Punxsutawney ;
Armstrong County, Long Run, Avonmore Station ; Kittanning ; Butter County, Ren-
frew; Branchton ; Lawrence County, Wampum; Mercer County, Mercer.

Delaware: Sussex County, Seaford, (S. N. Rhoads coll., exch. Ac. Nat. Sci..
Philadelphia).

Maryland: Kent County, Chestertown, (E. G. Vanatta coll., exch. Ac. Nat.
Sci., Philadelphia).

West Virginia: Hancock County, Congo; Brooke County, Colliers; Wetzel
County, New Martinsville.

406 MEMOIRS OF THE CARNEGIE MUSEUM

Ohio: Franklin County, Columbus, (E. B. Williamson coll.).

Indiana : Wells Count;/, Bluffton ; Twin Lakes ; Craigville ; Liberty Center ;
Uniondale; Blackford County, Hartford City; Dc Kalb County, NewVille (all coll.
by E. B. Williamson).

Iowa: Greene County, Cooper, (J. B. Hatcber coll.); Lee Count;/, Denmark (R.
L. Baird coll.).

Kansas : Douglas County, Lawrence, (S. Prentice coll.).

PREVIOUS RECORDS.

Type Locality : Washington, D. C, (Girard).

New Jersey : Trenton, Mercer County, (Abbott).

Maryland : Worcester County, (Faxon) ; Dorcbester County, (Faxon) ; Caro-
line County, (Faxon); Baltimore County, (Faxon); St. Mary County, (Faxon).27

Virginia : Accomac County, (Faxon) ; Northampton County, (Faxon) ; Alex-
andria County, (Faxon); Prince William County, (Faxon); Fredericksburg,
Spottsylvania County, (Faxon) ; Petersburg, Dinwiddie County, (Hagen).

North Carolina: Kingston, Lenoir County, (Faxon); Wilmington, New Han-
over County, (Faxon).

Pennsylvania : Derry, Westmoreland County, (Faxon) ; Pittsburg, Allegheny
County, (Williamson).

Ohio : Oberlin, Lorain County, (Osburn and Williamson) ; Knox County,
(Williamson); Columbus and Lockbourne, Franklin County, (Faxon) ; Montgomery
County, (Williamson).

Michigan : Detroit, Wayne County, (Faxon).

Kentucky : Louisville, Jefferson County, (Faxon) ; Bee Spring, Edmonson
County, (Faxon); Mammoth Cave, Edmonson County, (Hay).28

Indiana: Kendallville, Noble County, (Bundy) ; Riverside, Laporte County,
(Faxon) ; Kokomo, Howard ( Jounty, (Faxon) ; Mechanicsburg, Henry County,
(Bundy) ; Irvington, Marion County, (Hay) ; North Salem, Hendricks County,
(Hay) ; Greencastle, Putnam County, (Hay) ; Bloomington, Monroe County, (Hay);
Knox County, (Faxon).

27 Faxon also gives Deer Park, Garrett County, (Coll. P. R. Uhler). I hare doubted (19056, p. 136) the correct-
ness of this locality. In order to be sure I have asked Dr. P. R. Uhler of Baltimore to send to me the specimens upon
which this record was founded, and Dr. Uhler very kindly complied with my request. There were two males of the
second form and four females in the lot, all belonging to C. bartoni, and representing the typical form, with the excep-
tion of one male, which I have identified with the southern form of 0. bartoni robustus (see above, pp. 391 and 393).
Thus this record (Deer Park) for C. diogenes is to be dropped.

28 The localities in Jefferson and Edmonson Counties are given by Faxon with a ? ; but Hay's record of this species
from Mammoth Cave tends to confirm their correctness.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 4(17

Illinois: Chicago, Cook County, (Hagen) ; Evanston, Cook County, (Hagen) ;
Lawn Ridge, Marshall County, (Hagen); Abingdon, Knox County, (Faxon);
Decatur, Macon County, (Faxon); Belleville, St. ('lair County, (Hagen).

Wisconsin: "Abundant in Wisconsin," (Bundy) ; Racine, Racine County,
(Faxon); Green County, (Faxon); Appleton, Outagamie County, (Faxon).

Minnesota: Fort Snelling, Hennepin County, (Faxon).

Iowa: Davenport, Scott County, (Faxon); Delhi, Delaware County, (Faxon) ;
Belmond, Wright County, (Faxon).

Missouri: St. Louis, St. Louis County, (Faxon); Carroll County, (Faxon).

Kansas: Leavenworth, Leavenworth Count)', (Faxon); Lawrence, Douglas
County, (Harris).

Arkansas: Paragould, Greene County, (Faxon); Fayetteville, Washington
County, (Faxon).

Mississippi: Monticello, Lawrence County, (Hagen).

Louisiana : New Orleans, (Hagen).

Wyoming: Cheyenne, Laramie County, (Faxon).

Colorado: Clear Lake, (Faxon)2"; Boulder, Boulder County, (Harris)30.

REMARKS.

Cambarus diogenes occupies two areas, in the United States, which, according to
our present knowledge, are separated from each other ; a western and an eastern.
Both areas enter Pennsylvania, the one extending over a large portion of the south-
western territory of this state, the other being much smaller and restricted to the
southeastern extremity.

I have closely studied the material at hand, and have found certain differences
between eastern and western specimens, which however are very slight, and not
always constant; yet a tendency to a morphological separation between the eastern
and western forms seems to be indicated. Faxon (1885a, p. 72) has already called
attention to some of these differences.

The description given above refers chiefly to the eastern form of this species.
Specimens from western Pennsylvania show the following differences :

1. Areola in most cases not entirely obliterated. There are, indeed, cases in the
eastern form where the two lines bordering the branchial regions are not in contact,
but they are rare. In western Pennsylvania the latter condition is rather the rule,
although specimens in which both lines unite, forming in the middle only one line,

89 Location unknown to the writer.

30 Professor T. D. A. Cockerell has sent to me for examination a young male collected October 7, 1905, in a small
stream near Boulder. Although very small it clearly belongs to this species.

408 MEMOIRS OF THE CARNEGIE MUSEUM

are by no means absent. There is some variability in this character. Generally a
very narrow space is left between the two lines, which does not leave room for any
punctures. But it is a curious fact that the lareola is widest in specimens from
Fayette and eastern Greene Counties. In these localities specimens with an entirely
obliterated areola are exceedingly rare, and specimens with the areola so wide that
there is room for one irregular row of punctures are rather frequent, (Pennsville,
Dunbar, Smithfield, Fayette County; Rice's Landing, Greene County). Such
specimens with punctures on the areola, which is accordingly wide, are scarcely
found anywhere else. I possess only one from Pittsburgh.

2. Rostrum, in all eastern specimens at hand, with a more or less distinct acu-
men. In the western form there is a distinct tendency to render the acumen ob-
scure. Indeed there are many specimens which have the acumen exactly as in the
eastern form, but there are as many -where it is not marked, the margins converging
evenly to the tip. In such specimens the rostrum assumes a rather regular lanceo-
late shape, and appears somewhat more elongate and narrow. However, it is actu-
ally not longer than in the eastern form.

3. The swelling at the posterior ends of the postorbital ridges is sometimes more
distinct in western specimens.

4. In our western form the external orbital angle is rarely angular, but generally
blunt or rounded. In some cases it is very slightly developed.

5. There are specimens in western Pennsylvania where the posterior section of
the telson is longer and more tapering. In extreme cases this is rather striking, the
posterior section being distinctly longer than the anterior, and longer than wide.
This latter condition is never found in eastern specimens. On the other hand there
are many western specimens which do not differ in this respect from the eastern,
and many transitional conditions have been observed.

6. Antennal scale in the western form often slightly wider than in the eastern,
and with a stronger spine. This difference, however, is very slight.

7. Chil.r in the western form (PI. XL, Fig. 7) of slightly different shape, but this
difference always holds good, provided the chela has not been regenerated, and is
otherwise normally developed. The inner margin of the palm in the eastern form
has always two rows of tubercles, while on the upper surface there are at the best
only a few minute, scattered tubercles, chiefly near the base of the dactylopodite.
In the western form the two rows of tubercles are also generally distinguishable,
but often the inner row is irregular, and merges into the scattered tubercles present
upon the inner half of the upper surface. These latter tubercles are invariably
present, and are much more numerous than in the eastern form. The dactylopodite

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 409

of the western form is generally shorter. While in the eastern form it is at least one
and one-half times as long as the palm, in the western form this relation is the
maximum, and is due to a proportionally heavier and stronger development of the
palm as compared with the fingers. (This does not hold good for regenerated
chelae, in which the dactylopodite always is considerably longer in proportion to
the palm.)

The tubercles of the cutting edge of the dactylopodite are slightly different in
both forms. In the eastern the third or fourth is generally much larger, and, just
before it, there is a gap, as if one tubercle were suppressed. This produces a distinct
excision at the base of the finger, which is always more or less marked, even in
rather young individuals, so that the fingers always appear gaping at the base (PL
XL, Fig. 6). In the western form the fourth or fifth tubercle is larger than the
rest, but there is no distinct gap anterior to it, the three or four proximal tubercles
being rather equidistant. They decrease slightly in size from the first to the fourth,
so that a slight emargination is indicated. But this emargination and the large
tubercle following it are well distinguishable only in larger individuals ; in younger
.specimens they are rather indistinct, or entirely wanting, so that the basal gap is
absent, and the fingers are in contact all along their edges.

The differences in the relation of dactylopodite to palm, and in the tubercles of
the dactylopodite, give to the whole chela of the western form a different aspect,
the hand appearing rather more massive with shorter fingers. (See PI. XL, Figs.
6 and 7.)

8. The colors in western specimens are more vivid, and with more contrast.
Eastern specimens are more uniformly russet or olive-green, with no oil-green at the
base of the fingers, while the latter tint is very characteristic of the western form, at
least in specimens of a certain size. Old western specimens, when fresh, are rather
brilliantly colored.

I think I am able to recognize and to distinguish eastern and western speci-
mens, chiefly by the help of the characters of the chelae, if the latter are normally
developed ; but I do not know whether it would be advisable to distinguish both
forms by varietal names. The latter may be necessary in future, when the form- of
C. diogenes from the regions west of Pennsylvania have been more closely studied.
I think I am able to see certain differences in the western specimens in our collec-
tions, but the material is too poor to be sure of it.

No remarkable freaks or malformations have been observed in this species, with
the exception of a few color variations. They are the following:

1. In Fern Hollow, Pittsburgh, a specimen was found in a stagnant pool, 6 to

410 MEMOIRS OF THE CARNEGIE MUSEUM

10 inches deep, in yellowish brown mud. Its color was entirely yellowish brown,
mottled lighter and darker, and no trace of olive-green was present. This appa-
rently was a stray specimen.

2. A large male of the first form was found at Dunbar, Fayette County, the
ground-color of- which was salmon-color (VII, 17), the abdomen buff' (V, 13), whitish
on the sides. The red was brightest on the chelae, with traces of green between the
tubercles of the hand, and the lower side of the chelre and body were dirty brown-
ish yellow. This is) apparently a case of albinism.

IV. ECOLOGY AND GEOGRAPHICAL DISTRIBUTION.
A. Ecology.31

Satisfactory conclusions as to the relation of geographical distribution to the
physical conditions of the country can only be expected, if we know all about the
ecological laws governing the different species. With reference to the seven species
of crawfishes present in Pennsylvania, we shall see that the ecological conditions
are quite varied, and the single species behave very differently. Thus it is necessary
to discuss these facts first, before we attempt to study the distribution.

Three main types of ecological conditions may be distinguished among our
crawfishes. We possess species which generally live in the larger rivers ; other
species which favor the opposite extreme, preferring the groundwater, where it is
not far from the surface, and appears in the shape of springs and swamps ; and
intermediate between these two conditions is a species which selects the smaller
streams for its home. We may conveniently call these " the river species," " the
mountain-stream species," and "the burrowing species."

1-. The River Species.
Cambarus limosus, (Jambarus projnnquus, Ckimbams obscurus.
Although principally living in the larger rivers of the state, these species are
not entirely restricted to them, being able to live in any larger body of water, run-
ning or stagnant, providing it is permanent. Thus these forms, in some cases, go

31 "Ecology," the science of the "relation of organisms to external conditions,'' is the oldest terra, created by
Ha:ckel (" Oecologie," in " Generelle Morphologic dtr Organixme," 1866). The term " Rionomics," which is often used
in its place, was first introduced by E. Ray Lankester (in the article: Zoology, in Encyclopaedia Britannica, 9th ed.,
1888, p. 803), and subsequently, but independently, (as " Bionomie") by J. Walther (Einleitung in die Geologie ah
Mstorische Wissensrhafi" 1. Bionomie des Meeres, 1893, p. XX). The term "Oecologie" was revived chiefly by E.
Warming, ( Plantesamfund. Grundtraek af den fekologiske Plantegeografi, 1C95). The term " Ethologie " introduced
by F. Dahl (Verh. Ges. Naturf. & ;Erzte, Bremen, LXI1I, 2. 1891. p. 123) has a wider sense, including also what
we here call " life-history. "

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 411

far up stream into the smaller tributaries of our rivers, sometimes almost to their
sources. The only condition which stops them in an up-stream migration seems to
be the character of the watercourse, which must not he too rapid and rough. This
is a very important fact, and largely explains the absence of these species in the
mountainous regions of the state. On the other hand, these species are by no means
averse to quiet bodies of water, such as ponds and lakes, and although the parts of
Pennsylvania where these species are found are singularly free from lakes, craw-
fishes are almost regularly found in them, and even in artificial ponds, reservoirs,
etc. It has been observed that in ponds and lakes these species seem to thrive
exceptionally well.

Among the three species belonging to this class, there are certain differences.
Cambarus limosus of the eastern part of Pennsylvania has its main abode in the
quiet streams of the lowlands. It goes up stream for a certain distance, but rarely,
and only under exceptional conditions, (see below), is it found in streams which are
somewhat rough. Its center, at least for this state, is in the lower part of the Dela-
ware River, where it is under the influence of the tides. Here it prefers the muddy
banket, living among the water weeds, and congregates often in large numbers at the
mouths of small streams tributary to the river. In fact the latter places are the
most favored, since this species loves to hide under stones, and it is chiefly at the
mouth of streams that stones are found in this part of the Delaware. Further up
stream, beyond the reach of the tide in the Delaware, and in its tributaries (Nesh-
aminy Creek, Schuylkill River, Brandy wine Creek) and in the Susquehanna and
Potomac drainages, this species is generally found hiding under stones, as was
first reported by Abbott (1873, p. 80) with reference to the Delaware River at Tren-
ton, N. J. But such is not the exclusive habitat of C. limosus. It is very often
found in quiet ponds, in ditches or canals, where there are no stones to afford con-
cealment. In these places it frequents patches of weeds ( Vailisneria, etc.), often in
considerable numbers. From such places (ditches of the Delaware meadows at
Trenton) it was reported by Faxon (1885a, p. 88). C. limosus is generally found in
very shallow water, but sometimes at a considerable depth. I captured a few speci-
mens in a quiet cove of the Delaware River at Penns Manor, Bucks County, at the
woodwork of a pier, at a depth of from six to eight feet (Sept. 15, 1905), and fre-
quently got numbers of it in creeks and canals (Schuylkill ('anal, Manayunk ;
Delaware and Raritan Canal, Princeton), in the water weeds, at a depth of from two
to four feet. Although this species loves to hide under stones, and although it
scoops out the dirt under stones, it is by no means a burrowing species. The hol-
lows made under stones are very insignificant, and I have never observed that it

412 MEMOIRS OF THE CARNEGIE MUSEUM

makes holes in the banks of streams. Faxon (1885a, p. 89) reports that Mr. Uhler
found this species near Cumberland, Md., in " holes in the bottom and sides of a
canal," but whether these were made by the crawfish, or were cracks and joints
between stones, remains doubtful. In the lowlands in Maryland this species,
according to the same authority, is found under stones in rivers and creeks. We
may say of C. limosus that it is of all the species of this state the one which most
decidedly prefers the quieter water of large rivers, canals, and ponds ; that it
likes to hide under stones, but is not at all averse to muddy bottoms and masses of
vegetation.

The latter trait distinguishes it from the species of the western rivers, Cambarus
obscurus, which dislikes muddy bottoms (without stones) and vegetation. In fact,
this is so general a rule, that it is vain to look fur ('. obscurus in any part of a river
which has no stones. Only in rare and exceptional cases have I found it not hiding
under stones, apparently being forced to do so by necessity. I observed this in two
cases : in the Allegheny River at Larabee, McKean County, and the Shenango
River at Linesville, Crawford County. In both cases the river runs through peaty
soil (through the Pymatuning Swamp at Linesville), and it was only after a long
search that the species was discovered, when I struck places where stones were lying
in the water. But it was interesting to note that at both places the supply of stones
was apparently not sufficient to accommodate all the specimens, and so a number
of them had to be content with a shelter afforded by the peaty banks, where they
had built short, horizontal holes, not more than a few inches long, close to the edge
of the water. These holes are apparently only temporary, and are often abandoned
and changed, since a number of them were seen on the banks above the present
water level, which were built and inhabited at a previous higher stage of the river.
C. obscums does not love vegetation. The patches of Dianihera mucricana, so fre-
quent in our rivers, do not harbour many crawfishes, although they are not entirely
absent from them ; but they do not hide under these plants and their roots, but
under stones. They always scoop out a hollow under the stone selected, and bring
out the mud and gravel, throwing up a small rubbish pile in front of the hole,
which, however, is soon obliterated by the current, The crawfishes are easily found
by noting these rubbish piles. They rarely go into deep water (possibly only in
winter), but always are close to the banks in shallow water ; but on the other hand
they never go out of water. In the mountains of western Pennsylvania this species
is generally absent, and it is apparently the roughness of the streams which causes
its disappearance. The stones, under which it hides, must be rather permanent in
their position, and must rest upon a bed of mud, sand, or gravel, to afford congenial

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 413

conditions. In the mountain streams the rocks are rolled over very frequently, Dot
only at high stages of water, but also under ordinary conditions, and this apparently
does not suit the tastes of this species, and may be even directly dangerous. Since
the general direction of the migration of this species in western Pennsylvania was
and is upstream, it is evident that falls and rapids in the mountain regions present
effective barriers, In the ponds and lakes of western Pennsylvania, connected with
the Ohio drainage, this species is very abundant, but here also it always selects
stones under which to hide.

Cambarvs propinquus is restricted to Lake Erie and the lake drainage. It lives
in the tributaries of the lake, exactly under the same conditions as C. obscurus. In
the lake itself it has been found on two occasions. Dr. D. A. Atkinson collected a
number at Presque Isle in the bay, but particulars as to their habitat were not
recorded. The only other specimen from the lake was collected by myself on the
sandy and gravelly beach near Miles Grove, thrown out by the surf, but alive. It
does not seem to be very abundant in the lake, or at any rate seems to favor
only certain places, and we may presume that places with stones and rocks on
the beach and not too much exposed to the surf are the proper localities in which
to look for it.

Cambarus 2)ro]>inqm(s sanborni in Ohio and West Virginia is found under exactly
the same ecological conditions as its representative forms in Pennsylvania.

Nothing was known hitherto as to the ecological habits of G obscurus and pro-
pinquus, except the short notice of Hay (1896, p. 498), that in Indiana G. propin-
quus lives "in the smaller streams hiding under stones, concealed in short burrows
along the banks, or resting quietly on the hottom."

2. Tin' Mountain Stream Species: Catnbarus bartani.
Conspicuously differing from "the River Species," Cambarus bartoni favors the

rough streams of the mountains, hills, and the uplands generally, and is absent from
the large rivers. The size of the stream may vary. In fact it goes up to the very
sources and is found in the springs. The amount of water may lie very small. In-
deed it is frequently found in streams which dry out superficially during the hot
season. But in such cases water is always present at a certain depth. The varying
and often scanty supply of water forces this species to accommodate itself to these
conditions, and thus it has become to a certain extent a burrowing species. It
always selects stones to hide under, and in larger streams with a permanent supply
of water is satisfied to scoop out a hollow under the stone after the fashion of C.
obscurus. But very often the burrows are more complex, consisting of a hole going

414 MEMOIRS OF THE CARNEGIE MUSEUM

down to a depth of a foot or even more. These burrows are found along the banks
of the streams, and the opening is often not in the water, but away from it, but
rarely more than a few feet. The deepest burrows are found in late summer and
fall, when the small streams are almost or entirely dry. Then necessity compels
the crawfish to dig deep to reach the underground water. I have observed burrows
eighteen inches deep in a vertical direction (see Plate XLI, Fig. 1). Under these
conditions a considerable amount of dirt (mud, sand, gravel), is removed from the
hole, and this is piled up in more or less regular mounds at the entrance of the hole,
often assuming the shape of " chimneys," which may be fully equal in size to those
of the typical chimney-builders. Here we see the origin of this habit. G. bartoni
is not an habitual chimney-builder, but is content to hide under stones and to scoop
out shallow holes when the stream has plenty of water. But when the supply of
water becomes scant it has to dig down to reach it, and the burrows and mud-piles
are the natural consequences of the attempt of the crawfish to accommodate itself to
these peculiar conditions.

The manner in which the burrows are constructed, and the "purpose" of the
chimneys will be discussed below when we come to consider the true burrowing
species, and it may be remarked here, that everything said with reference to the
latter holds good also for < '. bartoni.

The roughness of a stream presents no obstacle to the presence of C. bartoni.
Indeed, it prefers small streams which descend in cascades and fall from the hillsides,
provided the rocks lying in them are stationary enough. It goes to the very upper-
most springs and is frequently found there associated with G. monongalensis or G.
carolinus, and also may be found near C. diogenes. I have observed cases where C.
bartoni occupied holes, which were apparently built by specimens of these other
species, and am able to give the following instances. Digging for C. monongalensis
at West Brownsville, Washington County, I found in a large and wide hole a female
C. bartoni (with eggs). The individual was much too small for this hole. To all
appearances an old (abandoned?) hole of G. monongalensis was here occupied by
C. bartoni, Similar observations were made at Avonmore Station, Armstrong
( 'ounty, where in the swampy ground of the valley of Long Run a colony of C. diog-
enes was found, and several C. diogenes were taken. In two holes, however, a half-
grown specimen of C. bartoni was found, and again these holes were much too large
for them. This place was about fifteen yards distant from the stream in which G.
bartoni was abundant. Another similar case was observed at Oreekside, Indiana
County.

Going down stream C. bartoni remains abundant, as long as the character of the

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 415

stream is maintained; after that it becomes scarce, but it does not disappear entirely,

and in western Pennsylvania is frequently found associated with ('. obscurus, in
eastern Pennsylvania witli C. limosus. In the large rivers it is generally entirely
absent, and, if found, it is at the mouth of small streams, or at places where there are
springs on the banks. Here it becomes evident that temperature plays an impor-
tant part. The mountain streams, which are the favorite haunts of ('. bartoni, are
characterized all the year round by a rather uniform, hut comparatively low tem-
perature. In winter the temperature of the water goes down to just above the
freezing point, hut generally remains slightly higher (in running water about 35°-
40° F.); in summer the maximum of these streams rarely goes above 60°, and does
so only temporarily for a few days, while in the larger streams it remains for weeks
above 70°, and may go up to 80° or even more. (The temperature of the Ohio
River at Baden, Beaver County, on August 26, 1905, was 78° F. on a compara-
tively cool day.) That it is temperature which affects distribution is evident in
summer, when in western Pennsylvania in the warm water of the rivers C. obscurus
is found, but where there are springs on the banks discharging perceptibly cooler
water into the river, G bartoni suddenly appears.

It may be mentioned that I once found this species under very peculiar condi-
tions. At New Hagerstown, Carroll County, Ohio, I discovered numerous burrows
in the black muck of a swampy meadow at the bottom of a small valley, which I
took first for burrows of G diogenes. But I was unable to get any diogenes, every
hole investigated being occupied by G. bartoni (about half a dozen were taken).
This part of the meadow was close to a hillside, at the foot of which were numerous
springs with a few stones, also sheltering specimens of C. bartoni. On the other
side of the valley, which was about 100 to 200 feet wide, was a small stream with
sandy and gravelly bottom, and a few stones, where also a few G. bartoni were
present. The largest number of specimens was present in the swampy meadow,
which is rather exceptional, but finds its explanation in the scarcity of stones in
this locality.

The variety C. bartoni robustus in general agrees with the typical form as con-
cerns ecological conditions, especially in that it prefers rough, rocky streams. How-
ever, it was found preferably in streams of a larger size, avoiding tin' smaller head-
waters. As Williamson (1901, p. 11) puts it : "at the headwaters" (Squaw Run,
Allegheny County, is taken as an instance) "bartoni is found ; following down the
stream robustus is noticed ; then an occasional obscurus; till finally bartoni becomes
rare and disappears; then robustus disappears; and further down C. obscurus is the
only species."

416 MEMOIRS OF THE CARNEGIE MUSEUM

Since those parts of the stream which are inhabited by G. bartoni robustus are
always well supplied with water, this form does not need to make extensive har-
rows, and I have never observed regular chimneys.32

The ecological conditions under which C. bartoni occurs were to some extent
previously known. Goodman (1833 (1842), p. 293) gives a good account of them33
as observed in small streams near Philadelphia. According to Abbott (1873) it is
found near Trenton, N. J., burrowing in the muddy banks of ditches and small
streams, rarely of the river (Delaware). This, however, is not the usual condition,
as we have seen above. Faxon (1885a, p. 63) says that it prefers cooler waters of
mountain regions or uplands, living under the stones in clear streams and in
springs, which is the usual condition under which it is found in this state. How-
ever, that there are variations in its habitat, occasioned by exceptional conditions,
is seen from the case mentioned above from the state of Ohio, from Abbott's account,
and from the observations of Dr. J. Sloan as reported by Faxon (/. a), according to
which, in southern Indiana, it is found in ponds and still water, not in running
streams. This is, however, not always the case in Indiana, since, according to Hay
(1896, p. 489) it is found in "springs and streams of clear running water, where it
hides under stones or digs short burrows into the banks."

3. The Burrowing Species.
Cambarus carolinus, Gambarus monongalensis, Cambarus diogenes.

a. General habitat.
The burrowing species are always found at a certain distance from open water,
although often in close proximity to streams, ditches, or ponds ; but never, under
normal conditions, in them. Exceptions are very rare, and only accidental, and
found chiefly in the case of young individuals which have not settled down perma-
nently, or of individuals which have been disturbed.34 These species, however,
always depend on the presence of water, but it is the groundwater which is inhabited

32 Shufeldt ( 1896, p. 27) figures a chimney of C. bartoni robustus from near Washington, D. C. Why Shufeldt attrib-
utes this chimney to this form, I do not understand. He says that he studied burrows near Washington, "many of
these were of C. diogenes, others were of C. bartoni robustus, which I found abundant in Montgomery County, Md."
The specimen which built the chimney figured was not taken by Shufeldt, and he says that in the vicinity another bur-
row was opened which contained a C. diogenes. According to the description of the hole belonging to the chimney, it is
too deep and complex to belong to C. bartoni, and I do not see any reason for not regarding it as belonging to C. diogenes.

33 The crawfish hole, eight to ten inches deep, with a wider chamber at the end, under stones in a small stream,
with the opening in the water, undoubtedly belongs to this species.

"A case where young specimens of C. diogenes were found in numbers in open water by Dr. D. A. Atkinson will be
discussed below. (See V. ) This case is also to be regarded as exceptional.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 417

by them, and thus they are found at places where the groundwater is near the sur-
face, in springs and swamps. In order to reach the water these species have to dig
a hole in the ground, which often goes down through a considerable amount of dry
soil, but it is always filled with water at the bottom.

The three species belonging to this class differ slightly with regard to the selec-
tion of their localities. C. carolinus chooses the mountains, and is found in springy
places on the highest parts of the Alleghany plateau. The most favored localities
are high valleys with a " hard pan," that is to say a layer of stiff clay below, which
serves to keep the groundwater within a few feet of the surface. In such places the
surface is often apparently dry, but upon digging down fresh and clear spring-water
is found at a depth of one to three feet, and the holes of this species go down to the
" hard pan " in order to reach the water. C. monongalensis favors similar conditions,
yet it does not live in the mountains, but on the Toothills west of the Chestnut
Ridge. (The physiographic classification of these features will be discussed below).
In this region extensive valleys with clay bottoms are rarely found, and thus
C. monongalensis is content with the more restricted deposits of clay found on the
hillsides. Such localities, however, are very abundant in this region, and wherever
there is a spring and a certain amount of clay this species occurs. It prefers the
cool spring-water, and if the springs collect to form a small swamp, this species is
found on its upper margin, not in the swamp.

G. diogenes does not haunt springs to the same extent as C. carolinus and C.
monongalensis. It is sometimes found under similar conditions as the other two
species, but generally at places where a spring or small stream spreads out to form
a swamp. It is also abundant in swampy ground along the borders of ditches and
streams, and in swamps formed in depressions of the valleys of the large rivers
(abandoned ox-bows). In the formerly glaciated area of the state it prefers kettle-
holes. Like C. monongalensis it rarely occurs in the soft mud of swamps, but
generally along their borders, where the firmer ground affords a better chance
to dig more permanent holes. In consequence of the habit of preferring swamps to
springs, C. diogenes is generally found at a lower elevation than C. carolinus
and monongalensis, where it comes into contact with them. The two last named
forms occupy the region of the clear and cool spring-water, while C. diogenes appears
a little further down stream, where the water is not so clear, and in summer
not so cool. In and near swamps the water in the holes of C. diogenes is often
stagnant and muddy (even sewage is not much objected to by this species), while
in the holes of the other two species there is always fresh and clear spring-water
bubbling up.

418 MEMOIKS OF THE CARNEGIE MUSEUM

b. Shape of the burrows. (See PI. XL, Figs. 8 and 9, PI. XLI.)

The burrows of these species (and also of C. bartoni) are to a certain degree alike,
although they are very variable in depth and shape, so that there is very little uni-
formity. Only a few features are common to them. From the more or less dry
surface they go down to the groundwater, where there is generally a kind of a pocket
or widened chamber (PI. XLI, Figs. 5 and 6). The width of the hole corresponds
to the size of its inhabitants. It seems that one and the same individual perma-
nently uses the same hole, although one and the same hole ma}' be occupied by dif-
ferent individuals in succession, for an old abandoned hole may be occupied by a
young specimen. This happens chiefly in localities where the holes are much
crowded. There are places where the ground is fairly honeycombed with them,35 and
under such conditions a new hole may interfere with an old one. when a young
specimen after attaining the proper size begins to build its own burrow, as it invari-
ably does. If the old hole is abandoned the young specimen may take advantage
of it, while in the alternative case, a fight ensues which ends in the expulsion or de-
struction of the weaker.

Each hole is always occupied by owe individual only, with two exceptions. The
first is the case of mating, couples, when one adult male of the first form and one
female are found in one and the same hole. The second is when the young of a
certain size are associated with the mother in the hole of the latter.

The holes have all manner of shapes (see PI. XL, and PI. XLI). They may con-
sist of a single shaft only, or may be more complex, branching off in various direc-
tions, and may have more than one opening at the surface. The chamber may be
well marked or indistinct, and there may be several chambers. The chambers may
be simple widenings of the hole, or may form side pockets. The direction of the
descending shaft is rarely more or less vertical, and if vertical in the upper part, it
generally soon assumes a slanting direction, and sometimes it is irregularly spiral.
Shorter or longer branches may go off at the sides, and these may end blindly or
may ascend to the surface. At the bottom side-branches may be absent, or may be
developed to a considerable extent, running either horizontally or vertically. On
steep hillsides, or along the banks of ditches, the general direction of the burrows is
very often horizontal, the outward opening being lateral (PI. XLI, Figs. 2, 3, 7).

The depth of the holes depends on the distance of the level of the groundwater

35 1 have seen this on the largest scale in the case of C. earolinus in the valley of Upper Decker's Creek near Reedsville,
Preston County, W. Va. The whole valley, for two or three miles, offers favorable conditions for this species, and
thousands of chimneys may be seen everywhere, coming up even between the railroad ties of the Morgautown & King-
wood Railroad.

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 419

from the surface. The holes are driven down by the crawfish to such a depth that
a good supply of water is at the bottom at all seasons. Where the water appears at
the surface, or is very near to it, the holes are sometimes hardly a foot deep. Gen-
erally they are considerably deeper, as much as two and three feet. They certainly
at times go even deeper than this, but I never tried to dig at places where such con-
ditions prevailed, that is to say, where from all appearances the level of the ground-
water was more than three feet from the surface. Such conditions were not infre-
quently met with in the case of C. carolinus.

In a general way we may say that the holes of G. barUmi are very simple
(PL XL, Fig. 8; PL XLI, Fig. 1). Among the true chimney-builders the holes
of C. diogenes are also rather simple, consisting often of a single shaft with a pocket
at the bottom (PL XLI, Figs. 5 and 6). In C. monongalensis they are decidedly
more complex (PL XLI, Fig. 2), and the highest degree of complexity is reached
in C. carolinus (PL XL, Fig. 9).

The shape of the burrows of C. diogenes was first described by Girard (1852, p.
89), who called attention to their variable character. Tarr (1884, p. 127) has given
sketches of burrows of this species, and also observed their variability. Of C. caro-
linus, only the fact that it is a chimney-builder was known (Faxon, 1885a, ]>. 71).
The burrows of C. monongalensis (as dubius) were described by Williamson (1901, p.
12), and he emphasizes their complexity as compared with those of C. diogenes.

c. Cons/ruction of the burrows and of the chimneys.
Although the "chimneys" or mud-piles at the mouths of the burrows have
often been described and their purpose discussed, (Girard, 1852; Tarr, 1854; Shu-
feldt, 1896; Harris, 1903), the manner in which the crawfish excavates the burrow
and piles up the mud in front of it had never been correctly observed. Abbott
(1885) describes how Mr. J. DeB. Abbott saw the crawfish (C. diogenes) engaged in
building its chimney, and states that it comes out of its hole " bearing on the back
of its right claw a ball of clay mud, which by a dexterous tilt of the claw was
placed on the rim of the chimney." This description, as we shall presently see, is
apparently founded upon correct observation, but the observer witnessed only the
final act, and drew from it a wrong inference. The old observation of Goodman
(1833, (1842), p. 293), that C. bartoni brings out of its hole an "armful of rubbish
and throws it over the side of his cell, and down the stream," should be quoted,
since, although referring to another species, it is pertinent and applies well to the
regular chimney-builders as regards the mode of carrying the mud.86

36 The way of carrying the mud out of the holes seems to be identical in all burrowing species. It has beeu observed
in a similar form by Mr. W. S. Sutton in C. pilasus Hay, as described by Harris (1900, p. 272). That the crawfish uses

420 MEMOIRS OF THE CARNEGIE MUSEUM

I have repeatedly observed the digging and the removal of the dirt out of the
burrow. Of course it is impossible to see the digging going on in the field inside of
the burrow, and consequently this was observed in the laboratory in the case of speci-
mens of C. diogenes and monougalensis kept in captivity in large glass jars partly filled
with clay and water. It is not difficult at all to see them at work, and after they have
been brought into the laboratory the specimens begin to work within a short time,
digging out the mud, carrying it upward and plastering it all over the walls of the
jar. After some time (days or weeks), their activity lessens, and not much digging
is done, producing the impression that they have become discouraged in the effort
to construct something similar to the burrows in the field.

In digging the chelse of the first pereiopods are used. The fingers are slightly
spread out, so that they are about parallel, thus acting as forks for digging. They
are pushed vertically down into the mud on both sides at the same time, and a
lump of mud is thus loosened and lifted upward toward the ventral face of the
body. In lifting the chelipeds are bent toward the body (the region of the mouth),
and finally the ball of mud is appressed to the anterior part of the body and held
in position by the chelse. Very likety also the third maxillipeds take hold of it,
but it was impossible to ascertain this. In this position, as Goodman expresses it,
carrying an "armful " of dirt (or rather twoarmfuls), the crawfish walks slowly and
deliberately to the mouth of the hole. I have repeatedly observed it coming out in
nature.37 It advances to the top of the chimney and deposits the mud pellet upon
the rim, finally pushing it into the proper position with the upper (outer) surface
of the claws. This latter act apparently was seen by Mr. J. DeB. Abbott ; but
according to my experience the mud is not brought up upon the back of the claw,
but held, as described above, between the folded claws and the anterior part of
the body.

After having been disturbed in the field, the crawfishes often begin to work again
within a short time, and it is chiefly on such occasions that I have seem them at
work, with the exception of one case, when I saw a large male of G. diogenes at work
on the evening of April 30, 1905, (Sunday), in Nine-Mile Run, Pittsburgh. The

the " lateral tail-fins and telson " in any way, as suspected by Shufeldt (1896), in the sealing up of the orifice of the
burrow, is hardly possible.

37 The following particular instances maybe mentioned : C. bartoni in a spring near Burgettstown, Washington
County, Aug. 4, 1904 ; C. cnrolinun at Indian Creek, Fayette County, July 11, 1904 ; at Ohiopyle, Fayette County,
July 12, 1904 ; at Myersdale, Somerset County, August 11, 1904 ; C. inonmigalensis in Fern Hollow, Pittsburgh, May 6,
1901 ; at Edgewood Park, Allegheny County, May 9 and 21, 1904 ; April 21, 1905 ; Monaca, Beaver County, June 30,
1904; at Cheat Haven, Fayette County, September 6, 1904 ; at Cameron, Marshall County, W. Va., May 1, 1905 ; at
Morgantown, W. V., May 16, 1905 ; C. diogenes at Dunbar, Fayette County, September 7, 1904 ; in Nine-Mile Run,
Pittsburgh, April 30, 1905.

OKTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 421

usual time for working Beema to be at night, and I have repeatedly observed that
holes and chimneys disturbed or destroyed on one day exhibited signs of recent
action on the following day. The crawfishes also seem to work occasionally on
rainy or cloudy days; at all events, on such days it is easier to induce them to come
to the mouth of the hole.

As to the purpose of the "chimneys," different opinions have been expressed.
Abbott (1884) believes that the chimneys are designed, and that the crawfish intends
to build just such structures, that is to say, rather regular subcorneal mud towers.
He maintains this against Tarr's view (1884, p. 127), that the chimneys are not a
necessary part of the burrows, and that they simply are the result of the digging.
Subsequent writers have rather inclined towards Tarr's idea, for instance Shufeldt
(1896, p. 89), who says that it is easier for the crawfish to build a chimney than to
carry the mud away from the hole, and that "it is the most convenient and safest
way to get rid of the pellets, besides being the least troublesome, and the method by
which they are the least likely to roll back into the burrow." Harris (1903, p. 605)
thinks that the chimneys very likely are only "the result of the easiest method of
disposing of the material removed in excavating the burrow."

I must indorse the latter opinion, and for the following reasons. Regular chim-
neys, although claimed by Harris (I. c.) to be "usually" present and well built,
are by no means so frequent as believed by most authors. Of course they are
abundant in each colony of chimney-builders, and attract the attention of the
observer. However, according to my experience well built chimneys are rather
scarce compared with the total number of holes existing in a particular locality. In
the majority of cases only more or less irregular and shapeless mud-piles are found
at the mouths of the holes, and it is only under certain favorable conditions that they
assume the shape of " chimneys." These conditions occur when the upper part of
the hole is more or less vertical (see PI. XL, Fig. 9 at J ; PI. XLI, Fig. 2 at C; Figs.
7 and 8), so that there is opportunity for the crawfish, in bringing up the pellets of
mud, to deposit them rather uniformly all around the rim of the chimney. Sup-
posing that it is always the lowest part of the rim at which additional pellets are
deposited, which is altogether a very likely supposition, if the crawfish wants to get
rid of the pellets as quickly as possible, the chimney must grow regularly. If. how-
ever, the mouth of the burrow opens in a slanting direction or horizontally, more or
less one-sided mud-piles will be the result. (See PI. XLI, Fig. 2 at A ; Figs. 3 and
4). Further, much depends on the character of the material brought up. If it is
fresh clay (not disturbed before), as will generally be the case when the crawfish is
digging out a new hole, the pellets will be firmer, stick better to the rim of the chim-

422 MEMOIRS OF THE CARNEGIE MUSEUM

ney, and will remain in position, thus favoring the construction of a "well-built"
chimney. On the other hand, when the mud is very soft, chiefly so when the craw-
fish is not digging new holes, but only cleaning out the old ones, the pellets are not
firm, and the more liquid mud will flow down the outside of the chimney and ren-
der it lower and broader and, consequently, less " well-built." This latter fact also
explains why young specimens often construct the neatest and most elaborate chim-
neys (Abbott). Young specimens, when they begin to work, bring out undisturbed,
firm, and sticky clay, and the pellets are more likely to remain where they are
placed on the rim of the chimney, which thus becomes very regular. Old speci-
mens, on the contrary, live in holes which are practically finished, and when they
work it is rather a process of " housecleaning " than of " housebuilding." The mud
removed is more liquid and less stick}', and thus the chimneys are shapeless and
irregular.

Very often the opening of the chimney is found closed. Abbott believes that
the closing is merely the result of the accidental falling in of the rim. This may
indeed happen, but in other cases it is plain that the crawfish closes the aperture
intentionally, and Girard (1852) regarded this as the completion of the work of
chimney-building. Shufeldt and Harris likewise believe that the crawfish itself
seals up the burrow. This is my own opinion, and with Girard I think that the
sealing up is the final act characterizing the completion of the burrow. Sealed
burrows are very often found (see PI. XL, Fig. 9; PI. XLI, Figs. 2, 3, 4), chiefly in
summer and fall, and it is in many cases evident, by the material used (see PI.
XLI, Fig. 6) that the shutting up was done by the crawfish by depositing pellets
in the orifice. Often the "stopper" is not at the orifice itself, but a certain distance
(5 to 6 inches) below. In fall the stopper is made rather substantially and fills the
upper end of the hole for a distance of 6 to 10 inches (see PI. XL, Fig. 9b at /)), and
such a filling cannot be accounted for by accidental falling in.

In my opinion the construction of the hole is the chief aim for which the craw-
fish works. For the removal of the clay and dirt an opening on the surface is
needed ; but when the burrow is completed this opening is shut up again, and the
crawfish is content to remain inside, possibly for weeks or even months. This
affords protection for the crawfish and its young from enemies (snakes).. Females
with eggs or young are almost always found in closed holes. It also affords neces-
sary quiet and seclusion during the moulting process (soft shells are generally found
in closed holes). It furthermore protects the hole from the disturbing influences
of rain and frost. Of course it would not be advantageous to have the hole perma-
nently sealed, since the crawfish wants to get out now and then (for mating, for

OKTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 423

food), but this is necessary only at long intervals (even for food it is not absolutely
imperative to go out frequently, see infra), and the stopper is easily removed.
During winter a more effective stopper is provided by the crawfish, and it remains
for three or four months shut up in its hole.

The chief activity in chimney-building is in spring.3" During winter frost de-
stroys or damages the upper parts of the hole, and the rebuilding necessitates a good
deal of work, and large mud-piles are accumulated in consequence (4 to 12 inches
high, 12 to 18 inches in diameter). But after the hole has been restored to a satis-
factory condition work ceases, and in summer not much fresh mud is brought out.
Occasionally new chimneys are seen in summer, and the activity may be resumed
at any time if necessary. Besides young specimens remain active all through the
summer. In G. diogenes, as we shall see, it is chiefly in midsummer that the
young begin to build their own holes. In the other species this may take place at
any time from spring to fall, and thus the new and often very regular chimneys of
small specimens may be seen at any time during the warm season.

General activity again begins with older specimens late in the fall, and this lias
a very interesting cause, and my attention was called to it by Mr. F. E. Kelly of
Pittsburgh, but I have confirmed it by subsequent observations of my own. It is
evident that the deepest parts of the holes are occupied and used by the crawfishes
only in winter ; these parts go down to about three feet, and thus are entirely out
of reach of the frost. In summer these parts are abandoned and the crawfish in-
habits only the upper parts of its burrow. In digging for crawfish in summer I often
followed the main hole to a considerable depth, finally discovering that this hole
was filled with soft ooze and mud, and that no crawfish was in this part ; further
careful investigation generally revealed a side branch at a higher level, which was
clear of mud, and here the crawfish was captured. In the fall the deeper, aban-
doned part of the hole (see PI. XLI, Fig. 7 at c), which fills up during the summer
with dirt, forming at the bottom of the hole a soft, pulpy mass, is reclaimed by the
crawfish in order to go deeper down out of reach of frost ; the mud is consequently
removed, and the necessity of cleaning out these deeper parts of the hole is the
cause of the renewed activity in the autumn (PL XLI, Fig. 7). Before Mr. Kelly
communicated to me Ins discovery of this fact I had not paid attention to it, but
was able to verify it in the summer and fall of 1905. The fall activity takes place

38 Young specimens begin first, as soon as the frost is out of the ground. New chimneys of C. diogenes were seen
on March 23, 1905, in Nine-Mile Run, and the activity wasgeneial on April G, 1905, ( Renfrew). The first signs of new
chimneys of C. monongalenm* were seen at Edgewood Park on March 18, 1905 (frost only partly out of the ground ) ; the
activity was general ou March 31, (Colliers, W. Va.), and April 4 (Edgwood Park).

424 MEMOIRS OF THE CARNEGIE MUSEUM

in the month of November (in the neighborhood of Pittsburgh), after the first kill-
ing frosts.39

It is evident from the foregoing observations and considerations that the " chim-
neys" are not necessary parts of the burrows. They are simply the result of the
work of the crawfish, and only represent the material removed from the holes,
which must be carried somewhere, and is most conveniently disposed of right at
and around the mouth of the hole. The regular shape of the chimneys is simply
due to the way the crawfish has to work under certain conditions, and to the phys-
ical properties of the clay.

Nevertheless there are certain advantages connected with the shape of the chim-
neys, which, however, are by no means always present, and, in my opinion, are not
originally intended. The uppermost part of the hole generally has the tendency
to be more or less vertical; an addition of 4 to 12 inches adds so much to the
length of the vertical canal, and the crawfish, when sitting at or near the mouth, is
able to suddenly drop down to escape enemies, which, as I have repeatedly
observed, it actually does. Thus a considerable length of the vertical part is
decidedly advantageous, giving the crawfish a chance to get more suddenly and
effectually out of reach of danger.

Another effect of the mud-pile is noticed when the hole opens horizontally on
sloping ground (banks of ditches). Here the mud-pile generally is semicircular,
convex toward the ditch, concave toward the mouth of the hole, and thus serves to
keep the water at a uniform level in the hole (PI. XLI, Figs. 2 and 3), for generally
in such cases the hole has spring-water flowing out of it. This may be advantage-
ous under certain conditions, since I have often found that by removing a pile of
mud of this character I was able to drain off the water from a considerable part of
the hole, thus making it distinctly uncomfortable for the crawfish, as is evidenced
by the fact that it often came out of the hole, as if to investigate the cause of the
sudden disappearance of the water.

39 In 1905 it began rather late. On November 8, in Fern Hollow and Nine-Mile Run, no fresh chimneys were
seen on the golf links, where C. dingenes is abundant. On November 22 fresh mud-piles were numerous at the ident-
ical locality, several sharp frosts having occurred in the meantime. The same dates and the same facts were ascertained
for C. mnnungalensis in Fern Hollow. Mr. Kelly's observations were made November 14 and 15, 1904, but in 1904 I
noticed fall activity as early as November 5 (Nine-Mile Ruu, ft diogenes). (See PI. XLI, Fig. 5.)

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 425

B. Geographical Distribution.

(Plates XLII and XLIII.)

1. Cambarus limosus.

a. Summon/ of Facts (see above, pp. 356-358).

This species belongs to the rivers, ponds, and canals of the lowlands of the At-
lantic Coastal Plain and the Piedmont region 10 in the states of New Jersey, Penn-
sylvania, Maryland, District of Columbia, and Virginia. It has not been reported
from the State of Delaware, but must certainly occur there also.

In Pennsylvania it extends up the rivers to a certain distance and in the Sus-
quehanna River enters the Alleghany Mountain region. In Maryland it goes up
the Potomac River, reaching the eastern extremity of West Virginia, thus also en-
tering the Alleghany Mountain region.

Thus it is found in Pennsylvania in the drainages of the Delaware, Susque-
hanna and Potomac Rivers ; but it decidedly prefers the region of the lower Dela-
ware, from the bend of the river at Trenton downward. Here it is exceedingly
abundant, as also in the lower and quieter parts of the Schuylkill River at Philadel-
phia. It goes up the Delaware and Schuylkill, and is found in their tributaries
within the Piedmont region, but here it is by no means as abundant as in the Dela-
ware. It seems to be absent in the great Alleghany Valley between the Susque-
hanna and the Delaware, but reaches the foot of the Blue Mountain between the
Susquehanna and Potomac, occupying the Cumberland Valley (part of the great
Alleghany Valley), and in the Susquehanna and Juniata it ascends even further,
far into the Alleghany Mountains (Center and Bedford Counties).

b. Origin, of the distribution of C. limosus.

In Pennsylvania.

In the Delaware River above Trenton this species goes up as far as New Hope-
in Bucks County; but is very rare there, (only one specimen was secured by the
writer after a prolonged search), and it seems that it does not go far beyond this
point, if at all. Professor A. E. Davison informs me that it is not found near
Easton, Northampton County, about ten to fifteen miles from the Blue Mountain,
and I was unable to find it in the Little Lehigh Creek near Emaus, Lehigh County.

40 As to the division of Pennsylvania into Coaxtal Plain, Plmhnani Plateau, Omit AVeghany V II- 1,. \tlrghnnii !/-.« •
tains and Alleghany Plateau, see Davis, 1889, p. 187, and Hollister, 1904, p. 10, map. Fig. 1 ; also Powell, I -90 p.
7Se< sen. and map, and Willis, 189u, p. 169,

426 MEMOIRS OF THE CARNEGIE MUSEUM

Mr. W. R. McConnell mentions (in his notes) the absence of crawfish in the Dela-
ware at Portland, Northampton County.

It is found, however, in small tributaries of the Delaware in the southeastern
half of Bucks County (Neshaminy Creek). In the Schuylkill River it goes up to
Reading (Girard and also McConnell), and slightly beyond (Maiden Creek) in Berks
County, but I have not been able to find it in the Schuylkill, where it comes out of
the Blue Mountain, (Shoemakersville). It has been reported from Brandywine
Creek in ( 'hester ( 'ounty.

It is known from a number of places in the drainage of the Susquehanna, but
they are all in the region of the Great Alleghany Valley or the Alleghany Moun-
tains. I was unable to find it in the Susquehanna in Lancaster and York Coun-
ties, (Pequea and York Furnace), and I do not think that it is present there on
account of the roughness of the river, which flows over a rocky bed in a channel
cut deep into strata, chiefly of the archaic age, belonging to the Piedmont Plateau,
from York Haven to the Maryland state-line and beyond. Such conditions are de-
cidedly unfavorable for this species, and it is rather strange that it should be found
at all above this rough part of the Susquehanna, which is about thirty to forty
miles long. I think that this species immigrated into these parts in very recent
times by way of the Susquehanna and Pennsylvania canals, which closely followed
the river from its mouth in Maryland to the New York state-line and the Juniata
up to Hollidaysburg, and connected it with the Schuylkill. These canals were
maintained and in use a long time, beginning as early as 1834, were abandoned
about 1890,41 and at present only remnants of them are seen. C. limostis is often
found in canals. First reported by Faxon from near Cumberland, Maryland, I
have found it in considerable numbers in the Schuylkill and Delaware and Raritan
( lanals. It is quite possible that the Susquehanna and Pennsylvania Canals afforded
this species the means of reaching the Susquehanna River in the region of the Great
Alleghany Valley south of Harrisburg. Its further distribution up stream is then
not strange, after the rough portion of the lower Susquehanna had been overcome,
or avoided.

The same may be true of the Schuylkill River. Although certainly originally
present in the lower part, it was the Schuylkill canal (once connected with the Penn-
sylvania canal) which possibly afforded an opportunity for C. limrmis to go up the
river as far as it does now, since the Schuylkill above Philadelphia is rather rough.

" The main line of the canal was completed in 1834, the Susquehanna Canal from Columbia to Havre de Grace in
1840; see Jenkins, 1903, pp. 275, 277, 282 and Klein, 1900, p. LXXIX ; see also Hoyt & Anderson,, 1905, p. 24. In
the latter paper fine views of the scenery of the lower Susquehanna are published (PI. 1, B, PI. 8), which convey a good
idea of the roughness of the water of the river.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 427

Similar conditions seem to have played a part in the distribution of this species
in the Potomac River. It has been reported from an old canal (Chesapeake and
Ohio) four miles south of Cumberland, and I have found it in the Potomac at
Cherry Run, West Virginia, and at South Cumberland. At both places it was scarce,
and I am much inclined to believe that in this region (western Maryland and east-
ern West Virginia) it got into the river from the canal. Originally its distribution
in Maryland was very likely similar to that in Pennsylvania, belonging only to the
Coastal Plain and the Piedmont Plateau.42

Of the tributaries of the Potomac in southern Pennsylvania those which empty
into the Potomac east of the Alleghany Mountain region also possess this species.
It has been found in the drainage of the Monocacy River at Gettysburg, Adams
County, and that of the Conococheague Creek in Franklin County, the latter local-
ity again belonging to the Great Alleghany Valley. Further west, within the Alle-
ghany Mountains, it seems to be absent. I did not find it in Big Cove Creek and
Tonoloway Creek, Fulton County, and it is not in the collections made by Mr. H.
A. Pilsbry for the Philadelphia Academy in Sideling and Town Creeks, Washington
and Alleghany Counties, Maryland. This supports the view that the presence of
this species in the Potomac as far up as Cumberland is due to the existence of the
canal. Above Cumberland, where the canal ends, C. limosus is positively absent in
the Potomac drainage in Pennsylvania as well as in Maryland and in West Virginia.

Thus it seems that C. limosus belongs originally only to the larger rivers of the
southeastern section of our state, and that its real center for Pennsylvania is the
Delaware. It has spread, however, upstream, and has approached the Alleghany
Mountain region, even entering the latter in the Susquehanna River. This upstream
dispersal is apparently not everywhere due to natural migration, but has been
favored in recent times by canals. The present northwestern boundary, disregard-
ing the Susquehanna River, is marked by a line (see PI. XLIII) running from New
Hope, Bucks County, to Maiden ('reek and Reading, Berks County, thence to Bain-
bridge, Lancaster County, Carlisle, Cumberland County, and to Williamson, Frank-
lin County. This line, generally speaking, runs parallel to the Blue Mountain, and
it is very likely that the differences in the physical features of the Piedmont Plateau
and the Alleghany Mountains have something to do with the distribution of this
species, although the real cause cannot any longer be clearly seen, the original con-
ditions being apparently obscured by several factors. For it should not be forgotten
that the streams from the Susquehanna to the Delaware, issuing through the Blue

"The Chesapeake and Ohio Canal forms a continuous waterway from Washington to Cumberland , and was com-
pleted in 1850, see Hulbert, 1904, p. 160, and map opposite p. 80.

428 MEMOIRS OF THE CARNEGIE MUSEUM

Mountain from the Anthracite basin, are largety charged with mine-water, and in
this section of the state (Berks and Bucks counties) we see that G. limosus does not
so closely approach the Blue Mountain, while in Cumberland and Franklin Coun-
ties, where the streams are clear, it goes to the very foot of the mountain.

Of course we cannot any longer ascertain what the original conditions were, and
thus it is hardly profitable to enter into any further speculations. It is probable
that the original range of this species has been reduced on the one hand by pollu-
tion of the streams, and has been extended on the other hand by modern river im-
provements. How far this holds good in detail, remains doubtful.

General origin of the distribution of G. limosus.

Aside from the more recent dispersal of this species just discussed, we are
prompted to inquire how this species was able originally to reach the parts where
it is now found.

As the writer has pointed out in a former paper (19056, p. 108, 111, 114, 127)
('. limosus stands rather isolated geographically as well as morphologically. It
belongs to an ancient group of the subgenus Faxonius, probably the most ancient,
which consists of five species. The other four species are entirely removed geograph-
ically from G. limosus, and are found in the central basin of the United States, in
Kentucky, Indiana, and Missouri, that is to say, about four hundred miles to the
west of the range of G. limosus, with the Appalachian System between them. We
have to deal here with a marked case of discontinuity of distribution in the limosus-
group. Since, as has been shown by the writer in the paper referred to, we locate
the center of the subgenus Faxonius in the central part of the Mississippi drainage,
('. limosus must have reached its present home by migration, and there are several
ways by which it may have gone.

The most direct route is across the Alleghany Mountains. We may suppose that
the limosus-group once extended in the Ohio drainage up into western Pennsylvania
and West Virginia, and that it was able by some means to cross the divide into the
Atlantic drainage. This does not appear impossible, inasmuch as in the mountains
stream-piracy has taken place on a large scale during all ages (Davis, 1889). In
fact all of the larger rivers now running into the Atlantic have captured large tracts
originally belonging to the interior drainage, and the divide has been continuously
shifted westward.

On the other hand, considering the ecological peculiarities of C. limosus, this as-
sumption does not appear very likely. The habit of living in larger streams in
rather quiet water would not favor a migration across the mountains, and if this

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 429

form actually came by the way indicated some traces of its former existence should
have been left in Pennsylvania, Maryland, or the Virginias, chiefly since there was
no competition by any other species, river-forms being absent in the Alleghany
Mountain region. Thus the direct route across the mountains seems to be out of
the question, and this is further rendered probable by another consideration.

C. limosus being ancient, its migration eastward must have taken place at a re-
mote epoch, certainly at an earlier time than that of a group which is more advanced,
namely, the propinquus-growp. As we shall see below, the latter existed already in
Preglacial times, and thus we are forced to place the origin of the limosus-growp at
least as far back as the Tertiary. During this time, however, the Ohio in its present
form did not exist. There was Spencer River,48 in West Virginia and western Penn-
sylvania, and another river (Old Kanawha)44 in West Virginia and Ohio, running
northward to the Erigan River, which transversed the basin of Lake Erie.45 And
further the present upper Susquehanna (North Branch) is apparently new. It must
have taken in Preglacial times a northward route toward the St. Lawrence basin,
possibly also to the Erigan River (White, 1896, p. 376). All these rivers flowing
noAward in Pennsylvania and Ohio were different in character from what the
ri#irs of this region are now. Their fall was slight, and they were rather sluggish.
This is positively known of the Spencer River (or the Old Monongahela), which
must have been practically at base-level (White, 1896, p. 377). If this was the case,
nothing is opposed to the assumption that C. limosus (or its ancestral form) once was
an inhabitant of some of these rivers. But then we see that its eastward migration
cannot have been in a direct route, but must have gone on in a roundabout way>
chiefly by the old Erigan River.

If the Erigan River was tributary to the Mississippi system, this is easy to
imagine. If it drained to the St. Lawrence Gulf, as Spencer believes, we must
assume an earlier crossing of the continental divide by this form, wherever this was
situated (Indiana?), and then again a crossing of the divide between the Erigan
River and the Atlantic coast drainage.

Be this as it may, we are forced to move the old range of the Rmoms-group to
the north, into the Erigan River drainage, and this gives us the means of explain-
ing the discontinuous range of this group. If it were at one time present in an area
extending from Kentucky and Indiana through Michigan into Ontario, and if we
assume that it crossed over into the Atlantic drainage somewhere in northern Penn-

" See Foshay, 1890, p. 368 ; Leverett, 1902, p. 89.

"See White, 1896, p. 376 ; Leverett, 1902, p. 100 ; Tight, 1903, map, Plate I, Plates 16 and 17. (Teays Kiver.)

"See Spencer, 1881, map 2, and 1894, p. 293.

430 MEMOIRS OF THE CARNEGIE MUSEUM

sylvania or New York, the advancing ice of the Glacial Period must have entirely
covered a large part of this range. In the central parts, in Ohio and western Penn-
sylvania, it was impossihle for these forms to retreat southward, these parts being
occupied by another vigorous group of river-crawfishes, as we shall see below (pro-
pinquus-gromp), and only in the east and west a chance to survive was left. The
eastern remnant is the present G. limosus, the western is the group of species found
now in southern Indiana and Kentucky.

How G. limosus reached the Atlantic Coastal Plain from the Erigan basin is very
hypothetical. One suggestion may be made. Not only does the North Branch of
the Susquehanna seem to be a reversed river, but the West Branch has captured
a large part of the original drainage of the Alleghany Plateau in Potter, Cameron,
and Clearfield Counties. Davis (1889, p. 248) believes that this happened largely
in Pretertiary times, since he thinks that the Alleghan}' Plateau belongs to the
Cretaceous peneplain. However, Campbell (1903, p. 2<S0) has shown that there
are two old base levels in northern Pennsylvania, an older one (Cretaceous), iden-
tical with that of Davis, and a younger one (1,600 to 2,200 feet) corresponding to
the Harrisburg peneplain of Old Tertiary age. Since the headwaters of the West
Branch of the Susquehanna are carved into this second peneplain, it is probable
that during Tertiary times the stream-piracy of the Susquehanna was going on
rather vigorously. If we assume that G. limosus in Tertiary time existed in this
part of the Erigan River drainage, namely in the Old Upper and Middle Alleghany
Rivers,46 which did not belong to the Old Monongahela or Spencer River, it must
have been possible for it to get into the Susquehanna drainage in consequence of
this stream-piracy in Tertiary times. This, however, is a mere suggestion. There
is no other evidence for it but the bare fact that stream-piracy has gone on in this
region. I mention it here only to show that the crossing over of this species into
the Atlantic drainage is not altogether unthinkable.

After arriving in the coastal plain C. limosus was cut off in the Glacial Period
from its allied forms in the west. But it survived, and in Postglacial times was
able to advance again. But the Postglacial dispersal cannot have amounted to
much, since the increasing roughness of the streams, caused by the Postglacial eleva-
tion of the country, was not favorable to a northward expansion. We do not know
the exact northern boundary of G. livwsus outside of our state. It is found in New
Jersey as far north as Morris County, yet we do not know whether it reaches Rari-
tan and New York Bays, and the Hudson River. No positive record from New
York State is at hand (see De Kay, 1841, p. 23, and Paulmier, 1905, p. 117).

"See : Carll, 1880, pp. 333 and 336, map, PI: 2 ; Leverett, 1892, pp. 129 and 132 ; Tight, 1903, map, PI. 1.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 431

An alternative supposition might be entertained. G. limosus might have arrived
in its present home coming from the south by way of the Atlantic Coastal Plain.
This, however, does not seem probable. First of all, the distribution of G. limosus
does not extend southward beyond Virginia, and even in Virginia it is known only
from a few localities. Southward no representative of this group is known on the
coastal plain, and, if C. limosus had come from the south, traces of this migration
might be expected. On the other hand, if it came from the north, as we here
assume, the fact that it did not spread beyond Virginia may be accounted for by the
presence of another group of this genus, the subgenus Gambarus (blandingireection),
in the southern parts of the coastal plain, which, like C. limosus, prefers ponds and
sluggish streams. Indeed both species ( ( ' limosus and G. blandingi) are found actually
associated at the same localities (by Faxon, 1885o, p. 88, at Trenton, New Jersey,
and by the writer in the Delaware and Raritan Canal at Princeton, New Jersey),
but we must bear in mind that in New Jersey, and also in Maryland and Virginia,
C. blandingi is an intruder, its chief domain being in the Carolinas.
The following are conclusions from the above considerations :
Cambarus limosus is an ancient species, characterized by morphological and
geographical isolation. The most closely related forms are found in Kentucky and
southern Indiana. An attempt to explain the presence of G. limosus at its present
location has to connect its range with that of these related species. A connection
by way of the Atlantic Coastal Plain southward is out of the question. Thus only
the connection across the Appalachian system remains. The fact that the rivers
just west of the mountains in western Pennsylvania are occupied by a more
advanced group of species (propiriqims-groTip) of a subgenus which is certainly of
Preglacial age, as we shall see below, leads us to the conclusion that the limosus-
group also must be not only Preglacial, but older than the propiwjuus-group. But
at that time there was no direct way from the lower Ohio, where its center of dis-
tribution was situated, into western Pennsylvania and across the mountains, the
Ohio having no existence as yet, and the general drainage in this region being to
the north. This leads us to assume a former more northern range of the limosus-
group, extending into Preglacial Canada; and this assumption furnishes an expla-
nation why it was possible for the Glacial Epoch to cut the range of the limosus-
group in two, leaving no representative of it in the region now drained by the
middle and upper Ohio. Gamba/rus limosus is a Tertiary type, mid /'/ reached its pres-
ent area coming from the west awl by way >>/ the north, being driven smith along tin
Atlantic Coastal Plain by the advancing ice of the Glacial Period. It survived during
the Glacial Period in the region of the lower Delaware River mid Ghesapeah Hag. while

432 MEMOIRS OF THE CARNEGIE MUSEUM

all the rest of the former range of the group was covered by ice and its representatives were
destroyed, with the exception of a small remnant in the southwestern portion of the range,
in southern Indiana and Kentucky, outside of the glaciated area. The reason why
this group was destroyed in the glaciated area, and was not able to retreat
southward and to survive in the intervening parts (Ohio, western Pennsylvania,
and West Virginia), was that here the rivers were occupied by another group of
the subgenus.

The above is a mere theory, and it remains doubtful by which way C. limosus
reached the Atlantic Coastal Plain. The assumption that it was by way of the
Erigan River and the St. Lawrence basin satisfactorily accounts for the facts, but
this is the only point directly in favor it. However, the study of the distribution
of G. limosus is not yet finished, since the actual boundaries of the distribution,
chiefly to the north and south, are not positively known. But this does not con-
cern us at present, since they are not situated in the State of Pennsylvania.

In G. limosus we have a species which survived during Glacial times in a part of
the Atlantic Coastal Plain which is well to the north, not far from the southern edge
of the ice. Of course this forms a part of Adams' (1902, p. 121) southeastern center
in its widest sense, lying at its northeastern extremity. Although surviving not far
from the edge of the ice, C. limosus cannot be considered as belonging to the tundral
biota (Adams, 19U5, p. 58), but it belongs very likely to the second wave (north-
eastern biota), Avith a slight suggestion of the third wave (southeastern biota) (/. c,
pp. 58 and 62). As Adams indicates, the first and second waves of Postglacial dis-
persion had their glacial homes in very narrow belts parallel to the southern edge of
the ice, while the southeastern (and southwestern) biota covered in Glacial times
wide tracts of country. The second wave largely invaded the coniferous forest-belt
of ( 'anada, while the third wave was more stable and did not spread so far north-
ward. With regard to its geographical location during Glacial times, G. limosus
should be classed with the northeastern biota ; and with regard to its stability in Post-
glacial times, with the southeastern. But we are to consider that a Postglacial north-
ward dispersion was rendered difficult in this case by the physiographical features
of the country. The coastal plain with its sluggish streams and stagnant ponds
disappears in northern New Jersey, the uplands (Piedmont Plateau) reaching the
coast in the vicinity of New York Bay (see McGee, 1 888, PI. 2) ; this did not offer
advantageous conditions for this species, and thus it remained within comparatively
narrow limits in a corner, into which it was pushed in Glacial time. G. limosus is a
Tertiary relic at the northern extremity of the coastal plain, which has not been able
to expand its area to any considerable degree in Postglacial times.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 433

2. Cambarus propinquus, Cambarus propinquus sanborni and Cambarus obscurus.
a. Summary of Facts. (See pp. 362-363 ; 368-360; 372-37:!.)

If we desire to arrive at a proper understanding of the distribution of ('. propin-
quus, C. propinquus sanborni, and ('. obscurus, they must be discussed together.

The area occupied by these three forms (see PI. XLII, Fig. 3) includes eastern
Iowa, southern Wisconsin, northern Illinois, Indiana, Michigan, Ohio, northeastern
Kentucky, northern West Virginia, western Pennsylvania, western New York, and
parts of Canada (Ontario and Quebec). In the western and northern part of this
range C. propinquus is found; C. propinquus sanborni occupies the larger part of
Ohio and parts of Kentucky and West Virginia; while C. obscurus has its chief
domain in western Pennsylvania, passing southward into West Virginia and north-
ward into New York. Thus it is apparent that the three forms occupy different
sections of the general area of the group, propinquus being western (and northern),
sanborni central, and obscurus eastern. As far as observations go all three forms are
rather (sharply separated geographically, although they come into contact at the
edges of their ranges. This is especially true, as we have seen, in our state and the
adjacent portions of Ohio and West Virginia, while in western Ohio and in Indiana
nothing is known of the boundaries of the forms represented there.

In Pennsylvania only two of these forms are found (PI. XLII, Fig. 2). C. propin-
quus is restricted to Lake Erie and its drainage ; C obscurus belongs to the ( >hio sys-
tem, and is found everywhere in the western section of the state, in the Ohio, Mo-
nongahela, and Alleghany Rivers and their tributaries. The boundary toward the
east is formed by the divide between the Alleghany and Susquehanna systems, and
farther south generally by the Chestnut Ridge (with exceptions to be discussed
below). Northward this species crosses over into the Genessee drainage, and extends
into New York. It also crosses over into the Lake Erie drainage in Pennsylvania.

Along the western border of the state it passes beyond the state line into Ohio,
the drainage belonging in the northern part to the Beaver River. Furthermore it
goes down the Ohio and is found in all creeks running from Pennsylvania through
the Panhandle of West Virginia as far south as Fish ('reek in Greene County,
Pennsylvania, and Marshall County, West Virginia. Fish ( "reek falls into the Ohio
a little below Moundsville, West Virginia, and contains only the typical form of
C. obscurus.

Going further down the Ohio conditions suddenly change. In Fishing Creek,
Wetzel County, West Virginia, which empties into the Ohio near New Martinsville,
about thirteen miles below the mouth of Fish < "reek, C. propinquus sanborni appears.
But the form here found is not typical. As we have seen above, it inclines some-

434 MEMOIKS OF THE CARNEGIE MUSEUM

what toward C. obscu/rus, and one individual has been found which represents
typical C. obscurus. In Middle Island Creek near St. Mary's, Pleasants County,
West Virginia, which is about twenty-five miles further down the Ohio, the few
specimens collected seem to be typical C. propinquus sanbomi.

Thus it appears that C. obscurus goes down the Ohio River to about Mounds-
ville, West Virginia. All the tributaries of the Ohio in the Panhandle possess this
species, and very probably it will be found also in Ohio on the opposite side of the
river. But crossing over the divide between this part of the Ohio and the Muskin-
gum-Tuscarawas River in Ohio, we again find G. propinquus sanbomi in the drain-
age of the latter. The western boundary of C. obscurus consequently is formed by
the divide just mentioned, but this line crosses the Ohio River between Mounds-
ville and New Martinsville, West Virginia (PI. XLII, Fig. 2, and PI. XLIII).

Further to the south in West Virginia in the drainage of the upper Mononga-
hela this species has not been traced. It surely goes up the Monongahela beyond
the southern boundary line of Pennsylvania, but how far has not been ascertained.

The fact that G. obscurus is found also in the Potomac drainage, in Wills Creek,
between Hyndman, Bedford County. Pennsylvania, and Ellerslie, Alleghany County,
Maryland, deserves special mention, and will be commented upon elsewhere.

b. Origin of the distribution of C. propinquus, propinquus sanbomi, and C. obscurus.

In order to get a fair understanding of the distribution of these forms, we must
take notice of the Preglacial physiography of the region in which they are found,
for, as we shall see below, Ave are led to believe that these forms are of Preglacial
age, and survived during the Glacial Period in the southern parts of the drainage
systems, which now constitute that of the Ohio.

First of all, we should bear in mind that at the end of the Tertiary Period before
the ice pressed down from the north, the Ohio River in its present form did not
exist. In the whole region, drained now by the middle and upper Ohio, the drain-
age was at that time not to the west, but to the north, and it was collected by a
river running in a northeasterly direction toward the present Gulf of St. Lawrence,
(the Erigan River or Ancient Grand River).47

Disregarding some smaller streams, for instance the Old Middle and Old Upper
Alleghany, which do not concern us here, three main rivers, tributary to the Erigan
River, have been traced with more or less accuracy, and the evidence for their ex-
istence, although fragmentary, leaves no doubt as to the general correctness of the

"This is the opinion of Spencer (1881 and 1894). Others helieve that this river drained toward the Southwest,
into the Mississippi ; see (Jrahau, 1901, maps, p. 44 and 45 (Dundas River).

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 435

main features of this drainage, which differs so strikingly from that which exists
to-day.

The easternmost of these rivers was the Spencer River, or Old Monongahela, or
Old Upper Ohio.48 which drained southwestern Pennsylvania, northern West Vir-
ginia, and a small part of eastern Ohio. West of it was the Old Kanawha River,
or Old Middle Ohio, or Teays River (Leverett, 19U2, p. LOO, map, p. 101 ; Tight,
1903), which drained parts of West Virginia and Kentucky, and the larger part of
central Ohio. The old Muskingum-Tuscarawas River belonged to this drainage,
the Muskingum River not flowing southward, but westward and southwestward
from near Zanesville, Ohio, to ( 'ircleville, Ohio, thus joining the Old Kanawha
(Newark River; Tight, 1903, PI. 1).

The divide of the Old Kanawha to the westward was formed by the Cincinnati
uplift, and was situated according to Leverett (1902, p. 100) near Manchester, Ohio,
on the present Ohio River. Beyond this divide we have the Lower Ohio system
(Leverett, p. 109). The Preglacial lines of discharge in this region are rather
obscure, but according to Leverett and Newsom (1902, p. 168, PI. 6) it is probable
that a large part of the present system of streams was tributary to the lower Ohio in
Preglacial times, but that a small number of them may have had a northward dis-
charge through the Great Miami basin in western Ohio (Leverett, p. 116). There
are distinct indications of a northward drainage in the vicinity of Cincinnati (Cin-
cinnati River, Tight, 1903, PI. 1). This possibility is also admitted by Newsom
(1902, p. 181).

We may take it for a well established fact that in Preglacial times at least two
rivers existed in this region, the Spencer and the Old Kanawha, which did nut
drain into the Ohio and Mississippi in a southwestern direction, but Mowed north-
ward into the Erigan basin. Westward there was very likely a third river ("Old
Miami") running in a similar direction; but in this region we arrive at the old
Preglacial divide between the Lower Ohio and the Erigan River. It remains
doubtful whether the latter drained to the St. Lawrence Gulf or to the Mississippi
by the way of the present Wabash.

Assuming the theory of the former existence of an Old Miami (or Cincinnati)
River, we see that there are certain interesting relations of these three old rivers to
the present distribution of the three forms of Cambarus under discussion.

Of course, we must disregard those parts of the ranges of these forms which lie

,sSee above, p. 429. Descriptions are given by Foshay, 1890, White, 1893, and Leverett, 1902, p. 88 (with map
on p. 89). Additional evidence has been furnished by Hice, 1903, p. 302. Another name is Hittsbuigh River i Tight,
1903, PI. 1).

43(3 MEMOIRS OF THE CARNEGIE MUSEUM

in the formerly glaciated area, for these are due to Postglacial expansion. But
looking upon the localities south of the terminal moraine (PI. XLII, Fig. 3) we see
that only a few are known for C. propinquus, and these are all in southern Indiana
(Brown, Monroe, and Green Counties),49 and belong very likely to the old Lower
Ohio drainage, but in the region where it comes into contact with the supposed Old
Miami River (or possibly some other river flowing north in the State of Indiana).
Since we have reason to believe (Ortmann, 19056, p. 114) that the center of radia-
tion of the subgenus Faxonius, to which the propinquus-group belongs, is in the
central basin formed by the three great rivers (Missouri, Mississippi, and Ohio), C.
propinquus distinctly points toward this center, of which southern Indiana forms
part. This is the more interesting since we see that it is the most primitive species
of the propinquus-gvoup which most closely approaches the original center. In
Preglacial times ('. propinquus belonged to the northeastern extremity of the old
Ohio drainage (Lower Ohio), and in this region there apparently was a chance for
it to cross over the continental divide into the Atlantic (St. Lawrence) drainage.
If, however, the Erigan River drained to the Mississippi, the presence of this
species in the Lower Ohio and in the lower part of the Erigan River is more easily
accounted for by direct communication of the waters.

Taking up the distribution of G propinquus sanborni, we observe that until
recently only one locality was known to the south of the drift, namely, the type
locality in Carter County, Kentucky, which is undoubtedly in the drainage of the
Old Kanawha River. In addition, I have discovered a number of localities in
eastern Ohio (Carroll, Harrison, and Stark Counties), and in northern West Virginia,
which belong to the same drainage (Newark River and Marietta River, tributaries
of the Old Kanawha), which are also outside of the glaciated area (at Canton,
Stark ( Ymnty, Ohio, close to the edge of the drift).

The chorological facjs about the distribution of G. propinquus and C. propinquus
sanborni are very meager, and not at all satisfactory; but as far as our present
knowledge goes, all known localities of G. propinquus sanborni, outside of the drift,
are in the drainage of the Old Kanawha, while none of the known localities of C.
propinquus are in this drainage, but are situated to the west of it. Now, this
mutual relation between distribution and Preglacial drainage becomes more evident
when we look upon G obscwrus, the distribution of which I have studied more
closely.

"They are close to the southern edge of the drift, and it is a little doubtful whether they are inside or outside of
it. Disregarding the Illinoisan drift, they are surely outside of the Postillinoisan glaciation, as is also a locality in
Franklin County, Ind.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 437

Leverett (1902, p. 89, Fig. 1) has given a map of the Old Monongahela River,
which is reproduced on Plate XLII, Fig. 1, and alongside of it, Plate XLII, Fig. 2,
I give a map of the present distribution of C. ohscurus. It is evident at a glance
that there is close correlation between C. ohscurus and this old river. The most
important features are furnished by the western boundary. The divide between
the section of the Ohio which runs along the Panhandle of West Virginia and the
Muskingum-Tuscarawas drainage is the old divide between the drainages of the
Spencer River and the Old Kanawha. This divide crosses the present Ohio just
above New Martinsville (see Leverett's map, p. 90, Fig. 2: "probable early di-
vide"). As I have found (p. 434) this old divide coincides with the present boundary
between C. ohscurus and G. ohscurus sanborni. It also is significant that it is not the
Tertiary (Preglacial) divide, which is located by Tight (1903, PL 11) just below New
Martinsville, nor the "later divide" of Leverett (I.e.), but just the one which
existed at the beginning of the Glacial Period. We shall have to return to
this topic.

Thus it is clear, first, that the original separation of these two forms was brought
about by the fact that they belonged to different river systems ; second, that we
must assume the Preglacial age of the propinquus-grouj) ; and third, that the distri-
bution of these crawfishes furnishes additional evidence for the correctness of the
view of the Old Monongahela and Old Kanawha, as held by Leverett (and others) ;
and with reference to these crawfishes it seems to me that the following theory is
rather well founded.

In Preglacial times, the propinguus-group, coming front the south /rest (lower Ohio)
reached the Erigan River drainage {either directly or by crossing a divide), of which it
became characteristic.™ It entered, consequently, also the southern tributaries of this
river, and owing to the fact that there were time main tributaries, this group developed
the tendency to split up into as many geographical forms. These were apparently the
conditions when the Glacial Period began.

The chief effect of the advancing ice was that the northern parts of the range of
this group were covered by ice. Only in the region of the headwaters of these
rivers, to the south of the edge of the ice, was there a chance to survive, and sur-
vival here occurred. Both the Old Monongahela and the Old Kanawha were

50 We have seen that a similar dispersion very likely took place in the case of the tinuwiw-gronp. The latter being
more primitive, we must assume that it formed a first and earlier wave of immigration from the Lower Ohio into the
Erigan drainage, while the propinquus-gronp came later. This movement is still going on. There is evidence of B sub-
sequent Postglacial wave (later than the Postglacial migration of 0. propinquus) also starting from the Lower Ohio, and
represented by the rusticus-group, which again has all the morphological marks of a more recent type than the proptn-
jnus-group. But this is outside of the scope of the present paper.

438 MEMOIRS OF THE CARNEGIE MUSEUM

dammed up by the ice, and transformed into lakes (Lake Monongahela of White,
1896, and Lake Ohio, cf. Jillson, 1893, p. 19, and Map, PI. 5, with the necessary
restrictions), and this led to the result that the colonies of crawfishes belonging to
the southern (upper) parts of these rivers became sharply separated from each other,
and I think that the tendency toward the formation of three species (C. propinquus, G.
sanborni, ('. obscurus) is direct!;/ due to this process ant] to physiographical conditions
'prevailing in the earlier part of the Glacial Period (Kansan or Prekansan, cf. Hice,
1903, p. 300).

Finally these lakes were connected and drained off toward the southwest, thus
forming the present Ohio River (Postkansan, but before the Wisconsin stage, cf.
Hice, 1903, p. 299); the areas of the three forms of crawfishes were reunited, but
the different parts of the new Ohio drainage are occupied by different forms of the
prcjpincpi its-group, remaining in their original areas; the upper Ohio is character-
ized by C. obscums, the middle Ohio by ' '■ propinquus sanborni, and the lower Ohio
by C. propinquus.

But additional changes took jda.ce in Postglacial times. According to the pres-
ent distribution these must have been greatest in the case of G. propinquus.
Almost the entire range of this form lies within the glaciated area, and thus it is
beyond question that its present distribution is largely due to the Postglacial migra-
tion northward and northeastward.51 This migration possibly began at an earlier
date than in the case of the other two forms. We know that in southern Indiana
and southwestern Ohio an early retreat of the southern border of the ice took place,
as is indicated by the presence of Illinoisan drift south of the early Wisconsin border
(cf. Leverett, 1902, PI. 2 and PI. 11). There also was considerable recession of the
ice of the Maumee- Miami glacial lobe in the earlier and later Wisconsin stage, while
in central and eastern Ohio and western Pennsylvania (Scioto glacial lobe and
Grant River glacial lobe) only in the later Wisconsin stage did recession take place
(cf. Leverett, ibid., and PI. 13 and PI. 15). Toward the end of the later Wisconsin
stage large lakes began to form in front of the receding ice, and this happened first
in the western part of this region. The first lake thus formed was Lake Maumee
(Leverett, p. 710 ff, IT. 20 and 21), which had an outlet toward the west and south-
west (Fort Wayne outlet; see also Grabau, 1901, p. 58). Lake Maumee was situ-
ated in northwestern Ohio, in the present Huron-Erie basin, and thus we see that
the latter was opened first in its western part to an immigration from the southwest

51 A loss of territory must have occurred in the south, C. propinquus losing ground in competition with C. rusticus
which was pushing on from the south. This matter does not belong to our present investigation, but attention should
be called to it.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 439

(Indiana) at a time when this basin was entirely covered by ice further east, thus
being closed more or less to an immigration from the central parts of Ohio (drain-
age of middle Ohio), and being closed entirely to an immigration from eastern
Ohio and western Pennsylvania (drainage of upper Ohio).

This explains why G propinquus, which survived in southern Indiana, had the
first chance to spread northward and to enter the future Huron-Erie basin by way
of Lake Maumee. The subsequent stages of this lake (Lake Whittlesey, Lake Warren,
etc.), are all direct continuations in time of Lake Maumee, and so it is not astonish-
ing that G propinquus, after the final establishment of the St. Lawrence drainage,"
is found all over this region, not only in the Lake Huron and Lake Erie basins, but
also farther down, in Lake Ontario and the Lower St. Lawrence drainage. In the
occupation of this whole region G. propmquus was not interfered with by the other
forms, since no opportunity was given to G propmquus sanborni and G obscurus, to
enter the Erie basin, the drainages of the middle and upper Ohio remaining perma-
nently changed to the southwest, away from Lake Maumee, a condition which ob-
tains, with very slight changes, up to the present time.

However, G propinquus sanborni as well as G obscwrus, have entered the hake
Erie drainage. With regard to the first, it may be sufficient to state that it is found
in Lorain County, Ohio, in rivers and creeks running into the lake, and this is ap-
parently due to a comparatively recent immigration under similar conditions as in
the case of G obscwrus in Pennsylvania. The latter species has been discovered by
the writer in Crawford and Erie Counties, Pa., in streams Mowing to Lake Erie,
associated with the Lake Erie form, G. propinquus. Thus G. obscurus must have
crossed the divide between the upper Beaver (Shenango) River and Alleghany
River (French Creek) on the one side, and Lake Erie (Conneaut and Elk Creeks) on
the other, and the question is by what means this was accomplished.

It is only natural that ('. obscurus, surviving during Glacial times in south-
western Pennsylvania and West Virginia, migrated up the drainage of the upper
Ohio, chiefly the Beaver and Alleghany Rivers, in Postglacial times, for after
the end of the Glacial Period this system formed a unit", and no serious barriers
to the dispersal were, or are, present. Thus it was easy for this species to go up

52 The change of the westward drainage to an eastward took place toward the end of the Glacial Period, as soon as
the ice receded far enough to uncover Lake Ontario (Lake Iroquois), thus permitting the water to drain oil through the
Mohawk, and later through the St. Lawrence. This was accompanied probably by a depression of the lan.l in the
Northeast, culminating in the marine invasion of theSt Lawrence valley (Champlain submergence). (SeeGrabau, 1901,

p. 59 e t seq. )

» As to the formation of the present Alleghany out of the former Lower, Middle, and Upper Alleghany, see Lev-

erett, 1902, p. 129 ti seq.

440 MEMOIRS OF THE CARNEGIE MUSEUM

toward the bead-waters of these rivers and to closely approach the divide toward
Lake Erie.54

This would favor a direct crossing of the divide by actual migration over land,
and indeed the river-species are able to survive when out of the water for a consid-
erable time under certain circumstances, as I have ascertained by experiments.
During hot and dry weather it is hardly possible to keep them alive for more than
an hour or two ; but in cook cloudy, and damp weather I have found that speci-
mens suspended on a string on an open veranda55 were not dead after seven hours,
and restored to water, recovered entirely. This might at least render a
migration over land possible, but I do not think that it actually takes place,
since it has never been observed, either by others or by myself, that G. ob-
scurus, or any other species classed ecologically with the river-species, leaves
the water voluntarily. On the other hand it is possible that G. obscurus may
undergo a passive transport from one drainage to the other, as for instance by birds.
However, I do not believe that the crossing of the divide toward Lake Erie is due
to the latter cause. It seems to me highly improbable, not that birds should be
able to carry crawfishes for a long distance, but that it should happen that a bird
should take up a crawfish in one stream, carrying it to another safe and sound, and
drop it there without hurting it. Birds do take crawfishes 56 and sometimes carry
them short distances, but this always results in serious injury, even if the specimen
is not immediately eaten. Thus, even though we may admit that crawfishes might
be transported by birds without being injured, such cases must necessarily be
extremely rare, and do not happen often enough to effect the establishment of a
species in a drainage system from which it was originally absent.

There are other considerations which make the assumption of passive transfer
improbable in our case. Toward the east G. obscums is (with exceptions to be dis-
cussed below) rigidly restricted to the Ohio drainage, and nowhere crosses into that

51 At Linesville, Crawford County, I found this species in the very headwaters, almost in the springs running into
Shenango River just south of Summit, which is on the divide.

55 Particulars of one of the experiments (I have made a series) are as follows : November 9, 1905. Cloudy day.
Mean temperature : 34° F. Light breeze from West-South-West, and light snow in afternoon. Specimens of C obscurus
suspended on strings on veranda with southern exposure. Beginning of experiment 9 a. m. One specimen taken in at
2 p. m., another taken in at 4 p. m., and put into water. Both alive and vigorous next morning, and were kept alive
till December 18, when they were thrown into alcohol.

In midsummer, on hot days, I often observed that the vitality of C. obscurus becomes very low after they are only a
short time out of water. They may die within an hour, without having been subject to any other injury than that
caused by the removal from the water.

56 Mr. W. E. C. Todd informs me that remnants of crawfish are quite usual in the nest of the kingfisher. I have
seen, in the collection of the Department of Agriculture, Harrisburg, a specimen of O bartotii, taken from the stomach of
a kingfisher.

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 441

of the Susquehanna. If transport were at all probable we should expect to find
that it had taken place here, as well as in the region of Lake Erie.

Further, and this is the most important objection to the transport theory, while
C. obscurus has invaded the Lake drainage, not only in Pennsylvania, but also in
New York (Genessee River), in no case has the opposite taken place namely, that
C. propinquus has invaded the Ohio drainage. If the crossing of the divide were
due to passive transport, the same cause should have acted in both directions ; but

C. propinquus is entirely absent from the Ohio system.

The latter objection holds good also with reference to another assumption, that
G. obscurus may have crossed into the lake drainage by the aid of the old canal
which connected the Beaver River with Lake Erie (Erie extension of Beaver canal).
This canal (see Jenkins, 1903, p. 288, 289) was in part used as early as 1834, and
was completed in 1844 ; it was abandoned in 1871, and it cannot be denied that by
it C. obscurus might have been able to reach the Erie drainage. I would not hesi-
tate to accept this as correct if it were not for the fact that C. propinquus haa not gone
in the opposite direction. 6("'° Precisely in the region of this old canal my collections
are very complete, and are supplemented by those of others (Messrs. O. E. Jennings,

D. C. Hughes, and W. R. McConnell), so that I am positive about the absence of
C. propinquus.

On the other hand, we have seen that the specimens of G. obscurus from the
tributaries of the lake seem to approach more closely those of Beaver River than
those of French Creek. This would be in favor of the canal-theory, the canal run-
ning from Newcastle by the wa}T of Shenango River to Conneaut Creek (Jenkins,
/. c), while French Creek was not so closely connected with it (although there was
a "French Creek feeder"). The absence of G. propinquus in the Beaver drainage
may be due to the fact that in Erie County, the canal was not so closely connected
with the streams running to the lake, and that thus the lake species could not get
into the canal ; or else G. propinquus being the weaker species of the two could not
make any headway against the more vigorous G. obscurus.

There remains another theory, namely, that the migration of C. obscurus into
Conneaut and Elk Creeks is due to stream-piracy. The latter has undoubtedly
taken place in this region in Postglacial times. The Postglacial divide between
Lake Erie and the Ohio was formed originally by moraines of the late Wisconsin
stage (Lake escarpment morainic system. See Leverett, 1902, PI. 18; also Carll,
1880, PI. 1) or by higher elevated parts of the non-morainic drift lying immediately

5S" It should, however, be borne in mind that the discharge of the water from the canal was downward toward the
lake and thus that migration might in that direction have been easier than in the opposite. — Editor.

442 MEMOIRS OF THE CARNEGIE MUSEUM

in front of this morainic system. The fall of the creeks running northward to Lake
Erie from this divide is much more considerable than that of those running south-
ward, and thus it is clear that erosion on the northern slope must have been more
efficient than on the other side. The consequence is that the tributaries of Lake Erie,
at least some of them, have worked back through the original divide, and have cap-
tured parts of the original Postglacial drainage of the Ohio. This is most evident
(see PL XLIII) in the cases of Conneaut and Elk Creeks, and it is just in these
creeks that I found C. obscurus associated with G. propinqwus,51 while in Walnut
Creek, which has apparently not entirely cut through the original divide, C. ob-
scurus is not found.

Thus it is possible that the presence of G. obscurus in the Lake Erie drainage is
due to stream-piracy. Both species, C. obscurus- and jyropinqwus, are associated here,
but it seems that they are antagonistic to each other to a certain degree. In the
tributaries of Conneaut Creek I found G. propincpms exclusively, while Conneaut
Creek itself contained both, but G. obscurus prevailed, and it appears as if the latter
had driven out the other species, which took refuge in the smaller tributaries.

We might expect to obtain some light upon the question, whether C. obscurus
reached the Lake Erie drainage in consequence of stream-piracy or by the help of
the canal, by the analogy offered in the Genessee drainage, but conditions seem to
have been not entirely identical here. The type locality of G. obscurus (see PL
XL1I, Fig. 3) is the Genessee River at Rochester, Monroe County, New York, where
this species also is found associated with G propinqwus. Mr. W. P. McConnell has
discovered G. obscurus in the upper Genessee drainage near Ulysses, Potter County,
Pennsylvania, The material consists of numerous males of the first and second
form and of females, and there is not the slightest question that this is the true C.
obscurus, no trace of G. prvpiuquus being present here. How did this species get
from the Alleghany drainage into that of the Genessee ?

The drainage of the Genessee River lying entirely within the glaciated area, this
must have happened in Postglacial times. Fairchild (1896, p. 423) has shown that
during the recession of the ice the Genessee basin was occupied by a lake, which had
its outlets in different directions successively, draining either to the Susquehanna or
to the Ohio. He distinguishes ten stages, and the sixth was the last in which the
water flowed to the Susquehanna ; in the seventh and eighth stages Genessee Lake
became connected with Lake Warren, which drained to the west into the Missis-
sippi basin (but not into the upper Ohio), and finally the St, Lawrence drainage was

SI The sources of Elk Creek are in a tamarack swamp, which also drains to the south, to French Creek, so that some
kind of a direct connection may be present. I have not visited this swamp.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 443

established. Thus we see that in the beginning Genessee Lake was connected
repeatedly with the upper Ohio (Alleghany River) drainage, but it is not probable
that C. obscurus immigrated at this time, for then it ought also to have reached the
Susquehanna drainage, since the lake discharged its water into the Susquehanna
(through the "Burns outlet") subsequently to the last connection ("Cuba outlet ")
with the Alleghany River. (See Fairchild, 1896, map, PL 19.)

After this a discharge toward the upper Ohio was never re-established. But we
know that stream-piracy has taken place in this region (headwaters of the Genessee),
and although in some cases the Alleghany River seems to have captured parts of
the Genessee drainage (Oil Creek has captured the head of Black ( 'reek ; see Lev-
erett, 1902, p. 207), the opposite has positively also taken place, for instance, Knight
Creek and Van Campen Creek have captured, according to Fairchild, small lakes
that once discharged towards Oswayo Creek, a tributary of the Alleghany. This may
have happened after the sixth stage of Lake Genessee, when there was no longer
any connection with the Susquehanna s}Tstem, and would explain the presence of ('.
obscurus in the Genessee River and its absence in the Susquehanna.

The eastern boundary of ('. obscwrus in Pennsylvania is formed, generally speak-
ing, by the divide between the Ohio drainage in the west and that of the Susque-
hanna and the Potomac in the east. This is most evident in the northern part of
this line, in Potter, Mclvean, Elk, Clearfield, Jefferson, and Indiana* Vanities. This
species goes up the Alleghany River probably into Potter County, for it has been
found not far away from the county line at Larabee, McKean County. It has not
been found in the drainage of Clarion River in Elk and Jefferson Counties, but this
is very likely due to the excessive pollution of this river. There is hardly a water-
course known to me in Pennsylvania which is in a worse condition than < 'larion
River in Elk County. The wood-pulp mills at Johnsonburg, the tanneries at Ridg-
way, the chemical factory at St. Mary's discharge refuse into it, and Toby Creek
adds sulphur water from the mines above Brockwayville (Jefferson County). Simi-
lar conditions prevail in Red Bank and Sandy Lick Creeks in Jefferson County, but
I have been able to ascertain the presence of this species near the head of Sandy
Lick Greek at Dubois, Clearfield ( 'ounty (about 10 miles from the divide).58 In
southern Jefferson County, C. obscurus is not present in Mahoning Creek at Punx-
sutawney (although C. bartoni was there), this creek being slightly polluted by mine-
waters ; but I found it here in a pond connected with the creek. In Indiana
County it is present in all creeks running to the Alleghany and < ionemaugh ( Little

5BI found this species here on June 16, 1905. Only two living specimens were taken, but numerous dead ones were
lying in the creek. Apparently some injurious substance had been quite recently introduced into the water.

444 MEMOIRS OF THE CARNEGIE MUSEUM

Mahoning Creek, Crooked Creek, Two Lick, and Yellow Creeks). Crossing over the
divide in this region into the drainage of the West Branch of the Susquehanna, no
trace of this species is found. I hunted for it in vain in Sinnamahoning Creek in
Cameron County, in the West Branch and its tributaries in Clearfield, Cambria,
and Indiana County (near Cherry Tree), and in Clearfield Creek in Cambria County.

In this whole region (headwaters of the West Branch) stream-piracy has taken-
place on a large scale, the whole basin of this river having been taken away from
the original Alleghany drainage. But G. obscurus has not been taken over. Ac-
cording to Davis (1889, p. 248, see also above, p. 430) this stream-piracy fell largely
into Pretertiary times, and although we are to assume that it continued during sub-
sequent times (p. 430), it must have been rather slow, and insignificant, chiefly so in
Glacial and Postglacial times, which alone are to be considered in the case of C. ob-
scurus. Although this species was present in the Alleghany River drainage, it did
not go up into the headwaters, remaining away from the actual divide for a distance
of about ten to twenty miles. Under these circumstances, as stream-piracy was only
going on at the headwaters, no good opportunity was offered for this species to cross
the divide.

In Cambria County the continental divide bends to the east, and is transferred
to the main chain of the Alleghanies (Alleghany Front) ; but the eastern boundary
of C. obscurus does not follow it. Here it is the Chestnut Ridge which constitutes
the boundary, beginning in southern Indiana County, and continuing through
Westmoreland and Fayette Counties to the southern state-line. Generally C.
obscurus does not pass beyond this ridge into the higher parts of the Alleghany
Plateau, but there are two exceptions. It is found in the Loyalhanna River in the
Ligonier valley, and in Indian Creek, and in this region it is not the Chestnut
Ridge, but the Laurel Hill Ridge which forms the eastern boundary. In the Cone-
maugh River and the Youghiogheny, this species has not been able to pass up-
stream beyond the Chestnut Ridge, since both rivers become very rough above this
point, and this roughness apparently existed also at the end of the Tertiary Period,
when the rivers descended, through the Chestnut Ridge, from the elevated Old Ter-
tiary peneplain to the late Tertiary base-level, at which they were then flowing.59

59 According to Campbell (1903, p. 292) the peneplain of southwestern Pennsylvania, elevation 1200 to 1300 feet,
is identical with the Old Tertiary Harrisburg peneplain ; and according to White (1896, p. 377), the Old Monongahela
(with the Youghiogheny) of Late Tertiary age was about at base-level. Stevensou (1878, p. 259) has called attention
to an old terrace of the Youghiogheny at Connellsville, which apparently corresponds to the late Tertiary base-level,
200 feet above the present level (894 feet) at about 1100 feet above the sea. At Confluence it is 1820 feet high, thus
giving to the river between Confluence and Connellsville a fall of about 700 feet at the end of the Tertiary. At present
the fall of the river is only 432 feet between the points named. Although the identity of the old terraces is not demon-
strated, the difference of elevation is so great that a considerable fall of the Tertiary river is beyond doubt, and thns at
that time a barrier to the upstream dispersal of C. obscurus must have existed here.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 445

As has been repeatedly mentioned, C. obscurus has been discovered in Wills
( 'reek, a stream which belongs to the Potomac drainage, at Hyndman, Bedford
County, Pennsylvania, and Ellerslie, Maryland. This locality is entirely isolated
and about 40 to 50 miles distant from the nearest parts of the main range in West-
moreland and Fayette Counties, separated from the latter by that part of the Alle-
ghany Plateau which is included between Chestnut Ridge, Laurel Hill Ridge, and
the Alleghany Front. In this region, chiefly in the drainage of the upper Youghi-
ogheny and Castleman Rivers, ( '. obscurus is missing, of which fact I am quite posi-
tive, having searched for it in vain at the following localities : the Youghiogheny
River at Ohiopyle, Fayette County, same river and Laurel Hill Run, ( 'onfluence,
Somerset County ; Youghiogheny River, Selbysport, < rarrett County, Maryland ;
Castleman River, Rockwood, Somerset County ; Flaugherty Creek, between Meyers-
dale and Keystone, Somerset County.

Under such conditions stream-piracy is out of the question. For some time I
suspected that C. obscurus might be present in other parts of the upper Potomac
drainage, but this is not the case. I have investigated the Potomac River at Cum-
berland, Maryland, and above Cumberland (Rawlings, Alleghany County, Mary-
land), and further up, where it forms the boundary between Garrett County, Mary-
land, and Mineral and Grant Counties, West Virginia ; but I have not seen a trace
of this or any other river-species. Below Cumberland G. limosus turns up. Thus
the presence of C. obscurus in Wills Creek is very local, and restricted to only a
small part of the creek. I found it at Ellerslie, Maryland, but not below this point,
although I investigated the whole creek from Mt. Savage Junction to the Pennsyl-
vania state-line. At Hyndman it is quite abundant, but only below a point about
half-a-mile south of the railroad station, thus occupying only about eight or nine
miles of the creek.

These facts are rather strange, and, I believe, can only be explained by the
assumption of artificial introduction by human agency. I do not think that it was
necessarily intentional, but it may be due to accidental stocking of the creek with
this species, which is not altogether improbable, if we consider that in this region a
good deal of fishing is done, and that fishermen from places between Pittsburgh and
Connellsville go over this whole region, and frequently use crawfishes as bait, cap-
turing them in one part of the country, and carrying them for their purposes to
other parts. If C. obscurus has not been transported in this way to Wills Creek,
intentionally or accidentally, I have no other explanation to offer.

The above theory as to the origin of the distribution of the group of C. jiro/iin-
quus explains the facts, as far as I can see. Our knowledge of the distribution of

446 MEMOIRS OF THE CARNEGIE MUSEUM

C. propinquus and of C. propinquus sanborni is rather unsatisfactory, but none of
the known facts is opposed to our theory. With reference to C. obscurus in Penn-
sylvania, I think our assumptions are well supported. Cambarus obscurus is a Pre-
glacial form, belonging to the Old Monongahela (or Spencer) River, which survived dur-
ing Glacial times in the headwaters of this river (Lake Monongahela), and spread out,
in Postglacial times, over the whole of the Upper Ohio drainage (in addition to the
Ohio and Monongahela, over the drainages of the Beaver and the Alleghany Hirers), and
wns only clucked in its dispt rsal in the direction to/card the mountains by the roughness
of the streams. The Ohio River of Postglacial times opened a way down stream,
but C. obscurus was unable to spread in this direction, since these parts were occu-
pied by another closely allied species, C. propinquus sanborni. It has slightly
entered upon the territory of the latter (Fishing Creek), but has not been able to
crowd it out or to conquer it. Similar conditions prevail in the Lake Erie drainage,
which has been reached in consequence of stream-piracy, or else, by the help of
modern canals. Here it came into contact with C. propinquus. In both cases (in
West Virginia and Erie County, Pennsylvania) we see that the other species show
indications of an inclination toward C. obscurus. I believe we have to deal here with
hybridization, but this will be discussed later. Finally the species has crossed over
into the Lake Ontario drainage in the region of the headwaters of Genessee River,
presumably in consequence of stream-piracy. In the upper part of this system, in
Pennsylvania, it did not find any competition, and is alone represented there, while
in the lower part, at Rochester, N. Y., it is again found associated with C. propin-
quus. Further details with respect to these parts are not at hand.

Comparing the distribution of the propinquus-groap with Adams' scheme of
Postglacial dispersal of the biota of North America, we see at a glance that the whole
group belongs to his northeastern biota of the second, wave (Adams, 1905, p. 5cS). The
biotic preserve of this element, during glacial times, was not far from the southern
edge of the ice, in what is now the Ohio drainage, but it was restricted to this
western part, and was not extended east of the Alleghany Mountains. In Post-
glacial times this group advanced northward, forming part of the second wave,
which is most clearly seen in the present distribution of G. propinquus, which largely
entered the coniferous forest-belt in Michigan, New York, and Canada. The other
two forms (C. propinquus sanborni and C. obscurus) did not take much part in the
migration of this wave, since they found a barrier to the north in the shape of the
continental divide, and then, after they had crossed this divide at certain points,
they found competition in C. propinquus, which had populated the whole St.
Lawrence drainage at an earlier date (" biocenotic barrier"). In New York state.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 447

however, G. obscurus has advanced north in the Genessee basin to a considerable
distance.

3. Cambarus bartoni.

a. tin in nuir;/ of Facts. (See pp. 381-38G.)

Cambarus bartoni is very uniformly distributed all over the state, being, how-
ever, rather scarce in the extreme northwest in Erie < kranty, where it is replaced
by the form C. bartoni robustus. We shall discuss this later.

The species extends considerably beyond the limits of this state, chiefly toward
the southwest and northeast. In these directions it ranges from Tennessee and
North Carolina to New Brunswick and Quebec. Westward it reaches central Ken-
tucky and southern Indiana. The Atlantic Coastal Plain is apparently not invaded
by it to any considerable degree.

We clearly see that its range follows the main strike of the Appalachian system,
and knowing that ecologically this species is a form of the rapid and cool waters
of the uplands and mountains, living preferably in small streams and even springs,
we understand that the distribution must be entirely different from what we have
learned with reference to the river-species already discussed.

In Pennsylvania conditions seem to be favorable for this species everywhere,
possibly with the exception of a very narrow strip on the eastern border, along the
Delaware River (coastal plain); but even here it approaches the lowlands very
closely, the Piedmont Plateau reaching the river at many places.

In the mountains elevation is no barrier for it, I found it myself at 2,000 feet on
Laurel Hill Ridge, west of Jennerstown, and at 2,300 feet near Sandpatch, Somerset
County, and at other places at elevations not much less (Chestnut Ridge in West-
moreland County ; near Cresson, < 'ambria County ; Keating Summit, Potter ( ounty).
At Davis, Tucker ('ounty, West Virginia, I collected it in Blackwater River at 3,050
feet, and Faxon, 1898, p. 649, records it from Roan Mountain, North Carolina,

6,000 feet,

h. Origin of the distribution of C. bartoni.

The first point is to ascertain the center of radiation of this species. As I have
pointed out in a previous paper (1905, p. 121), we must regard the southern section
of the Appalachian system as the original home of the subgenus Bartonius, to which
this species belongs, and (/. c, p. 122) the advance and dispersal of the subgenus
took place over the eastern mountains of the United States, the axis of the dispersal
being directed from southwest to northeast.

We have reason to believe that the origin of this species falls into I'reglacial
times, it being rather primitive within the subgenus (at least in comparison with the

448 MEMOIRS OF THE CARNEGIE MUSEUM

diogenes-groxxp). If this is the case it very likely extended in the Tertiary at least
as far north as at present, but the advancing ice of the Glacial Period must have
driven it south again, and it must have survived in the mountains of Virginia,
West Virginia, and states further south. Possibly southern Pennsylvania formed
part of its preserve in Glacial times, for the peculiar preference of this species for
cold water admits this assumption. Be this as it may, it is certain that after the
retreat of the ice this species advanced, occupying or reoccupying the whole state
of Pennsylvania, and keeping on in its northward migration, until finally reaching
its present range.

This advance in a northeastern direction clearly agrees with Adams' third high-
way of dispersal (1902, p. 123) along the Appalachian chain, and C. bartoni also
belongs to the northeastern biota of the second, Postglacial wave (1905, p. 58). What
is interesting in this case is that an aquatic creature follows here the main strike of
the mountains, independent of the drainage systems. I have previously called
attention to this fact (19056, p. 129), and have pointed out that this is rather the
rule with the subgenus Bartonius (this has been observed already by Faxon, 1885a,
p. 179). There is not the slightest doubt that this peculiarity is connected with the
ecological laws governing this species. It lives generally in the region of the head-
waters of the streams, where the dendritic conformation of the drainage systems and
their mutual interlocking favors frequent shifting of the divides in consequence of
stream-piracy.

Moreover, C. bartoni is a form which habitually leaves the water. It is found
not only in small streams, but also in springs, often at places where there is a very
scanty supply of water, and this has forced it to often assume burrowing habits.
Like C. obscurus it is able to survive exposure to the open air for a considerable time,
provided the temperature is not too high, and thus it is easy to imagine that it may
cross over divides during rainy or cloudy weather, wandering from spring to spring
in the mountains. On the other hand, we see that C. bartoni is not entirely absent
from larger streams, and if once established in a small part of the drainage of a cer-
tain river it may easily be distributed over the rest of it by simply following the
course of the stream. Thus it is not strange that this species has occupied the whole
of the state, and this uniformity of distribution is chiefly due to the fact that the
whole of Penns}dvania is hilly or mountainous, offering everywhere favorable condi-
tions for this species. The general dispersal is due to two causes : first the ability of
this species to cross watersheds by active migration ; and second to stream-piracy,
which has apparently played a considerable part in its dispersal.

It should be noted, as we have seen above, that the size of this species decreases

ORTMANN: THE CRAWFISHES OF TF1E STATE OF PENNSYLVANIA 449

markedly in the eastern section of the state. This fact is significant, in so far as it
points out that the center of radiation for the state of Pennsylvania is rather on the
western side of the mountains than on the eastern (see Adams, 1902, p. 122, "fourth
criterion for the determination of centt rs of dispersal ").

4. Cambarus barioni robustus.

a. Sum in a r a of Facts. (See pp. 390-391.)

This form is found in Pennsylvania in a rather continuous area in the extreme
northwest, in McKean, Warren, Erie, and ( Irawford ( lounties, both in the Alleghany
River and the Lake Erie drainages. It is often associated with the typical ( '. barioni,
but has been found at different localities in Erie County without the latter: In
addition it is not rare in the northeastern part of Allegheny County in the Alle-
ghany River, and its tributaries, and has also been found in Chartiers Creek, in
southwestern Allegheny ( 'ounty. Here it is always associated with the typical form.

In no other part of the state has this variety been discovered, and it is very im-
portant to note that no trace of it has been found in southwestern, central, southern,
and eastern Pennsylvania. Although G. barioni is abundant in these parts, and
particular attention has been paid to the possible presence of C. bartoni robusti/UJ,
all attempts to find it have foiled, and I feel justified in asserting that it is absent here.

I am not so sure of this with reference to the region between < Irawford and Warren
Counties on the one side, and Allegheny ('ounty on the other. I have searched in
this section for C. bartoni robvMus, for instance near Tionesta, Forest County, at I >il
( 'itv, Venango ( 'ounty, in Mercer and northern and central Butler Counties, and near
Kittanning and Mosgrove, Armstrong ( 'ounty, but did not discover it. However, it
is possible that it is present along the course of the Alleghany River, in the river itself.
and some of its tributaries, in Forest, Venango, and Armstrong Counties. In some of
the places mentioned 1 did not strike streams which looked very favorable, being gen-
erally not large enough. Yet in Erie and ( Irawford Counties I sometimes found this
species in rather small streams. Conditions in Otter Creek, Mercer County, Slip-
pery Rock Creek and Thorn ('reek, Butler County, were apparently identical with
those under which it is generally found in Erie County, but this form was not found.

I>. Origin of the distribution '</ C. bartoni robvistus.

Considering that the true C. bartoni robustus is a northern form, being found
outside of Pennsylvania in northern < >hio, western New York, and Canada (St,
Lawrence Basin), its center of distribution seems to be at the northwestern edge of
the range of C. bartoni, in the St. Lawrence drainage. In Pennsylvania, however,

450 MEMOIRS OF THE CARNEGIE MUSEUM

it has crossed the continental divide, and has invaded the Alleghany River drainage
in McKean, Warren, and Crawford Counties, and possibly has come down the
Alleghany River as far as Allegheny County, spreading into some of its smaller
tributaries.60 This assumption seems plausible if we take into consideration only the
Pennsylvanian material and that from the St. Lawrence Basin. We would have
here a case of distribution which is entirely unique. C. bartoni robustus should
then be regarded as a Postglacial form, which originated in the St. Lawrence
drainage, and in Pennsylvania spread southward, coming from the north.

But there are objections to this view. C. bartoni robustus has been reported also
from Virginia, Maryland, and Kentucky, and this, of course, would not be in favor
of this theory. However, as has been said above (p. 392), I am inclined to believe
that this southern form is not the same as the northern. If this view should be
correct, I should regard C. robustus as a good species, and then the above opinion
would hold good.

But further, the morphological characters of C. bartoni robustus, as compared
with those of the typical bartoni, are distinctly more primitive. The shape of the
rostrum is decidedly more archaic, the original form of the rostrum in the subgenus
Bartonius being rather elongate, and not short and broad as in G. bartoni. The
frequent presence of distinct lateral spines on the carapace is undoubtedly a primi-
tive character ; and the ecological peculiarity of preferring larger streams than are
haunted by the typical form might also be regarded as a remnant of more primitive
conditions. This, of course, would be strange in a Postglacial form, originating within
the glaciated area, and we rather ought to expect a higher differentiation than the
original, typical form.

Until the question of the identit}' of our northern C. bartoni robustus with the
southern form, which bears the same name, is settled, we cannot form a final opinion.
If both forms should be actually identical, we might have to deal with two races of
G bartoni, an older one ( G. bartoni robustus), which possibly constituted a first wave
of migration from southwest to northeast, which was overrun and crowded out by
a later wave, consisting of G. bartoni typicus. Remnants of the older stock have
been able to survive only at a few, scattered localities in the south, while in the

60 The Alleghany River, between Sandy Creek and Verona, has been investigated repeatedly. It is a curious fact
that Dr. D. A. Atkinson collected here a large number of C obscurus on September 17, 1900, but not a single robustus.
I was at the same place on June 1, 1904, together with Dr. Atkinson and Dr. O. T. Cruikshank, but we did not collect
this form (conditions were unfavorable) ; on November 19, 1904, I spent a whole day there, collecting numerous C. ob-
scurus, and a few C. bartoni (typical ), but not a single robustus was seen. When I visited this place again, on September
7, 1905, I seoured within a short time six specimens of C. bartoni robustus, and on September 30, 1905, I found three fine
specimens a little further up the river, at Hulton, although I did not hunt very diligently. Is it possible that the
migration of this form down the river is going on ? Does it gradually become more abundant ?

ORTMANN : THE CRAWFISHES OF THE' STATE OF PENNSYLVANIA 451

northwest a more continuous and solid colony has remained. The scarcity or even
absence of the typical bartoni in Erie County, Pa., which in our state is the chief
domain of C. bartoni robustus, would support this view. C. bartoni has not yet in-
vaded this region to such a degree as to crowd out the other form. Though I must
confess that it does not strike me as very likely that the smaller form should be able
to conquer the larger one.

Further investigations on this question should be made outside of this state.

5. Cawibarus carolinus.

a. Summary of Facts. (See pp. 396-397.)

This species (see PI. XLIII) is found in Pennsylvania in the southern part of
the Alleghany Plateau, between the Chestnut Ridge in the west, and the Alleghany
Front in the east, preferring the high valleys in this region, but not going up to the
highest elevations of the mountains. Thus, although abundant near Meyersdale in
Somerset County, it does not go up the valley of Flaugherty Creek toward Sand-
patch. I have searched for it in vain between Meyersdale and Keystone, and at
Sandpatch. In a northern direction this species ranges in the valley between the
Chestnut Ridge and the Laurel Hill Ridge as far as southern Westmoreland County.
Here the northern boundary is formed by the cross-divide in this valley separating
the headwaters of Indian Creek from the Ligonier Valley. I am quite positive of
this boundary, since I have searched in vain for chimney-builders all over Ligonier
Valley from Idlepark (below Ligonier) to the sources of the Loyalhanna River.
Coming across the divide to Jones' Mills, within a short time I discovered this species.
In the longitudinal valley between the Laurel Hill Ridge and the Alleghany Front,
this species has advanced further north. It has crossed the divide between the < !aa-
tleman River drainage and that of Stony Creek (tributary to the Conemaugh), and
I found it near Listie and Windber, in Somerset County. At the latter place it seems
to attain its northern boundary. At all events I failed to find it near Lovett in Cam-
bria County, in the high valley of Laurel Run, which to all appearance affords con-
genial conditions for its presence being rather swampy in many places. I have also
searched for it unsuccessfully in the region of < 'resson, Cambria County, and at
several places further north.

The rest of the range of this species is entirely to the south of this state, in
Maryland, West Virginia, Virginia, North and South Carolina (disregarding the
isolated report from the Indian Territory, in which I do not put much faith). No
particulars are known about its boundaries, but in this region also it is restricted to
the mountains.

452 MEMOIRS OF THE CARNEGIE MUSEUM

b. Origin of the distribution of C. carolinus.

Generally, conforming to the subgenus Bartonius, the center of radiation of this
species is to be sought in the southern part of the Appalachian system. It has fol-
lowed in its migration the strike of the mountains, keeping to the higher parts of
the latter. Thus it has entered southern Pennsylvania, being restricted here to the
highest portions of the Allegheny Plateau.

The lowest elevation at which I found it is at Ohiopyle, Fayette County, 1,250
feet, and at Dunbar, Fayette County, 1,260 feet. (At the latter place a few strag-
glers— two specimens — were taken as low as 1,070 feet, associated with C. diogenes,
but here they had apparently come down from the top of the mountain, where this
species was abundant at 1,260 feet.) All other localities in Pennsylvania were
higher, generally between 1,500 and 2,000 feet.

The northern boundary of this species in our state is formed by two different,
opposite features in the physical geography. Between Chestnut and Laurel Hill
Ridges it is a cross divide of the longitudinal valley ; between Laurel Hill and the
Alleghany Front the deep erosion of the original longitudinal valley by the head-
waters of the Conemaugh River forms the boundary. We do not know much of the
geological history of this region, but it seems to me that the floors of these high
valleys with their extensive clay deposits form a part of a former base-level, namely,
that of the Old Tertiary peneplain identified with the Harrisburg peneplain by
Campbell (1903, p. 293). In northern Somerset and southern Cambria Counties
this has been eroded by the Conemaugh system, thus removing a good deal of the
clay bottoms, which seem to be an essential condition for this species, and conse-
quently the lack of this feature, or its interruption by the Conemaugh system at the
northern end of Somerset County, has formed here the barrier to the dispersal of C.
carolinus.

To all appearances C. carolinus is a Postglacial immigrant into this state. The
northern boundaries in both of the longitudinal valleys are rather insignificant, and
we should expect that C. carolinus, being a chimney-builder and able to leave the
water for a considerable time, should be able, like'C bartoni, to cross boundaries of
this character. We should even expect that it would be better fitted to do so than
C. bartoni. In fact G. carolinus must have done so repeatedly on its way from the
South, being found in the upper drainages of rivers running in different directions,
for instance, the upper Youghiogheny in Maryland, the upper Potomac in Mary-
land and West Virginia, upper Decker's Creek (tributary of the Monongahela), and
upper Cheat River in West Virginia.61 That it has been checked in Pennsylvania

61 As to stream adjustments and migration of divides in Garrett County, Maryland, See Abbe, 1902, p. 47, 53.

OKTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 453

by such minor barriers as an insignificant divide and a deeply eroded system of
valleys, renders it very probable that these obstructions are only temporary, and
may be overcome in time, and, on the other hand, that the immigration of this
species is rather recent, its northward migration being not yet finished, but only
temporarily stopped.

The fact that this species is restricted to a narrow strip within the mountains
is clearly due to its ecological habits. It prefers a certain altitude and clay bottoms.
The latter are found in Pennsylvania chiefly on the Old Tertiary base-level, and this
is represented to a large degree only within the mountains. East of the Alleghany
Front and west of the Chestnut Ridge only insignificant remnants of this base-level
are found, and thus this species is missing.

We do not know anything about the Preglacial history of this species, and the
facts at hand furnish no evidence with regard to this question. According to the
morphological characters, and compared with C. diogenes and G. monongalensis, we
must assume Preglacial age for it. Its immigration into Pennsylvania probably is
Postglacial, and thus it possibly belongs to Adams' third wave of migration, starting
from the southeastern center (Adams, 1905, p. 62). However, in analogy to C.
monongalensis it may belong to the second wave, and the northeastern biota (see below
under C. monongalensis).

6. Cambarus monongalensis.

a. Summary of Facts. (See pp. 400-401.)

Cambarus monongalensis occupies in Pennsylvania (see PI. XLIII) a continuous
area in the southwestern part of the state. Toward the east, beginning at the south-
ern state-line, the limit of the distribution is formed by the Chestnut Ridge as far as
the point where the Loyalhanna Paver cuts through this ridge in Westmoreland
County. From this point the boundary follows the Loyalhanna to the north, and
continues northwestward along the Kiskiminetas River. From the point where the
Kiskiminetas empties into the Alleghany, the latter river, and further down the
Ohio, form the northern boundary of this species, until the Ohio leaves the state in
Beaver County.

Within this area this species has been found wherever it has been searched for,
namely: in Greene, Washington, and southern Heaver Counties ; in the northwest-
ern section of Fayette County i in the larger part of Westmoreland ( 'ountv. and in
southern Allegheny County. It has also been traced beyond the boundaries of the
state in a western and southern direction : it is found all over the Panhandle of
West Virginia (Hancock, Brooke, Ohio, and Marshall Counties), and has also been

454 MEMOIRS OF THE CARNEGIE MUSEUM

found at Morgantown, Monongalia County. It undoubtedly goes further south in
West Virginia, but no records are at hand from these parts.

The writer was unable to discover this species in the state of Ohio (Harrison,
Carroll, and Stark Counties), and its absence north of the Ohio-Alleghany River is
well established (with one exception to be presently mentioned). Particular pains
have been taken to ascertain the latter fact. While it is very abundant in Alle-
gheny County, south of the Alleghany and Ohio Rivers, the writer has not in a
single instance found it north of them. He has searched in vain at many localities
in northern Beaver, northern Alleghen)r, in Armstrong, and Butler Counties, and
further north. At one single locality, however, on the northern side of the Alle-
ghany River it is present. It was found by Dr. D. A. Atkinson near Squaw Run,
at Aspinwall, Allegheny County (more correctly near Claremont). This seems
to be a very restricted locality. The writer did not visit it himself, but he hunted
all over the region around it from Aspinwall to Squaw Run, and beyond to Mon-
trose, Powers Run, and Harmarville, without discovering additional localities for
the species. Thus it seems that this locality is the only one on the northern side of
the river, and we are able, as we shall see below, to explain its presence there.

This species is generally found at elevations from 900 to 1,200 feet ; and it rarely
descends to 800 feet or less. The lowest altitude at which it was found is 790 to 800
feet at Colliers, Brooke County, West Virginia, and at about the same (estimated)
elevation it occurs in Fern-Hollow and Nine-Mile Run, Pittsburgh.

b. Origin of the distribution of G. monongalensis.

The distribution of this species outside of the state is very incompletely known,
and consequently we cannot form any opinion as to its center of dispersal. Con-
sidering, however, that it is clearly a form cognate to G carolinus, we may safely
assume that it also came from the south, from West Virginia. G. carolinus and G.
monongalensis seem to be two parallel species, closely connected genetically, the
one belonging to the Old Tertiary base-level within the mountains, (elevation 1,200
to 2,000 feet), the other to the hills west of Chestnut Ridge (elevation 900 to 1,200
feet), formed by the Tertiary erosion of this base-level. The areas of both are sepa-
rated by the escarpment of the Chestnut Ridge, and both have probably migrated
on parallel lines.

G. monongalensis must have invaded Pennsylvania and the Panhandle of West
Virginia from the south, being confined to the region between the Chestnut Ridge
and the Ohio River. That in this case a large river forms a barrier to an aquatic
creature is highly interesting, but is easily explained by the ecological habits of the

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 455

species. Living underground near springs, and positively avoiding even the smaller
streams, it is clear that a large river does not offer congenial conditions, and that it
even may hecome dangerous to single individuals when they are accidentally swept
into such a stream, they then being unable to get out and reach more favorable
locations.

The restriction of this species to a comparatively small area in southwestern
Pennsylvania is thus easily explained. The northward expansion was stopped by
the first large river flowing from east to west in this region.

A few additional points need discussion. Coming from the south, this species
migrated largely in the direction of the great tributary of the Ohio, the Mononga-
hela, and this river did not offer a barrier. It is different with the Youghiogheny.
The latter comes through the ( 'hestnut Ridge, and should form a barrier to the east,
preventing it from entering Westmoreland County and eastern Allegheny County,
On the other hand we see that this species has in one instance ei-ossed the Alleghany
River. I do not think that this is due to direct and actual crossing of the rivers,
but to a shifting of their courses, of which we have many evidences. The geological
history of the rivers of this region is as follows. The highest elevations of the
country between Chestnut Ridge and the Ohio River are very uniform, rarely going
beyond 1,200 or 1,300 feet. This seems to represent an old base-level, belonging to
Old Tertiary times, according to Campbell (1003, p. 292 ff.). This was again cut into
by a drainage system belonging to the Old Monongahela or Spencer River, which,
at the end of the Tertiary, was running again at base-level (White, 1806, p. 377), at
an elevation of about 900 feet (in the region of Pittsburgh), having eroded its valley
about 300 feet below the Old Tertiary base-level. This river was rather sluggish
and frequently formed ox-bows. The most important old channels, having regard
to the matters in hand, are in the first place those which are marked by a terrace
about 225 feet above the present river (at Pittsburgh), both along the Youghiogheny
and Monongahela, at McKeesport, Allegheny County, to the east of the present
rivers, which, consequently, have been shifted to the west (Jillson, 1893, p. 12, pi. 1).
East of Pittsburgh we have an old ox-bow of the Monongahela in the " Wilkinsburg
Valley " at about the same level (Jillson, ibid., p. 8 ff.). Here also the river has been
shifted to the west. These instances are sufficient to show that repeatedly and at
various places opportunities were offered to C. monongalensis to passively cross the
Youghiogheny and the Monongahela Rivers on account of the shifting of the latter;
and the same seems to be the case with reference to the Allegheny [liver in the
region of Squaw Run. According to Jillson (1. c, p. KM, there is a ten-ace 250 feel
high belonging to the same general level as those mentioned above, one to two miles

456 MEMOIRS OF THE CARNEGIE MUSEUM

north of the Allegheny in the region of Claremont. At the same place there is
another terrace north of the river, only 150 feet high, and consequently belonging
to a later period, so that here the final shifting of the river to the south took place
later than in the other cases.

These facts, if they at all influenced the distribution of G. monongalensis, and I
believe they did, give us a hint as to the time of the immigration of this species.
The shifting of the rivers must have taken place sometime during the Glacial
Period, for we know that during this time a considerable amount of erosion was
accomplished, the 900 feet level belonging to the beginning of the Glacial time
(Lake Monongahela stage).52

This leads us to the conclusion that G. monongalensis must have been present in
this region during the Glacial Period, at least during a part of it, and shortly before
these channels were changed.63 It is quite possible that this species had its preserve
in Glacial time not far from the edge of the ice in southern Pennsylvania and
northern West Virginia, and that it began to advance as soon as the ice of the Wis-
consin stage began to recede. This would fully explain the fact that this species
was able to cross first the Youghiogheny and Monongahela by the help of the west-
ward shifting of these rivers, thus opening a way into eastern Allegheny and West-
moreland Counties, and that it later crossed the Alleghany River at Claremont,
when its channel was changed to the present more southern position.

The question remains, why G. monongalensis, having crossed the Alleghany,
did not advance further north. It is found at Claremont (near Squaw Run), in
a comparatively restricted locality, which is not altogether favorable, being at a
rather low elevation. It has not been able to reach more favorable locations at
higher levels, the ascent being more or less difficult on account of the very precipi-
tous hillsides, and moreover it may not prosper here because of the presence of the
competing species, G. diogenes, which is quite abundant in this region. G monon-
galensis is here, so to speak, cornered, and surrounded by unfavorable physical,
ecological, and biocenotic conditions.

If this species existed in this region during Glacial times its Preglacial origin

62 The rivers were cut down even deeper than they are at present, but the valleys were rilled up again, 100 feet or
more (see Jillson, 1. c. ). According to Foshay (1890, p. 402), the chief erosion falls into the end of the Tertiary ; but
the presence of glacial material in the old river channels, 900 feet high (East Liberty, Pittsburgh), places the deepen-
ing of the valleys at a later period. Possibly it was connected with and subsequent to the draining off of Lake Monon-
gahela (Wnite, 1896, p. 3751. which happened sometime during the Glacial Period. Hice (1903, p. -29) places this
cutting down below the present channels between the Kansan and the Wisconsin stages.

63The same conclusion is reached when we consider that this species cannot have immigrated before the present
Ohio was formed, that is to say, shortly before the Wisconsin stage (see last footnote). Otherwise the Ohio would not
form its western boundary.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 457

becomes rather probable. This is in keeping with the morphological characters, as
compared with G. diogenes, for the latter, as we shall see below, is very likely also
preglacial.

We have no evidence as to the Preglacial history of 0. monongalensis. It may,
however, be said, that it must have come from the original home of the subgenus
Bartonius in the southern Appalachians. How far north it extended in Preglacial
times we do not know, but the advancing ice cannot have driven it back very far.
This is very probable because it is a form decidedly partial to cold water. With
reference to its Glacial-Postglacial migration it belongs to the northeastern biota and
the scco ml wave of Adams ; but its advance was apparently checked at an early date
by the Ohio-Allegheny River. „

It will be remembered that with reference to G. carolinus another view has been
expressed (p. 453). In the case of that species we do not possess any facts which
enable us to fix its time of immigration into Pennsylvania with the same proba-
bility as in the case of G. monongalensis. The present extension of the range of G.
carolinus in the southern mountains classes it rather with the southeastern biota. On
the other hand, we know nothing about the southern range of G. monongalensis, and
thus it is at present impossible to properly compare these two species. Their close
affinity, however, and the identity of the ecological conditions under which they
are found (aside from the difference in altitude) render it rather probable that the
parallelism observed between them in some respects may reveal itself also in others.

7. Gawbarvs diogenes.
a. Summary of facts. (See pp. 405-407.)

Aside from a narrow strip along the Delaware River, in Delaware, Philadelphia,
and Bucks Counties in eastern Pennsylvania, this species covers a large area in
southwestern Pennsylvania, namely all the region occupied by G. monongalt risis,
and, in addition, a belt of a certain width to the north of it (see PI. XLIII). Here
the eastern boundary is formed, as in the case of G. monongalensia, by the I 'hestnut
Ridge, but it is continued beyond the Loyalhanna River, extending into Indiana
County, and then it follows the divide between the Susquehanna and Allegheny
drainages as far north as the southern extremity of Jefferson County. From this
region the boundary runs in a westerly direction.

In Jefferson County 1 found this species at Punxsutawney, and I have seen
chimneys rather abundantly to the east of this place, when riding on the Buffalo,
Rochester, and Pittsburgh Railroad, about as far as Big Run, Jefierson County. But

458 MEMOIRS OF THE CARNEGIE MUSEUM

this species is not present in the neighborhood of Du Bois and Falls Creek in Clear-
field County, although favorable localities are numerous there. In the valley of
Red Bank Creek I have looked for it in vain near Brookville, Jefferson County.
Farther west the boundary becomes obscure, and is marked by the following locali-
ties : Kittanning in Armstrong County ; Renfrew and Branchton in Butler County;
and Mercer in Mercer County. At all events I found this species at the places
named, but not north of them. Since no apparent physical feature marks the
boundary in these parts, it remains doubtful whether this is the actual northern
limit of distribution ; but we can narrow down the zone in which it must be situ-
ated by naming a few more northern places where I searched for it in vain at
the proper places : Goodville, Indiana County ; 64 Templeton, Armstrong County
(swampy places of the Alleghany river-bottoms) ; Oil City, Venango County !
the region of the Pymatuning Swamp near Linesville and Summit, Crawford
County. It seems, however, that toward the west the boundary has the
tendency to run in a northwesterly direction, and in Ohio this species reaches
the Lake Erie drainage in Lorain County (Oberlin).

Within the region above denned this species is generally found at a slightly lower
altitude than C. monongalensis. It is, however, not preeminently characteristic of
the river-bottoms, as I formerly believed (1905a, p. 400), but is chiefly distributed
at an elevation of about 900 feet (more or less), that is to say, at about the level of
the valley of the Old Monongahela River of Preglacial times. At the foot of the
Chestnut Ridge it goes up to 1,200 feet and more, the highest point being Donohoe,
Westmoreland County, 1,260 feet, but on the other hand it descends to the river-
bottoms, between 600 and 700 feet, the lowest elevation observed being on the Ohio
river-bottoms at New Martinsville, West Virginia, about 600 feet. Thus C. diogenes
is quite abundant at about 900 feet, where C. monongalensis is decidedly rare ; above
this C. diogenes is rare, while C. monongalensis has its chief domain at this level ;
and below 900 feet C. diogenes is also abundant, while C. monongalensis is found
only in exceptional cases.

While the boundaries of this species in Pennsylvania are tolerably well known,
it is quite different with the rest of the range. It appears that the range is divided
into two unequal, discontinuous parts, an eastern and a western. The eastern com-
prises, aside from the small section of Pennsylvania along the Delaware River, the
whole or portions of New Jersey, Delaware, Maryland, the District of Columbia,
Virginia, and North Carolina. Here it seems to be found exclusively in the Coastal

"This is iu the same valley as at Punxsutawney, hut C. diogenes is positively not found here, since a splendid
place was found for it where it ought to have heen discovered if at all present in the neighborhood.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 459

Plain, not even entering the Piedmont Plateau. The western range begins in
southwestern Pennsylvania and northern West Virginia, and we have seen that it
here belongs chiefly to the late Tertiary base-level of the rivers. But in Pennsyl-
vania it has entered the glaciated area (Lawrence and Miner Counties), and thence
has spread westward over the states of Ohio, southern Michigan, Indiana, Illinois.
Wisconsin, southern Minnesota, and Iowa. Its main range is here in the glaciated
region. But it also occupies localities south of the drift, in Indiana, Kentucky,
Illinois, Missouri, Kansas, Arkansas, Mississippi, and Louisiana, extending westward
to Colorado.

b. Origin of the distribution of G. diogenes.

The first point to be ascertained is whether there is actual discontinuity between
the eastern and western range of this species. In western Pennsylvania I have
positively located an eastern boundary for this species. It is formed by the divide
between the Susquehanna and the Alleghany in the north, further south by the
Chestnut Ridge. In the northern parts of West Virginia I am also positive that it
is not found east of the Chestnut Ridge in Preston and Tucker Counties. We have
the report of Faxon (1885a, p. 71) that this species is found at Deer Park, in western
Maryland, but, as we have seen, this is erroneous (p. 400, footnote 27), and the species
is absent in this whole region. I have searched for it in vain in Somerset and
Fayette ( bounties (east of the Chestnut Ridge) in Pennsylvania, in Preston, Tucker,
and Mineral ( "ounties, West Virginia, and in Garrett and Alleghany Counties, Mary-
land. East of the Alleghany Front, in the Alleghany Mountain region, in the
Great Alleghany Valley, and the Piedmont Plateau it is positively absent. It has
never been recorded from anywhere within these physiographical divisions, and I
myself made special search for it in Bedford, Blair, Fulton, and Franklin Counties,
and in the eastern section of Pennsylvania, and further in the Potomac valley at
Cumberland and Hancock, Maryland, and ( 'berry Run, West Virginia. At many
of these places highly favorable localities were discovered, but no chimney-builders
were found. This is the more convincing since I succeeded with ease in demon-
strating the presence of this species on the alluvial flats of the Delaware River in
Pennsylvania.

Although our knowledge of the distribution of < '. diogenes in Virginia and North
Carolina is far from being complete, all known localities are on the < 'oastal Plain,
and thus it appears that there is actually a gap in the distribution formed physio-
grapbically by the Appalachian system and the Piedmont Plateau.

Our knowledge of the distribution in the west is also very defective, and more
particularly we do not know anything about its southern boundary in West Virginia

460 MEMOIRS OF THE CARNEGIE MUSEUM

and Kentucky. Thus it is difficult or impossible to arrive at any conclusion as to
its center of dispersal. But at this point certain morphological observations may
possibly afford some help. We have seen (p. 407) that in western Pennsylvania the
areola is often not entirely obliterated, a condition which is certainly more primitive
than the normal one. Such specimens are quite frequent in southwestern Pennsyl-
vania, while in the other parts of the range they are rather rare or entirely absent.
This fact, according to Adams' (1902, p. 122, 125) eighth criterion for the determi-
nation of centers of dispersal, points clearly to southwestern Pennsylvania. Here
the character of the areola is the least progressive, while in either direction from
this center, to the east and to the west, it is more progressive. This conclusion is
further substantiated by Adams' seventh criterion : " location of least dependence
upon a restricted habitat." We do not know much about the "habitat" of G diog-
enes in the west and south, but it is certain that in western Pennsylvania it is less
restricted than in eastern Pennsylvania. Along the Delaware River I found it ex-
clusively in the black muck of the alluvial flats, while in western Pennsylvania it
has a much wider range ecologically, being found in clay bottoms, on hillsides, near
springs, swamps, and even on sandy or gravelly soil.

Judging from these facts, and also from the general rule which holds good for
the subgenus Bartoniiis, that its center is in the Appalachian region, we may safely
assume that C. diogenes did not have its center on the Atlantic Coastal plain, nor in
the western parts of its range in the Mississippi basin, but that it is somewhere on
the Alleghanian Plateau ; and since southwestern Pennsylvania and northern West
Virginia are the only parts of this plateau occupied by this species, we have to place
its center here.

Here, as we have seen, it dwells chiefly upon the late Tertiary base-level of the
Old Monongahela drainage, and I believe this was its original habitat. We have
no means to decide whether it was already present in this region in late Tertiary
times ; but the simple fact that it does occupy an area, the physiographical features
of which have developed in Tertiary times, is in favor of this assumption. Further
on we shall become acquainted with another reason for this view. In the Ter-
tiary period its range very likely extended further north ; but the Glacial Period
must have restricted it, and its preserve was in the region indicated. In Postgla-
cial times it spread northward again, at least in Pennsylvania. Unlike C. monon-
galmsis, the rivers did not form a barrier, for this species largely descended into the
valleys, going down to the river-bottoms and the very banks of the river,'"'5 and thus

65It is found frequently on islands in the rivers (Neville and Twelve Mile Islands, near Pittsburgh). I have seen
chimneys on the river banks near Verona, and obtained specimens on the banks of the Kiskiminetas at Kiskiminetas
Junction.

OKTMANN: THE CRAWFISHES OE THE STATE OF PENNSYLVANIA 461

it should have been able to cross the latter. Consequently its range extends
beyond that of C. monongalensis.

It is doubtful what physical feature constitutes the northern boundary of this
species. In Jefferson, Armstrong, and northern Butler Counties, where the boun-
dary is apparently located, the late Tertiary base-level, to which this species
belongs, loses its identity. Possibly it was not developed at all, and this region was
not reduced to base level. So it might lie possible that the roughness of the coun-
try constitutes a barrier here, and this is supported by the fact that the boundary is
located further south in the Alleghany vrdley than to the east and west of it. On
the plateau-like regions in Indiana and Butler Counties, favorable localities are
abundant, while the narrow Alleghany valley, with the deeply cut valleys tributary
to it, do not offer congenial conditions.

It is different further west. In Lawrence and Mercer Counties this species has
invaded the glaciated area, and is found to the north of the terminal moraine
(see Lewis, 1884, p. 183 and 193, PI. 1 1 and 12), and here prefers the swampy depres-
sions formed by kettleholes. But a northern barrier at this point is not evident,
although a tendency to a northward extension seems to be indicated.

Having thus invaded the area of the drift, it is not astonishing that tbis species
spread over large tracts of the latter in Ohio, Indiana, and Illinois. Its presence to
the south of the drift in the Mississippi Valley would then be a continuation of this
westward migration, which finally varied toward the southwest and the south. I
have represented it as such (Ortmann, 19056, p. 1 23, PI. 3) in a previous paper. Nev-
ertheless this question needs further investigation.

There remains the eastern area of this species on the Atlantic < 'oastal Plain. In
the paper just referred to I have expressed the opinion (/. <■., p. 123) that it "de-
scended from the mountains" toward the east, but this apparently needs correction.
Of course the direct way from its supposed center to the Atlantic plain is from south-
western Pennsylvania and northern West Virginia across the Appalachian system
and the Piedmont Plateau to Maryland and Virginia. But the total absence of this
species from this region is against this assumption. There is no possible reason why
it should have disappeared from the Potomac valley, if it had once been present
there, favorable localities being abundant.

Comparing, however, the present eastern range of C. diogenes with that of G.
limosus, we are struck at once by the general similarity. Both species belong to
the Coastal Plain from New -Jersey southward, G. diogenes going a little further
south, and not quite so far north, while < '. limosus has entered the Piedmont
Plateau, and C. diogenes has not. This similarity induces us to assume a similar

462 MEMOIRS OF THE CARNEGIE MUSEUM

origin of distribution, and in that case C. diogenes would also have come from the
north, being driven back by the advancing ice.

This necessitates the further supposition that ('. diogenes is a Preglacial species
(another reason for this has been mentioned above), which extended before the be-
ginning of the Glacial Period further north, probably from western Pennsylvania
into New York or even beyond. This is not improbable, since the Alleghany
Plateau stretched considerably to the north in Preglacial times (see Powell, 1896, p.
80), and although the Preglacial features are largely obscured in this region, it might
have been possible for this species to cross over into the coastal plain from western
New York to southern New England or northern New Jersey, skirting the northern
extremity of the Appalachian system. The coming of the ice must then have re-
sulted in the obliteration of the northern connection of the range, thus dividing the
originally continuous area into a western and an eastern subdivision.

According to Adams' classification (1905, p. 58), G. diogenes belongs to the north-
eastern biota, but its dispersal in Postglacial times does not entirely agree with that
of the second wave. Indeed there is a slight indication of a northward advance in
Pennsylvania, Michigan, Wisconsin, and Minnesota, but the main direction was
westward from the Alleghany Plateau, and even southward. This is undoubtedly
due to the ecological peculiarities of this species (chimney-builder), it having found
no competition in the directions named. In the eastern section of its range a north-
ward advance is hardly noticeable. Here the species is more restricted ecologically
(apparently a higher specialization), favoring only the black mud of alluvial deposits,
and this very likely prevented its northern expansion. However, its exact distri-
bution in New Jersey is unknown.

8. Summary of the Studies on Geographical Distribution.
We have been able in the preceding studies to advance a theory for the dispersal
of each of the Pennsylvanian species of crawfishes. It cannot be denied that in cer-
tain points our ideas do not seem to be well supported, but this is chiefly the casein
instances where our knowledge of the extralimital distribution is defective. It is to
be hoped that similar investigations outside of our state may furnish additional evi-
dence to substantiate our conclusions, or if necessary, to modify them. This much,
however, is evident, that the facts of the distribution of our species are due to two
causes: (1) partly to the existing physiographical features of the country; (2) partly
to past conditions, which have now disappeared. On the accompanying map
(PI. XLIII), we notice the following particulars. In the eastern part of Pennsyl-
vania, along the Delaware River from Trenton, N. J., to Marcus Hook, Delaware

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 463

County, a small part of the Atlantic Coastal Plain enters the state. The Delaware
River runs along the escarpment of the Piedmont Plateau ("fall line," See McGee,
1888, p. 122), but at several places a narrow space is left, chiefly opposite Trenton,
in Bucks County, and below Philadelphia, where alluvial flats are found. These
we may include in the coastal plain, and they are characterized by the presence of
C. diogenes (together with ('. limosus).

The next physiographical divisions of Pennsylvania are the Piedmont Plateau
and the Great Alleghany Valley, reaching from the eastern escarpment of the former
to the Blue Mountain. These divisions form a unit in Pennsylvania. The divid-
ing line between them, South Mountain, being rather insignificant and obliterated,
chiefly toward the northeast.66 This fact is also expressed to a certain degree in
the distribution of the crawfishes. Aside from the generally distributed C. bartoni,
we have here C. limosus, which has invaded this region, coming from the lower
Delaware, Susquehanna, and Potomac. It seems to have spread all over the Pied-
mont Plateau, and also into parts of the Great Alleghany Valley, for instance, into
the Cumberland Valley (between South Mountain and the Blue Mountain, called
here the North Mountain, in Franklin and Cumberland Counties). It has also
been found in the Schuylkill and its tributaries in Berks County, but not as yet in
the Lebanon and Lehigh Valleys (northeastern continuations of the Great Alleghany
Valley). Whether the conditions presented here are original or not seems doubtful.
On the one hand it may be that the canals have served to distribute this species ;
on the other hand, pollution of streams may have restricted it. Be this as it
may, the fact remains that the physiographical divisions distinguished as the Pied-
mont Plateau and the Great Alleghany Valley possess a species of crawfish which
is not found elsewhere, except in the Coastal Plain and the Susquehanna Valley.

Then follows the Alleghany Mountain region, between the Blue Mountain and
the Alleghany Front (see "Willis, 1. a). In the southern and central parts of the
state this section is well marked. In the northern part its western boundary is
obliterated, the Alleghany Front losing its identity. But all the areas which
undoubtedly belong to the mountain region are uniformly characterized by the
presence of only the one species, Cambarus bartoni, with the exception that C.
limosus is found in the middle Susquehanna valley from Harrisburg upward to
Columbia and Center Counties, and in the Juniata valley up to lied ford County ;
and further, G. obscurus exists in Wills Creek, Bedford County. Both these ex-

66See Willis, 1896, p. 172, aud map p. 170-171. South Mountain is the continuation of the Blue Ridge of Vir-
ginia, while the Blue Mountain of Pennsylvania is not identical with the Blue Ridge, but is to the west (northwest)
of it.

464 MEMOIRS OF THE CARNEGIE MUSEUM

ceptional cases, however, do not represent, in my opinion, original conditions. This
is most evident in the case of G. obscurus in Wills Creek, where no other explanation
is possible except that of artificial introduction. With reference to the occurrence
of C. limosus in the middle Susquehanna and the Juniata, within the Alleghany
Mountains, I think in this case also a recent immigration took place, favored by
artificial means (canals). This is, however, somewhat doubtful, since it is impos-
sible at present to ascertain the normal and original conditions prevailing in eastern
Pennsylvania before modern improvements were introduced.66"

The region containing only C. bartoni goes beyond the boundary of the physio-
graphical division of the Alleghany Mountains. In the northwest the Susquehanna
has captured a large part of the drainage of the Alleghany Plateau, and the whole
basin of the West Branch of the Susquehanna is included in the section containing
only G. bartoni. But again a physiographic boundary is formed by the divide
between the Susquehanna and Alleghany river-systems.

All the rest of the state belongs to the Alleghany Plateau. Here much more
varied conditions prevail with regard to the distribution of crawfishes, and this is
chiefly due to the fact that this region was open to Preglacial and Postglacial immi-
gration from the south, southwest, west, and northwest. Aside from C. bartoni,
which is found everywhere, the following five species immigrated into this region :
C. propinquus, G. obscurus, G. carolinus, G. mononyalensis, and G. diogenes. Since
each of these species had a different center of radiation, a different geological history,
different ecological habits, and since, consequently, different barriers existed to the
dispersal, no two species possess the same range.

G propinquus came from the west and northwest in Postglacial times. It belongs
to the Erie basin, and is restricted by the divides of the latter. Thus it is confined
in Pennsylvania to the Lake Erie drainage in Erie and Crawford Counties.

. C. obscurus belongs to the Preglacial Spencer River or Old Monongahela, and to
Lake Monongahela of early Glacial age. The Postglacial dispersal includes first of
all the Ohio-Monongahela-Alleghany drainage, and thus generally the divides of
this basin constitute its boundaries, with a few exceptions. In the region of the
upper Youghiogheny and the Conemaugh the roughness of the streams has pre-
vented it from reaching the plateau to the east of the Chestnut Ridge. In Crawford
and Erie Counties it has crossed over into the Erie drainage (due to stream-piracy
or to artificial means), and in Potter County it has found a way into the Genessee
system (due to stream-piracy).

G. carolinus came from the south, along the high level plains of earl}' Tertiary

b6a The green color on the map, PI. XLIII, should be continued up the Juniata to Bedford County.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 465

age, being partial to extensive clay bottoms. Such conditions are found well
developed only between that part of the Alleghany Plateau which is enclosed
between the Alleghany Front and the Chestnut Ridge. Possibly also elevation has
played a part. Thus it is found only in Somerset and parts of Fayette and West-
moreland counties. The northern boundary is formed by rather insignificant and
possibly temporary barriers.

C. monongalensis is a form parallel to G. carol in as. It also came from the south,
and invaded southwestern Pennsylvania, keeping to locations of less altitude than
G. carolinus. Thus its eastern boundary is formed by the ( 'hestnut Ridge. Its
northward advance was checked by the first large river flowing east and west,
namely, the Loyalhanna-Kiskiminetas-Alleghany-Ohio.

C. diogenes seems to be similar, at least in western Pennsylvania, to C. mononga-
lensis, but it was able to cross the rivers northward. The northern boundary is ob-
scure, and may be not entirely due to topographical conditions. Attention may be
here called to the fact that the isotherms have a curve somewhat similar to that
formed by the northern and eastern boundary of this species.

Thus western Pennsylvania is divided into several sections characterized by their
crawfish-fauna, namely :

1. Area of 0. carolinus (containing C. bartoni and carolinus) : Somerset and
southeastern Fayette Counties.

2. Area of G. diogenes and (J. monongalensis (containing G. obscurus, bartoni,
monongalensis, and diogenes) : Greene, Washington, northwestern Fayette, western
Westmoreland, southern Allegheny, and southern Beaver Counties.

3. Area of C. diogenes without monongalensis (containing G. obscurus, bartoni,
diogenes) : northern Beaver, northern Allegheny, northeastern Westmoreland, west-
ern Indiana, southern Jefferson, southern Armstrong, southern Butler, Lawrence-
and southern Mercer Counties.

4. Area of C. obscurus (containing G. obscurus arid barto%i) : northern Jefferson,
northern Armstrong, northern Butler, northern Mercer, the largest part of ( Yaw-
ford, Venango, Clarion, and Forest, western Elk, northwestern Potter, McKean,
Warren, and southeastern Erie Counties.

5. Area of G. propinquus (containing 0. propi/nqims, obscurus, bartoni): northern
and western Erie and a small part of Crawford Counties.

These are the chief divisions, but there are a few minor differentiations. The
greatest variety prevails in Westmoreland County. Its western part (west of the
Chestnut Ridge) belongs to two of the above areas (2) and (3), divided by the Loyal-
hanna River. But besides the valley between the Chestnut and the Laurel Hill
Ridges presents three different conditions, namely :

46fi MEMOIRS OF THE CARNEGIE MUSEUM

(a) Conemaugh drainage, with only one species : C. bartoni.

(b) Ligonier Valley, with two species : G. bartoni and obscurus.

(c) Headwaters of Indian Creek, with three species : G. obscurus, bartoni, and
carolinus. This latter section also comprises the northeastern corner of Fayette
County.

The greatest number of species found in any one county is five, namely, G.
obscurus, bartoni, carolinus, monongalensis, and diogenes. This is the case in West-
moreland and Fayette Counties. They may be found in close vicinity oidy along
the escarpment of the Chestnut Ridge. For the rest four is the largest number of
species found closely associated, namely the four belonging to the second area, com-
prising the range of G. monongalensis.

In conclusion, attention should be called to the fact that the terminal moraine
in no case constitutes a barrier for any of the Pennsylvanian crawfishes. (Compare
Lewis' map, 1884, with our maps, PL XLII and XLIII.) Of course, for the river-
species the moraine would not be of any consequence, and of the burrowing species,
two, C. carolinus and C. monongalensis, do not reach it at all, while C. diogenes has
crossed it in the west. But instead of being a barrier the glaciated area rather
seems to offer more congenial conditions on account of the frequency of swampy
places (kettleholes).

The question remains, whether our survey of the state is to be regarded as com-
plete and exhaustive, or whether there might be other species within its limits. This
is suggested by Faxon (1885o, p. 165) as to Cambarus blandingi (Harlan). This
species is found in New Jersey at Trenton, on the Delaware meadows just opposite
the eastern extremity of Bucks County, Pennsylvania, associated with G. limosus
(Faxon, 1. c, p. 22 and 88). I have made a careful search for it in this part of the
siate, and visited this corner twice (Sept. 15, 1904, and at the same date, 1905).
Having collected this species previously in New Jersey, I was acquainted with its
ecological habits and knew where to look for it. I indeed found localities that
appeared favorable, but I failed to see any traces of the species. I think it is quite
safe to assert that this species is not found in this state.

The case of Lake Erie is a little different. We have records showing that in
Ohio the western extremity of the lake is inhabited by Cambarus rusticus Girard and
C. imnmnis Hagen. The former has been reported from Kelley's Island (Faxon)
and from Sandusky Bay (Osburn and Williamson), and from several tributaries of
the lake. The latter is mentioned from Lake Erie, off Lorain County (Osburn and
Williamson), and from tributaries of the lake as far east as Lorain County. C.
rusticus is not found anywhere further east, and it seems doubtful whether it finds

OKTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 467

congenial conditions in the lake. The specimens taken may have been accidentally
swept into the lake. Thus I do not believe that it will be discovered in the part of
the lake bordering upon Pennsylvania.

C. immimis, however, besides being found in northern Ohio, reappears in New
York. Faxon (1898, p. 654) has recorded it from a tributary of Oneida Lake, and
recently I have seen specimens, belonging to the New York State Museum, col-
lected by Mr. F. C. Paulmier in Rensselaer Lake, Rensselaer County. Thus its
presence in New York, upon which I cast some doubt ( 1 905/>, p. 134), is to be
regarded as firmly established. However, the connection of these eastern localities
with the western range has not been discovered. If a connection is present at all, it
is to be looked for in the Erie-St. Lawrence basin, and thus would possibly include
the lake shores of Pennsylvania. Yet this connection may not exist, and C. immunis
in New York may be a recent, artificial introduction, which is not altogether impos-
sible, since we know that the crawfishes used for food in the New York market come
in part from the lake regions (Milwaukee, see Ortmann, 1900, p. 1260), and thus
this species may have been introduced. But this question is by no means settled,
and we should try to obtain further facts.

Finally we may observe that the conditions now existing in the case of the
Pennsylvanian crawfishes may not be original, but may have been altered by human
agency. The possible influence of canals upon the dispersion of two specks, C.
Umosus in the east, and C. obscurus in ( Jrawford and Erie Counties, has been dis-
cussed in the foregoing pages, and the transplantation of C. obscurus into Wills
Creek has been stated to have apparently occurred, accidently or intentionally,
through human agency. No other cases of dispersion beyond the natural boundaries
by artificial means are probable. But on the other hand certain species may have
become extinct, at least in parts of their original range, through human agency. Of
this we have many instances, but in our state none has gone so far as to entirely
obscure the original conditions. We have pointed out above that the absence of C.
lirnosus in the Delaware and Schuylkill Rivers in the region of the Great Alleghany
Valley may be due to the pollution of the streams issuing from the anthracite region.
That these rivers, as well as the Susquehanna are considerably polluted partly by
city sewage, partly by mine-water, is sure (see Leigh ton, 1903, p. 112, and 1904,
p. 48), but whether the absence of C. Umosus in this region is due to this fact, or not,
cannot be settled.

It is different in the western part of the state. Here C. obscurus originally occu-
pied all of the Monongahela and Alleghany drainages west of the Chestnut Ridge,
but there are many streams in which it is now lacking, and in which we must assume

468 MEM OIKS OF THE CARNEGIE MUSEUM

its former presence. This is apparently due to the large amount of pollution in
these streams, chiefly by water from mines. The pollution of the Alleghany River
near Pittsburgh, although bad enough from a sanitary standpoint, and due in the
first instance to sewage (Leighton, 1903, p. 122) does not affect the crawfishes, for
they are very abundant here, and the Ohio below Pittsburgh is rich in crawfishes.
But there are many smaller streams contaminated by the waste of coal-mines.67
Such streams are recognized at a glance by the precipitate of reddish and yellowish
sulphate of iron upon their bottoms, and are invariably without life. This is most
evident in the Monongahela drainage of southwestern Pennsylvania (Washington,
Fayette, and southern Allegheny Counties), and also in many smaller streams in
Butler, Westmoreland, Indiana, and Jefferson Counties, where coal-mines are abun-
dant. The worst conditions prevail in certain tributaries of the Monongahela,
in the Monongahela itself, in the Loyalhanna below Latrobe and the Kiskimin-
etas, and in Red Bank and Sandy Lick Creeks. The Clarion River is also with-
out crawfishes in Jefferson and Elk Counties, but this is due chiefly to pollution
by sewage from wood-pulp mills and tanneries (see above, p. 443). In all these
cases it is evident that C. obscurus once existed here, since remnants of it are left at
many places in some of the clearer and not polluted side streams. Since this pol-
lution of the streams by coal-mines is bound to increase, C. obscurus certainly will
disappear from other streams. As we have seen above it was on the point of dying
out in Sandy Lick Creek at Du Bois in 1905 (p. 443, footnote 58). Another case has
been observed in Fern-Hollow Run, Pittsburgh. In the fall of 1903 I found a
small number of specimens of this species left over in some pools once connected
with the run ; a sewer had recently been built here, discharging its polluted water
into the run. In subsequent years this species was not again seen, and has entirely
disappeared, as also from Nine-Mile Run, which receives sewage from Wilkinsburg
and Edgewood.

It should be added that C. bartoni also is frequently influenced by the contamina-
tion of streams, but seems rather more resistant than C. obseu rus. In two cases this was
evident, namely, in Mahoning Creek at Punxsutawney, and in Slippery Rock Creek
at Branchton. In both cases the streams were only slightly polluted by mine-water,
and contained a certain number of specimens of C. bartoni, while C. obscurus was
absent. The latter existed at Punxsutawney in a pond connected with the stream,
and at Branchton in a smaller clear tributary, and consequently must have once
been present in the two creeks.

A stream or river polluted in a certain part becomes relatively clear and pure

67 As to the chemical processes goiDg on in the so-called " sulphur water," see Leighton, 1904, p. 24.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 469

again further down. This is generally seen in the Ohio below Pittsburgh. Although
the Ohio collects all the badly polluted streams of western Pennsylvania, it is itself
not unfavorable to crawfish life. The best instance for this is the ('heat River. At
Cheat Haven, Fayette County, this river is rather clear and teems with life, craw-
fishes and Unionidas being abundant. I was therefore astonished at the condition
of this river at Parsons, Tucker County, West Virginia. The water there has a foul
smell, and is utterly unfit for life, which is due to a wood-pulp mill just above Par-
sons on Shavers Fork. Between Parsons and Cheat Haven (about fifty miles) the
water has improved so far that life is not only possible, but is abundant, only the
blackish color of the water remaining as the last result of the contamination.

V. LIFE HISTORY.

Only a few scattered notes have been published on the life-history of any of the
American species of Cambarus, and some of them are rather doubtful. The most
complete account is that given by Andrews (1904) on the breeding habits of G. limo-
sus, but even this comprises only a small part of the life-history, and moreover, as
may be seen below, is in part rendered unreliable by the fact that the observations
were not made in the field, but in the laboratory.

My own observations have been almost exclusively made in the field, and were
only occasionally supplemented, or rather confirmed, in the laboratory. Since it
was my object from the beginning to watch the behavior of the crawfishes under
natural conditions, laboratory-work could not be depended upon, unless controlled
by field-work, and thus the former was neglected altogether.

With the exception of January and February,68 my work in Pennsylvania extends
over the whole year, thus including all seasons. The results are rather satisfactory,
and I am able to give a complete account of the seasonal life of no less than four
species, and by comparison with these the life-history of the other species of this state
may be inferred. Of course I have not been able to solve all questions. For
instance the question of the frequency of moulting in one and the same individual
remains open, since it can only be settled by observing the same individual contin-
uously; but this is impossible in the field. Nevertheless I have found means to
elucidate this question in other ways, although not with absolute accuracy.

The most numerous and most complete records I possess refer to the common
river-species of western Pennsylvania, C. obscurus, and of this I shall first give an
account.

66 From New Jersey I possess observations made even in January and February and referring to C. blamtingi, C.
limosus, and C. bartoni.

470 MEMOIRS OF THE CARNEGIE MUSEUM

1 . Cambarus obscurus.

Nothing whatever was previously known in regard to the life-history of this
species. I have observed it during the larger part of two seasons, the dates of actual
observation covering the time from March 28 to November 19 (in 1904 and 1905).

Beginning in spring (March) it is ascertained that the species is at this time quite
active, being found in the usual localities (under stones in rivers and streams), and
the specimens are of various sizes and conditions, but all agree in having a rather
dirty (mud-incrusted) shell, a sure sign that the shell is old and that no recent
moulting has taken place. There are occasional specimens with a very clean shell,
in which moulting bas occurred quite recently. This teaches us that during the
winter months as a rule moulting does not take place, but that it begins quite early
in spring, although only in the case of a few individuals. Males of the first form
are abundant at this time, while males of the second form are scarce, and it is chiefly
these newly moulted males which are of the second form. It seems, however, that
in exceptional cases rather young males (30 to 40 mm. long) may have gone through
the winter in the second form. The size of the males of the first form varies greatly ;
the smallest found by the writer in spring (May 2, 1905) measure 40 mm. in length,
but specimens between 40 and 50 mm. long are very abundant. All the males
between 30 and 40 mm. long are of the second form, but they are not abundant, as
has been stated. The smallest male found in spring was 31 mm. long. The condi-
tion of the females in early spring corresponds to that of the males, and in this sex
the minimum size is 27 mm. in length.

Very soon an important event takes place in the life of the females. Eggs are
laid. No signs of this were seen on March 28, 1905, and March 31, 1905, although
a large number of individuals were collected at these dates. But on April 6, 1905,
(in Thorn's Creek, Renfrew, Butler County), numerous specimens with eggs were
taken, some in the very act of spawning. I was able to observe in this species the
peculiar attitude assumed by the female, and the "apron," described by Andrews
for C. limosus (1904, p. 180, fig. 5 ; p. 182, fig. 6). The same was seen repeatedly on
subsequent dates in April, so that April is to be considered as the spawning season.
The number of eggs is rather large, one hundred to two hundred and even more,
but young specimens sometimes have considerably less.

From the beginning of April onward females with eggs are found very regularly
until the end of May. My dates are the following : April 6, 1905 ; April 10, 1905 ;
April 19, 1905; April 24, 1905 and 1906; May 1, 1905; May 2, 1905; May 3.
1899 (Williamson and Shafer) ; May 4, 1905; May 8, 1905; May 17, 1906; May
22, 1905 ; May 25, 1905. With one exception (April 1'5, 1905) I found females with

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 471

eggs every time I collected this species during this period. This rather precisely
fixes the time when the females are "in berry." They carry eggs during the
months of April and May, but at no other time of the year, and during this season
all females, with few exceptions, no matter whether they are large or small, have
eggs. The smallest observed with eggs was 40 mm. long. Of course in the begin-
ning of the spawning season larger females may also be without eggs, but later on
females larger than 40 mm. long are only very rarely found without them. The
latter generally are newly moulted (having soft or clean shells), showing clearly
that at the beginning of the spawning season they were very likely below the mini-
mum size for spawning. Of the few females under 40 mm. long none had eggs.

During the spawning season (April and May) a general tendency toward moult-
ing is observed in all specimens which are not females "in berry." Among the
sterile females, as well as among the males, the old, dirty shells disappear; newly
moulted shells become more and more frequent, and soft shells are frequently
observed. This moulting process in most individuals takes place in the first half of
May, but, as we have seen, some individuals begin as early as March, and in others
the process is delayed till the beginning of June. But by this time all specimens
have moulted under normal conditions, with the exception of the fertile females,
which moult after the young are hatched in June.

A remarkable fact in the case of the males is that this spring moult invariably
changes them to the second form.89 In consequence males of the first form become
scarcer and scarcer, till finally at the beginning of June all have disappeared and
only males of the second form are left. Another remarkable fact is that after the
end of the moulting season in spring no very large males are found. While large
males of the first form of over 70 and 80 mm. in length are quite abundant in March,
April, and the first half of May, they become very rare after that time, and the
males of the second form, which are then abundant, only in rare instances exceed
the size of 70 mm. in length, (only two cases on record). During the summer the
males are generally less than 70 mm. in length. Large males reappear after the
summer and fall moults begin, and then they are again of the first form.

The question arises what becomes of the large males (over 70 mm. in length),
which are rather frequent in spring. According to the records, we cannot assume
that they moult into the second form, for we should then find large males of the
second form in summer. Thus it is suggested that these large males die and disap-
pear. Of this I have found positive evidence in two cases. On June 6, 1904, I col-
lected in the Shenango River at Linesville, Crawford County, a large male of the first

"This change was first observed by Faxon (1884n, p. 42) in Cambarm runticus Girard.

472 MEMOIRS OF THE CARNEGIE MUSEUM

form, 81 mm. long, which was lying concealed under a rock in the usual position.
It was absolutely perfect, without blemish, and with an old, very dirty shell, thus
clearly showing that it had gone through the previous winter. This specimen was
barely alive and to all appearances in a dying condition. A similar instance was
noticed on April 24, 1905, in Wheeling Creek, Elm Grove, West Virginia. Here a
large male of the first form (84 mm. long) was found showing no signs of injury, in
fact in very beautiful condition, but barely able to move. It was kept in water,
but was dead the next day, while other specimens collected together with it were
none the worse for their journey to Pittsburgh. Thus it seems that the conclusion
is justified that the largest males of the first form, after having in the autumn at-
tained a certain maximum size, which may be different according to conditions, but
may be said in general to be about 80 mm. in length, go through the winter, but do
not moult again, and die a natural death in spring.70

The latest dates at which I found males of the first form in spring are as fol-
lows : June 6, 1904, Linesville (the case just mentioned), May 30, 1904, at Waynes-
burg, Greene County ; May 25, 1905, Alleghany River, Mosgrove, Armstrong
County. After the beginning of June, all through the month, and through a large
part of July, no males of the first form have been found, and through the remainder
of July they are scarce. (See below.)

The eges carried by the females hatch at the end of May and the beginning of
June. I found young under the abdomen of the mother on May 30, 1904, on June
5, 1906, and on June 6, 1904. The period during which the young stay with the
mother seems to be short. On May 25, 1905, I found eggs, not yet hatched. From
June 15, 1905, onward all through the remainder of the year I never observed a
female with eggs or young. Between these two dates I have only four records, May
30, 1904, June 2, 1905, June 5, 1906 and June 6, 1904, in three of which the pres-
ence of young ones with the mother was shown. Thus the period when young are
found under the abdomen of the mother is very likely the end of May and first half
of June (about three weeks), and the young crawfishes probably do not remain with
the mother much longer than a week.

Throughout June and part of July no males of the first form are present ; all
males are of the second form, but they are not very large, reaching a maximum size
of only between 60 and 70 mm. The females have got rid of their progeny, and
begin to moult. The old females may die like the old males, but I have no evi-
dence on this point, except as drawn from analogy to the males, and the fact that

70 A great mortality of males in spring (after copulation in captivity) has been observed by Andrews (1904, p.
175) in C. limoaus.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 473

very large females are rare in the latter part of June and in July, and do not become
more abundant till the beginning of August. In addition we now have a new genera-
tion of young crawfishes, hatched at the end of May and beginning of June. After
these have left the mother it is difficult to get them. They are too small, and are
often overlooked, and if captured in the net, are able to escape through the meshes.
But I have seen them at this time, although the first recorded captures are as late as
July 10, 1900 (by Dr. 1). A. Atkinson), size 26 mm., and July 24, 1905 (by myself),
size 21 to 23 mm. This young generation is easily distinguished from the rest by
its size, being considerably less than 30 mm. in length. The minimum length of
crawfishes in spring is 31 mm. in the case of males and 27 mm. in that of females.

In the middle of July further changes occur. A new period of moulting begins
for the medium-sized and older individuals, which is chiefly noticeable among the
males, since they now again assume the first form. The earliest date for the new
males of the first form is July 11, 1905 (Tionesta and Spartansburg). Altogether
four individuals were taken, all of which had soft shells, showing the fact that they
had recently moulted. Further dates are July 24, 1905 (Deer Lick), July
25, 1906 (Russelton), July 26, 1904 (Derry), July 27, 1906 (Shousetown), and
then in August and the following months they regularly occurred. At first
these males of the first form are scarce, but they become gradually more fre-
quent, till finally at the end of September and in October almost all males have
assumed the first form. This also holds good for the new generation born at the
end of May and the beginning of June. These young ones are about 20 to 23 mm.
long in July; in August I have specimens from 24 to 39 mm. in length; in Sep-
tember from 26 to about 50 mm. in length. About this time this generation be-
comes obscured ; for it seems that the rate of growth of the young crawfishes is very
different in different individuals, some gaining during June, August, and September,
only about 15 mm. in length, others more than twice that length. They are about
10 or 11 mm. long when they hatch. The same fact was observed by Andrews
(1904, p. 202) in C. limosub, with even greater differences in size (the length of the
young of the same generation in October being between 22 and 60 mm.).

It is ascertained from the above observations that young specimens, born in
early summer, already at the end of the first summer (September and October)
reach a size sufficient to prepare them for propagation, and the males of this genera-
tion as a rule show this by changing into the first form. The smallest male of the
first form, collected by myself in fall (October 6, 1905), is a specimen from Kittan-
ning, 38 mm. long, but specimens from 40 to 50 mm. long, and undoubtedly belong-
ing to this generation, are quite abundant at this time. Thus we see that by October

474 MEMOIRS OF THE CARNEGIE MUSEUM

the same conditions are established which were found in eaidy spring. Males of the
first form prevail, and those of the second form are scarce, and generally of a small
size, between 30 and 40 mm. long. Specimens of less than 30 mm. in length are
very rare and represented by individuals of the last generation, which have not
been able for some reason to keep pace in growth with their brothers and sisters.
The males are sexually mature, and appai-ently the females likewise, as we shall
presently see.

Copulation actually takes place now. 1 have quite often observed it in the field,
and made record of the following dates : September 5, 1906 ; September 28, 1905 ;
October 6, 1904 ; November 19, 1904. In addition couples were found apparently
preparing for copulation, but not in the act, on September 7, 1905, and September
30, 1905. Among the material collected by Mr. W. R. McConnell was a couple taken
in copula on September 5, 1905. In captivity I observed copulation on September 8,
1905, and November 22 and 23, 1904, and I have found that it is very easy to
induce couples to copulate about this time (September, October, November), pro-
vided that one male and one female are put in the same jar. In no other part of
the year, and, what is more important, not even in spring (March, April, May) does
copulation take place, either in nature, or in the laboratory. All my attempts to
induce specimens to copulate in spring have been in vain, and, of course, in June,
July, and part of August, copulation is impossible, the males not being in condition.
Copulation may occur in August, males of the first form being present, but possibly
the females are not in proper condition before September on account of the delay
of the spring moult. The smallest female seen in copulation was 43 mm. long;
and she undoubtedly belonged to the generation born in spring, thus establishing
the fact that the females like the males are generally sexually mature at the end of
the first summer, and that they are able to produce eggs the next spring. (See
above. The smallest female with eggs was 40 mm. long.)

The above observations and conclusions are founded upon the comparison of
large numbers of individuals, but no single specimen has been followed through all
the different stages of seasonable development. But to a certain extent it is possi-
ble to ascertain the changes through which one and the same individual has to go,
and to lay down its life history.

The young Camharus obscurus is hatched at the end of May or the beginning of
June, from eggs laid in April. The young stay with the mother under the abdo-
men of the latter, for a short time (about a week) ; then they shift for themselves
and develop during the summer rather quickly, moulting repeatedly.71 In Septem-

71 Compare Andrew's observations on C. limosus (1904, p. 190, ff.) : he distinguishes seven larval stages (each be
ginning with a moult), from the hatching, end of May, to the middle of July, when a length of 29 mm. was reached.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 475

ber or October they have attained a length of from 40 to 50 mm., and the males
have assumed the first form. Tbe females also are sexually mature, and copulation
takes place from September to November.72 During the winter no changes occur, and
in early spring they are in about the same condition as in the previous fall. In April
the females spawn, and it is remarkable that spawning takes place normally from
four to six months after copulation. The males generally go through the spring
moult in May, the females a little later in June. This brings up the size of this
generation to from 50 to 60 mm. Then the fall moult begins, lasting from August
to October, in which the specimens attain a size of over 60 mm. After the first sum-
mer only two moults, the one in spring and the other in fall, seem to take place.

At about this time, (October of the second year), the specimens are seventeen
months old. They go through a second copuhiing season, and through the follow-
ing winter, and again through the spring and summer with the same changes,
attaining by the two moults their maximum size of over 70 mm. in length. A
third copulating season follows, their age being now two years and five months. After
this they live until the next spring, when the old males die in April and May, and
the old females probably in June. This shows the life of the individual to be about
three years.73

This seems to be the usual life-cycle of this species. But there are exceptions,
which are primarily due to the fact that in the first summer the growth of single
individuals may be quicker or slower. Whether slow growth, inducing late devel-
ment, influences the total length of life cannot be ascertained, but it must lead to
the result that some specimens are not sexually mature at the end of the first sum-
mer, and that thus the first copulation is postponed a whole year ; for copulation
seems to depend directly on the season, and takes place exclusively in fall, but never
in spring. Furthermore it may be that in single cases life is prolonged an additional
year, as for instance in exceptionally large individuals (about U0 mm. long). But
we may safely assume that three years, or at the outside in exceptional eases four
years, is the duration of the life of this crawfish, and that an individual that lives
up to this age without having met with an accident has fulfilled its destiny and dies
a natural death.

A few additional remarks should be made with reference to egg-laying, moulting,
and copulation. The act of laying eggs is hard to observe, and I cannot improve
upon Andrews' observations on <'. limosus quoted above. The process of moulting,

72 Possibly beginning at the end of August and extending to January.

"Andrews (1904, p. 204) was able to trace C. limosm only to the third rammer, when the sole survivor of his
material reached the size of 90 mm.

476 MEMOIRS OF THE CARNEGIE MUSEUM

is also rarely observed. The old skin splits on the back, between the carapace and
the abdomen, and the crawfish pulls itself out gradually, leaving the empty shell
intact. I have not made any observations on the increase of size at the moulting
time, in the case of this species. (But see below under C. diogenes, monongalensis,
and G. bartoni.) None of my specimens kept in captivity went through this process,
and those found in the field immediately after or during the act were too flabby to
be measured. Empty shells have been found now and then, but the individuals
belonging to these had then lost their identity. Possibly, as has been observed in
G. monongalensis, the old shells are eaten up.

The copulation resembles throughout that of C. limosus, as described by Andrews
(1895, p. 867, and 1904, p. 166, Fig. 1, p. 168), and only a few remarks seem neces-
sary. The male with its claws takes hold of the claws of the female at the base of
the hand or the base of the fingers, and in shifting its position it often seizes several
of the other pereiopods of the female. The other legs of the male are lying on the
sides of the carapace of the female. All the legs of the female are lying close
together on each side, directed forward. The male uses one of the fifth pereiopods,
sometimes the right, sometimes the left, to elevate the copulating organs, and this
leg is laid across the sternum, sticking out on the other side behind the fourth perei-
opod. The use of the hooks is the same as in C. limosus.

After copulation the annulus of the female contains a " spermal plug," as
described in C. limosus. Copulation takes place repeatedly between the same
couple, and one male may copulate in succession with several females, and one
female with several males. This has also been observed by Andrews (1895, p. 867)
in C. limosus, who says copulation " may be repeated by either animal with some
other."

2. Cambams jyropinquus and Cambarus propinguus sanborni.

My observations on these two forms are fragmentary, but the dates at hand make
it certain that the seasonal history is identical with that of C. obscurus.

For G. propinqims I have only three observations dated in the summer ; the most
important being June 7, 1904 (Conneautville Station), when I collected a consider-
able number of this species. As in the case of C. obscurus no males of the first form
were found, and all males of the second form as well as females were of medium size,
between 47 and 60 mm. long. Specimens collected at the end of August, 1900, by
Dr. D. A. Atkinson at Presque Isle were in the same condition (no males of first
form), but there was with them a young female 27 mm. long, belonging apparently
to the generation born in June.

The rest of my material was collected on October 4 and 5, 1901, in Erie County.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 477

Here males of the first form were abundant, while those of the second form were
few and small (between 32 and 43 mm. long). The latter consequently all belonged to
the generation of that year. Some of them very likely would have changed into the
first form within a short time. The smallest male of the first form was 39 mm. long.

In addition I have seen specimens from Lake Erie, Lorain County, Ohio (Oberlin
Museum) collected May 1, 1892, which demonstrate the presence of males of the
first form in spring, and I have received from Mr. E. B. Williamson a couple col-
lected September 1, 1901, in Emmet County, Michigan, in the act of copulation.
The date is slightly ahead of my earliest date for this act in C. obmurus, but falls
into the same general season. All these dates perfectly agree with the rules laid
down for C. obscurus.

Of C. propinquus sanbomi I have collected material only in the early spring and
late summer. The specimens observed in spring (April 14 and 28, 1905), in the
Tuscarawas drainage, Ohio, correspond entirely to the spring condition of G. obscurus.
Generally they have a thick coat of dirt, showing that they have gone through the
winter without moulting. Most of the males are of the first form, but a few are of
the second form, and these have new shells. A large number of the females have
eggs. From Dr. Sterki I received a number of newly hatched young, 12-15 mm.
long, collected on June 18, 190(5, at Dennison, Tuscarawas County, Ohio. Among
the specimens collected on August 28 and 29, 1905, in Wetzel and Pleasants
Counties, West Virginia, many males of the first form were present, but also a con-
siderable number of the second form ; besides, there were a number of small speci-
mens of the generation of that year. Among the material of this form from Oberlin,
collected September 28, 1903, the same was true, and thus in this form also the
known facts agree with what has been observed in the case of C. obscurus.

3. Gambarus limosus.

This is the species on which Andrews (1895 and 1904) made his observations.
My own dates, which are supplemented by those collected by Mr. H. Gera and Mr.
W. R. McConnell, are comparatively few, but, as far as they go, show certain
discrepancies with Andrews' results, which need attention and explanation.

The largest number of observations I possess are dated in the month of Septem-
ber, when I collected this species at numerous localities in eastern Pennsylvania,
New Jersey, and eastern West Virginia in the years 1898, 1904, and 1905. At this
time the condition of this species entirely corresponds to that of C. obscurus. Males
of the first form are abundant and of all sizes. (Smallest, 37 mm. long, from Stony
Brook, Princeton, New Jersey, September 21, 1898, and 40 mm. long, from Gren-

478 MEMOIRS OF THE CARNEGIE MUSEUM

oble, Bucks County, Pennsylvania, September 20, 1904.) Males of the second form
are scarce, and most of them are of small size (between 28.5 and 40 mm. long).
Among the females also are many small individuals (as small as 25 mm. long).
These small specimens clearly belong to the youngest generation, born in the spring
of the year of capture.

Copulation was observed by Mr. H. Gera on September 4, 1905. I saw a
repeated copulation of the identical couple on September 10, 1905, and again, in
other specimens sent to me alive by Mr. Gera, on November 4, 1905. In addition
I have seen this species copulating in captivity in the Anatomical Laboratory
of Princeton University in January, 1899, (Ortmann, 1900, p. 1242). Thus the
copulating season is identical with that of C. obscurus, and lasts from September
into the winter, possibly January.

C. limosus goes through the winter,74 and is found in spring in the same condi-
tion as in fall (April, 1899, at Princeton). In May I found females with eggs (May
9, 1905, Potomac River, Cumberland, Maryland), and on May 30, 1898, I collected
females with young under the abdomen (Stony Brook, Princeton, New Jersey).
Thus the spawning season seems to be identical with that of C. obscurus. During a
part of the summer males of the first form seem to be absent. I record that in July,
1904 (specimens collected by H. Gera in Camden County, New Jersey, without
exact date), no males of the first form were found. In a large set preserved in the
Academy of Natural Sciences of Philadelphia, collected by H. W. Fowler in the
Delaware River at Holmesburg, Philadelphia County, on July 4, 1899, all the males
are of the second form, and this although there are specimens in this set over
80 mm. long. Among the collections of W. R. McConnell there is a set of this
species taken on July 10, 1905, at Milesburg, Center County, which contains two males
of the first form, with quite fresh shells. This date corresponds closely to the first
date (July 11), at which males of the first form of C. obscurus were observed.75

According to the above records it seems very likely that the seasonal history
agrees in every particular with that of C. obscurus. The mating season in fall, the
spawning season in spring, and the absence of males of the first form in early
summer (June and part of July) agrees with what we know of G. obscurus. Com-
paring this with the account given by Andrews, we find the following differences.

11 1 collected specimens in January, 1899, in the Delaware-Raritan Canal, near Princeton, New Jersey. All
the males were of the first form. Collecting was done by seining under the ice. The crawfishes were obtained in
water about four or five feet deep.

"Mr. McConnell collected a newly moulted male with a soft shell, of the first form, at Bloomsburg, Columbia
County, on July 19, 1905, and during August he has several dates for these males : August 10 (Reading) ; August 18
(Marion) ; August 21 (Greenpark) ; August 22 (Landisburg).

ORTMANN': THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 479

Andrews (1904, p. 166) places the normal mating season in the months of February,
March, and the beginning of April, and says that there seems to be an autumnal
pairing (October, November) " in place of or in addition to " the spring pairing. I
believe, however, that the autumnal pairing is the normal one, which may be
extended through the winter. Indeed I have observed it in January, but only in
specimens kept in captivity. Since Andrews' observations were made in the labora-
tory, it appears probable that the mating may be continued or repeated under excep-
tional conditions such as are offered in captivity, but that this is not normal. In
C. obscurus I am positive that under natural conditions copulation does not take
place in March and April.

The time of spawning, as observed by Andrews (1904, p. 176) agrees well with
our records (end of March and April), also the time of hatching (I. c, p. 187), late
in May.

As to sexual maturity, Andrews did not gather facts to show that females are
mature and oviposit at the end of the first year (1904, p. 206), although he observed
copulation at the end of the first summer. I observed, on November 4, 1905, copu-
lation taking place in specimens less than 45 mm. long, and found females " in
berry " of the size of 50 and 45 mm. (Cumberland, May 9, 1905). Since the same
fact has been observed in the case of C. obscurus it is certain that males as well as
females are sexually mature at the end of the first summer, and that the sexual
union is effective, the females spawning the following spring. However, in such
small females the number of eggs is generally very small (fifty or less).

Thus it seems that G. limosus agrees perfectly with C. obscurus in its life-history,
and that the only marked difference from Andrews' account concerns the mating
season. This is however apparently due to the fact that Andrews' observations were
made in the laboratory. The explanation for this is very likely to be sought in the
temperature conditions. The water used in tanks in laboratories has generally a
rather uniform temperature throughout the year, while under natural conditions the
temperature of ponds, rivers, and streams varies considerably in summer and winter.
I made a few observations with reference to C. obscurus. In January, under the ice,
the water is near the freezing point, say about 35° F. ; on April 6, when females
were found spawning, the temperature of Thorn Creek, Butler County, was 45° F.
On May 1 the temperature of Grave Creek, Marshall County, West Virginia, was
66° F. and spring moulting was going on. In midsummer I observed a temperature
of 82° F. in Bates Fork, Greene County, on July 21, and a temperature of 78° F. in
the Ohio, at Ambridge in Beaver County on August 24.

This gives a range of from about 35 to 80 during the year, and I have no doubt

480 MEMOIRS OF THE CARNEGIE MUSEUM

that the nice restriction of certain periods in the seasonal history is primarily due to
differences of temperature. It is only natural that an equalizing of the temperature
must tend to efface the seasonal periods.

4. Cambarus diogenes.

According to my observations, which extend over the period from March 22 to
November 17, this species also agrees in the main features of its seasonal develop-
ment with C. obscurus.

I have the following records for females with eggs : April 6, 1905 ; April 19,
1905; May 2, 1901; May 14, 1899, (collected by Atkinson, Graf, and Williamson) ;
May 21, 1906 ; May 22, 1905 ; May 27, 1904. On May 21, 1906, and June 2, 1905,
I found several females with newly hatched young under the abdomen. In no
other part of the year has this been observed, and thus the spawning and hatching
season is well fixed (April, May, and the beginning of June), and is found to be
identical with that of C. obscurus.

The number of eggs is considerably less than in the case of C. obscurus, and
generally falls considerably short of one hundred. The spawning does not take
place outside of the burrows, but inside of them, and this was most evident in a
female collected on April 6, 1905, (Renfrew, Butler County), in which the eggs were
quite fresh, with traces of the " apron " still visible. This female was dug out of its
hole, as were all the rest with eggs or young ones.

After hatching the young remain a short time under the abdomen of the mother.
But soon they leave her, yet remain in the same hole. I have repeatedly found
young specimens in the same hole with their mother, namely, on June 13, 1904 ;
June 15, 1905; July 6, 1905; July 19, 1905. The smallest were about 10 mm.
long. These young specimens generally occupy a separate part of the burrow, and
are often found near to and inside of the mouth of a closed chimney. They remain
in the hole until they attain a length of 20 mm., which happens toward the end of
July. Then they leave the hole of the parent crawfish and begin to build their own
little holes and chimneys. I observed this on July 26, 1904, at Deny, Westmore-
land County, when I discovered a female 20.5 mm. long in a small hole of its own.
At the same date I found a larger one, 30.5 mm. long, which may have belonged to
the same generation. On August 4, 1901, at Francis Mine, near Burgettstown,
Washington County, I found numerous young specimens between 20 and 29.5 mm.
long, all in their own holes. On August 22, 1905, at Squaw Run, Allegheny County,
I discovered two young specimens 22 and 28 mm. long. On August 26, 1905, at
Baden, Beaver County, specimens 31.5 to 42 mm, long were found apparently under

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 481

the same conditions. As late as September 5, 1904, (Smithfield, Fayette County),
and October 6, 1905, (Kittanning, Armstrong County), I found two very small speci-
mens (24 and 20.5 mm. long) in small holes. Never after July 19 have I found
young ones in the hole of the mother, so that it is quite sure that at the end of July
they invariably shift for themselves when they have attained a length of about 20
mm. The largest found in the hole with its mother was 18.5 mm. in length, on
July 19, 1905. Since young specimens found in the same hole, apparently being
brothers and sisters, often have a different length (15 to 18.5 mm. in the case just
mentioned), and since, as said above, specimens of only 20.5 mm. in length are found
as late as October, the rule is established in the case of this species also that the in-
dividuals of the same litter grow up at a different rate.

With regard to the presence of males of the first form, the same conditions seem
to prevail as in the case of G. obscurus. These males are frequent in spring. I
found them at the following dates : March 23 ; April 2, 6, 15, 16, 24, 30 ; May 2, 3,
13, 14, 21, 22, 27, 29 ; June 2, 15. Then follows a gap of over a month to July 20.
Within this period I made observations upon the following dates : June 16, 18, 26,
27 ; July 6, 16. At none of these dates did I discover a male of the first form. It
is true that the material in this species is less abundant, a dozen specimens collected
on one day representing a rich haul ; but it is nevertheless remarkable that during
the period just mentioned, in which particular pains were taken to get males of the
first form, none were secured. But after this they again appeared regularly, namely :
on July 20; August 7, 8, 22, 26; September 5, 7, 15, 19, 21 ; October 6, 9, 11, 18,
24 ; November 5. This makes it evident that in early summer (end of June and
beginning of July) there is a time when no males of the first form are present.

However, males of the second form are found at any time in the year as fre-
quently as those of the first form. In this respect C. diogenes seems to differ from
C. obscurus. This seems to be due to the fact that C. diogenes attains sexual
maturity at a later age than G. obscums and the river species in general. The
smallest male of the first form ever found measures 55 mm. in length (August 22,
1905, Montrose). It is hardly possible that this individual should belong to the
generation born in June of the same year, since the latter are known to be at that
time about 30 or at the utmost 40 mm. long. We may assume that C. diogenes,
like G. obscurus and C. limosus, may attain at the end of the first summer a length
of about 40 or 50 mm., but these individuals do not then assume the first form as
the river species do. The same is true of the females. The smallest seen in copula-
tion (or associated with a male) was 63 mm. long, and the smallest female with
eggs was 62 mm. long.

482 MEMOIRS OF THE CARNEGIE MUSEUM

Young specimens less than 30 mm. long have not been found in spring, with
one exception, when a female of 25.5 mm. in length was found on May 27, 1904, at
Squaw Run. This, however, apparently was an exceptionally delayed individual
of the generation of the previous year. It was found under unusual conditions,
under a stone on the banks of the creek, evidently removed from its original habi-
tation by winter or spring floods. No additional specimens were seen in close
proximity.

I cannot say much in the case of G. diogenes in reference to the regular moult-
ing periods which were observed in G. obscurus to take place in spring and fall. I
have repeatedly found soft shells, and on April 24, 1904 (Nine-Mile Run, Pitts-
burgh), at a place where a large colony of this species was present, I picked up
numerous cast-off claws,76 which would indicate an early spring moult. But these
claws may have accumulated during winter and spring. The rate of increase at a
moult was measured in one instance. A female, 52 mm. long, captured on April
6, 1905, was kept in captivity, and moulted on July 16. After this process it was
54 mm. long. This cannot be regarded as entirely normal, since the specimen was
kept under unfavorable and artificial conditions.

The copulating season of this species also falls in the autumn. I have only
twice observed a male and female in copulating attitude, but in both cases they let
go when disturbed. This was on November 5, 1904 (Nine-mile Run, Pittsburgh),
and on October 24, 1905 (Branch ton, Butler County). The first couple was found
in water inside and near the mouth of a comparatively simple hole. The male was
70 mm., the female 81 mm. long. The second couple was found a little deeper, but
not over a foot, also in water. The male was 61 mm., the female 63 mm. long-
Mr. F. E. Kelly reports a similar observation made by him on November 14, 1904.
Besides on two other occasions I found males and females associated in couples in
the same hole. Three cases were observed on September 5, 1904, at Smithfield,
Fayette County, and two cases on August 26, 1905, at Baden, Beaver County.
Since it is an absolute rule that under ordinary circumstances only one specimen
occupies a hole, these finds are significant, and, inasmuch as in all these cases it
was always a male of the first form which was associated with a female of good size
(over 63 mm. long), it is evident that this association was connected with the mating
process. Whether the male visits the female, or vice versa, I do 'not know. In
every case the pair was easily captured, being lodged not far from the entrance of
the hole. In some of these cases I was struck by the simple character and small
depth of the burrow, and it may be that the couples dig out small, temporary holes

76 After moulting the shell is generally eaten up, with the exception of the big claws.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 483

for the mating time. This, however, needs further investigation, and possibly, if
found to be the case, may, nevertheless, not be the general rule.

In no other part of the year were similar observations made, and this fixes the
mating season for the months of August, September, October, and November,77 agree-
ing with what we have observed in the case of C. obscwrus.

Thus we see that the seasonal cycle in the life of C. diogenes corresponds closely
to that of the river-species. The only difference is in the time when sexual ma-
turity is reached, and it seems that in the case of C. diogenes this does not occur
earlier than at the end of the second summer. Whether this influences the dura-
tion of life is not known. Nevertheless the fact that this species frequently, or even
regularly, reaches a size superior to that of ('. obscwrus, specimens of over On mm. in
length being quite often found, suggests that this crawfish may live more than three
years, possibly four or five.

The resemblance of the life-history of this burrowing form to that of the river
species is due, I believe, in large part to the similarity of conditions of temperature.
As has been stated, C. diogenes lives near stagnant water and swamps, in places
where there is generally not much fresh and cool water, although such places are
not strictly avoided, and where the temperature of the water is subject to consider-
able seasonal changes. In winter and spring the water in the holes is rather cool
(43° Fahr. on March 23, 1905, in Nine-Mile Run), while in midsummer it becomes
when stagnant, almost lukewarm.

The above observations are in part at variance with those made by previous
writers on the same species. Girard (1852, p. 88), near Washington, D. C, found
females with eggs in March and April, which agrees with our dates, making allow-
ance for the difference of climate between Washington and western Pennsylvania.
Girard also noticed the fact that as a rule only one individual was found in each
hole and mentions as an exception that in one burrow a male and a female were
found together. However, he neglects to tell the exact date of this find (his obser-
vations were chiefly made in spring). In one case, he says that a male was seen
walking over the surface of the ground, as he believes, in search of the female.
But in this instance also no date is given.

Tarr (1884, p. 127) never found male and female together (in May, near Wash-
ington), and always only one individual in each burrow, and lie never found speci-
mens outside of the holes.78 He further believes that the burrowing crawfishes re-

77 It possibly extends further into the winter, as in the case of the other species discussed.

71 Although I have myself never seen a specimen of ('. diogenes walking over the ground, this must sometimes oc-
cur, for males and females must come together in the mating season, and the holi s do not communicate underground.
According to Williamson ( 1901, p. 12), C. diogene* and O, mommgnlemis are nocturnal, and that they come out of their
holes at night is shown by the fact that Mr. Rhoads captured some of them in traps set out over night for rodents.

484 MEMOIRS OF THE CARNEGIE MUSEUM

treat to the streams in the winter, and in spring construct holes for the purpose of
rearing their young, and that impregnation takes place after the winter has passed.
These ideas are not supported by any evidence, and are, as we have seen, above, in-
correct. His opinion that the same burrow is not occupied for more than one year
is also not supported by our observations. The time of hatching of the eggs is given
as about the middle of May (p. 128), which agrees with our dates. Faxon (1885a,
p. 74) reports that according to Mr. P. R. Uhler the female during the period of
incubation goes into pools, ditches, etc. This, however, is contrary to the observa-
tions of Girard, Tarr, and myself. All these particulars refer to the eastern form of
G. diogenes, on the coastal plain, and it seems that with regard to the spawning sea-
son and the spawning habits this form agrees with that of western Pennsylvania,
always considering the slight difference in climate which makes this season begin a
little earlier in the Atlantic lowlands.

The observations made on the western form show more marked differences.
Bundy (1877, p. 171) reports the discovery of a female with eggs nearly ready to
hatch, near Mechanicsburg, Henry County, Indiana, on January 1, 1875. Hay
(1896, p. 491) found that the breeding season in Indiana is in early spring, and ob-
served copulation on April 2, 1892. At this time the specimens leave their bur-
rows, and are frequently found in open ditches and streams. The eggs were laid
from April 18 to April 30. He also repeatedly saw females with well grown young
in small streams. According to Harris (1890, p. 267) a female with eggs was found
in Kansas on May 3, 1891,79 apparently in an open ditch, as is shown by the sub-
sequent sentences.

With the exception of Bundy's record these dates show April and May to be the
normal spawning season of the western form also. The observations of Hay and
Harris, that C. diogenes frequents open ditches in spring, and that it copulates in
spring, are, however, entirely at variance with the habits of this species in western
Pennsylvania. This is not the case here, and I have never seen specimens outside
of their holes in spring. My observations began as early as March 23, at a time
when the frost was hardly out of the ground.

That this is also not the general rule in northern Indiana is shown by notes sent
to me by Mr. E. B. Williamson from Bluffton, Wells County, Indiana. Mr. Wil-
liamson writes to me : " As to C. diogenes congregating in ditches and the like in
spring I have no evidence here. The large males of diogenes can be expected in
almost every little stream. They move about on the bottom restlessly, not lying
concealed. Often the current catches them and they roll over and over, but they

"Sic. Note the discrepancy between this date and the date of publication.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 485

always keep moving. Diogenes also wanders across country at this time, in day-
light as well as at night. I have found several crushed in wagon roads, limit r
these circumstances I have yet to find a female." In another letter Mr. Williamson
says pertinently that the specimens taken hy him in spring in open ditches are all
old males, "in which the death instinct had developed."

Thus it seems evident that the western form does not agree with the form found
in Pennsylvania in so far that in early spring the specimens seem to habitu-
ally leave their holes. Whether it is only old males when about to die (analogous
to what we observed in the case of G. obscurus) which wander about, or whether the
females with eggs also are found in open water, and further, whether copulation
normally takes place in spring, are assumptions which remain to be proved. The
observations of Bundy, Hay, and Harris are surely correct, but it remains to be
ascertained whether they represent exceptional cases, or whether they are the rule.
Moreover it is not improbable that in the western form the seasonal cycle is slightly
different, since it lives under somewhat different surroundings. Hay (1896, p. 491)
reports that during the dry months of the summer C. diogenes seems to lie at the
end of the burrow (which contains hardly any water) in a sort of a stupor. I never
observed anything like this in Pennsylvania, the holes of G. diogenes being always
well filled with water at the bottom, and the crawfishes being very lively.

That observations on the habits of this species should always be considered care-
fully with reference to all accompanying facts is evident from the following case :
Dr. D. A. Atkinson found a number of specimens on April 20, 1905, in open pools
near Westview, Allegheny County, Pennsylvania, in a region where this species is
abundant. These pools were in the course of an old, abandoned mill-race, which
dried out late in summer. All these specimens, seven in number, were young,
measuring from 33 to 52 mm. in length, and consequently belonging to the genera-
tion of the previous year. Now, bearing in mind the fact that the late summer and
fall of 1904 and also the winter of 1904-5 were characterized in our region by an
extreme lack of precipitation so that all streams were exceptionally low till the
middle of March, 1905, when a flood (March 20 to 25) restored the normal condi-
tions, it is very likely that this mill-race was dry in the summer and fall of 1904,
when these young specimens began to make their own burrows. They selected this
place as a favorable one, and remained there all through the winter, a few smaller
floods, one on January 13 and another on March 9, not disturbing them, till the big
flood filled the mill-race again for a longer time, Such conditions, however, do not
suit this species, and consequently the specimens came out of their holes, and were
found, at least for a time, in the open pools, till they had selected more convenient

486 MEMOIRS OF THE CARNEGIE MUSEUM

locations in the neighborhood. Thus this case must be regarded as exceptional, not
as a regular or normal episode in the life of the species.

5. Cambarus barton i and I dambarus bartoni robustus.

In all the species discussed so far we have found a regular seasonal period in the
life-history, marked chiefly by a distinct mating-season in fall, a spawning-season in
spring, and a season in early summer when no males of the first form are present.
But it is entirely different in the case of C. bartoni. In this species none of these
seasons is recognizable.

As to the mating period, I have observations on only two dates. On May 27,
1901, I found a couple in copula in Squaw Run. Here I was able to make a close
observation. The act of copulation is similar to that in the case of C. limosus, as
described by Andrews (1901), but the male does not take hold of the anterior walk-
ing feet of the female with its cheke, and its second Ipereiopods are clasped around
the carapace of the latter, lying in the cervical groove, and almost touching each
other on the back of the female. In this case it was the fifth pereiopod of the left
side, which was stretched across the sternum in order to elevate the copulatory
organs. The male of this couple was 67 mm. long, the female 73 mm. long. The
other observation occurred on October 6, 1905, when I found two couples together
at Weskit, near Kittanning. The male of the first couple was 63.5 mm., the female
59 mm. long. In the other couple both male and female were 63 mm. long. Both
couples separated when captured, and thus I cannot give particulars.

These two dates are so far remote from each other that it seems hardly probable
that they belong to one and the same breeding season. It is possible that one of
them is exceptional, but I have no means of deciding this. On the other hand, as
we shall see presently, spawning takes place at such different times of the year
that very likely the mating-season is also irregular.

Females with eggs have been found on the following dates : July 6, 1905 ; July
10, 1905 ; July 20, 1904 ; July 29, 1905 ; August 1, 1905 ; August 9, 1904 ; August 10,
1905. The number of eggs was between seven and one hundred and thirty-three,
the smallest number being found in the smallest individual, 59 mm. long. In ad-
dition I took a number of females with young under the abdomen. The following
records are at hand. At Princeton, New Jersey, in February, 1898. The exact
date is not recorded, but it was toward the end of the month. The length of the
female was 48 mm., the number of the young was ten.80 Further : March 31, 1905 ;

"•This number is unreliable, but represents as many as were secured. In some oases quite a number of tbeyounj.'
dropped oft when the mother was captured.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 487

length of mother 71 mm., ten young; August 18, 1904, length 59 mm., ninety-two
young; September, 20, 1905, two cases, one 53 mm., long, with thirty-nine young,
and another 55 mm. long, with thirty-five young ; November 8, 1905, 73 mm. long,
with one hundred and eleven young; November 22, 1 905, two cases, one 84 mm.
long, with seventy-five young, and another 67 mm. long, with sixty-eight young.

This extends the spawning season over the following months : February, March,
July, August, September to November. Since young were found in February and
November, these must have been in the egg-stage at least a month before they were
captured, and this would add January and October. Thus we have only interrup-
tions in December and from April to June. The gap in December may easily be
filled, and be due only to the incompleteness of our investigations in winter, but
the gap in April, May, and June may be real.

If there is any spawning-season in C. bartoni it would cover nine months of the
year, from July to March. This, however, is entirely different from what we have
seen in the river species, where the spawning season falls exactly in the months
where no spawning has been odserved in C. bartoni. And besides, this gap may be
partly filled in G. bartoni, for I have found very young specimens. (between 10 and
20 mm. long; the newly hatched young are 9 to 11 mm. long) on the following
dates: May 16, 1905 (13 to 14 mm.); May 25, 1905 (11 mm.); June 2, 1905 (about
J7 mm.) ; June 12, 1905 (14 mm.) ; June 17, 1905 (15 mm.) ; August 22, 1905 (10
to 11 mm.).

The conclusion is that very likely C. bartoni has no defined spawning-season,
but may spawn at any time of the year, and that accordingly the mating-season is
also not restricted to a particular part of the year. The latter is further substantiated
by the fact that males of the first form are found practically all the year round. I
have the following dates: March 21, 28; April 19; May 7, 9, 17, 21, 25, 27, 30;
June 2, 3, 6, 12, 13, 16, 23, 24; July 10, 12, 18, 26, 29; August 1, 10, 18, 22, 26;
September 11, 16, 20, 21, 30; October 5, 6, 10, 12, 17, 24, 31; November 8, 22;
December 25. The only two months missing are .January and February, when no
collecting was done. On the other hand males of the second form are also abund-
ant all the year round, and were found, with the exception of January and February,
in every month.

Under these circumstances it is impossible to say anything about the life-cycle
of the single individual, since different generations cannot be traced. But one
thing should be mentioned. The males of this species do not seem to attain sexual
maturity as early as the river-species. The smallest male of the first form ever
found in eastern Pennsylvania is 49 mm. long, and in western Pennsylvania 50

488 MEMOIRS OF THE CARNEGIE MUSEUM

mm. long. The smallest female with eggs or young is from New Jersey (Princeton),
and is 48 mm. long. From the eastern part of our state I have seen none smaller
than 55 mm. long, and in the western part the minimum is 59 mm. in length.
This is considerably above the minimum size of sexually mature specimens of C.
obscurus and agrees better with C. diogenes.

In one case I have been able to observe the increase in size which takes place
upon moulting. On July 11, 1905, I found at Tionesta, Forest County, a female
in the act of shedding, and succeeded in keeping her alive till the new shell was
hard enough to be measured. The old shell was 32 mm. long, and the new one 36
mm. in length. In this case the crawfish withdrew from the old shell through a
crack that appeared on the dorsal side between the carapace and the abdomen.

We have seen above that the regular seasonal cycle observed in the river-species
is probably due to the regular and considerable changes of temperature taking place
in the rivers. G. bartoni lives in small streams, which generally are much cooler in
summer than the larger ones, and this apparently explains the difference in the
seasonal history. The temperature conditions under which 0. bartoni is found, are
more uniform throughout the year, and consequently no regular seasonal periods
in the life are observed.

No previous observations on this species have been published, except William-
son's note (1899, p. 47), that this species was found with 3'oung under the abdomen
on March 28, 1899, at Columbus, Ohio. This lack of information is rather singular,
considering the extreme abundance of this form in the eastern part of the country.

Cambarux bartoni robustus very likely is identical in its life-history with the
typical form. I have made observations at only a limited number of dates, but
they tend to show that there are no marked seasonal periods.

The following dates for the capture of males of the first form are at hand : May
27, 1904; July 11, 1905; August 22, 1905; September 18, 1900 (Atkinson collec-
tion); September 30, 1905; October 4, 1904; October 6, 1904; November 14, 1903
(Mus. Oberlin). The smallest male of the first form measures 63 mm. in length.

Males of the second form were taken in the months of May, June, July, August,
September, and October. They were abundant in every case, considering the number
of specimens secured.

Copulation was never observed. A female with eggs was found on July 11,
1905, at Spartansburg, Crawford County. It was 84 mm. long, and the number of
eggs was 228, more than twice the number of those usually observed in the typical
G. bartoni. Young specimens, less than 20 mm. long, were taken on May 27, 1904
(18 mm.) ; and were numerous in a lot collected by Miss G. Kinzer on August 27,
1905 (9 to 16 mm.).

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 489

For the months of December, January, February, March, and April, no records
are at hand.

6. Cambarus monongalensis and Cambarus carolinus.

The temperature conditions under which these two species are found are similar
to those observed in the case of C. bartoni, and they are even more uniform, con-
sidering the fact that both are exclusively restricted to spring-water, avoiding even
small streams. In 1905 I took the following measurements of the temperature of
the water in the holes of C. monongalensis (May 16 at Morgantown, the rest at
Edgewood Park); March 18 (ground still frozen in places); 39° F. ; May 16, 58° F. ;
July 8, 63° F. ; August 18, 68° F. The range is considerably less than that given
for C. obscurus (35° to 80° F., see above, p. 479). We consequently should expect a
similar irregularity in the seasonal history as in C. bartoni, differing markedly from
the third chimney-builder, C. diogenes. This is indeed the case.

My observations are rather complete with reference to C. monongalensis, covering
the time from March 18 to December 26. During no period within this time were
males of the second form absent or scarce, but males of the first form were also
almost regularly found ; the following are the dates for the latter : March 18 ; April
4, 21, 24 ; May 1, 6, 9, 16, 21, 24 ; June 3, 30 ; July 6, 8, 20, 24 ; August 7, 13, 18 ;
September 10; October 9, 10; November 8.

Copulation was not observed; but in one case, May 6, 1904, (Fern Hollow), a
male of the first form and a female (55 and 72 mm. long respectively) were found
together in the same hole. The smallest male of the first form ever found was 53
mm. long.

Females with eggs were obtained on June 25, 1906 (O. T. Cruikshank) ; June
28, 1905, (three specimens); June 30, 1904, and July 20, 1905. These few obser-
vations would tend to restrict the spawning-season to the months of June and July,
but this conclusion is not admissible, since newly born young are found at various
other parts of the year. I have such (less than 20 mm. long) taken from the hole
of the mother, at the following dates: April 4, 19)5 (13 mm. long) ; April 29, 1905
(18 mm. long); May 1, 1905 (17 to 20 mm. long); May 2, 1905 (16 to 21 nun.
long); May 6, 1904(19 mm. long) ; June 11, 1904(19 to 20 mm. long); August
18, 1905 (14 mm. long) ; September 24, 1898 (19 mm. long and above, collected by
Rhoads and Williamson); October 29, 1904 (19 mm. long). This extends the
spawning-season considerably, but it has the appearance of being interrupted
during the winter.

The smallest female with eggs is 58.5 mm. long. The number of eggs is between
thirty-eight and seventy-nine, which is considerably less than in the river species,
and also on the average slightly less than in C. bartoni and diogenes.

490 MEMOIRS OF THE CARNEGIE MUSEUM

In this species the young of one litter seem likewise to grow at a different rate.
Thirteen young found with the mother in the same hole on May 2, 1U0-3, were
between 16 and 21 mm. long. Ten young found on June 16, 1904, were from 20.5
to 32.5 mm. long. Twenty-two young, found on July 20, were from 22 to 27 mm.
long. Twenty-four specimens dug out with the mother by Rhoads and William-
son81 on September 24, 1898, are from 19 to 29.5 mm. long.

Thus it seems that C. monongalensis agrees well with G. bartoni. No well-marked
spawning-season is present. At any rate the spawning-season extends over a very
large part of the year, and, correspondingly, no well-marked mating-season can be
distinguished. Males of the first form may be found at any time, and also males of
the second form. Sexual maturity is delayed, males turning into the first form
comparatively late, and the females also are not mature before they have reached a
larger size than the river species. In all these respects, except sexual maturity, C.
monongalensis differs from C. diogenes.

There are further differences from C. diogenes in the development of the young.
We have seen that young C. diogenes remain in the hole of the mother for some time
after they have left the abdomen of the latter. When hatched they are about 9 or
10 mm. long, and leave the mother very soon, since free individuals have been found
only 10 mm. long. When they have grown to about 20 mm. in length, they leave
the hole of the mother.

In 0. monongalensis, however, they stay considerably longer in the hole of the
mother. The exact time cannot be ascertained, but we can draw conclusions from
their size. Free young specimens from 13 to 25 mm. in length are always found
with the mother. The smallest specimen which had begun to make a hole of its
own was 26 mm. long (August 7, 1905, Fern Hollow). Another was 29 mm. long
(October 28, 1905, Edgewood Park). Specimens over 30 mm. long generally have
built their own burrows. But there are exceptions. As we have seen, young up to
32.5 mm. long have been found with the mother; and further, on July 24, 1905
(Deer Lick), I took out of one hole fifteen young, measuring from 27 to 33 mm. in
length, and out of another hole three .young measuring 37, 39.5, and 40 mm. in
length. In these cases the mother was also in the hole. Although in the last two
cases conditions seem rather abnormal, it is certain that the young of C. mononga-
lensis remain longer in the hole of the mother than those of C. diogenes. While the
latter begin to shift for themselves when about 20 mm. long, young specimens of
C. monongalensis do not do so before they reach 25 or 30 mm. in length, and may
even postpone this step till they have attained a larger size (maximum 40 mm.).

81 Williamson ( 1901, p. 12) says that there were forty-seven young ones : only twenty-four are now in the collec-
tion of the Carnegie Museum (Cat. No. 74. 25). Possibly this discrepancy is due to a misprint.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 491

I once observed the change of a male from the second form to the first form.
The specimen was 58 mm. long, and was dug out of its hole on August 18, 1905, at
the type locality, Edgewood Park. It was kept in ajar in water, and had moulted
on August 30. I had been away on an excursion on the two preceding days, hut on
August 27 it had not yet shed. When the fact was discovered it had eaten tin-
larger part of its old shell, only the claws remaining, which were also eaten up sub-
sequently, all but the finger-tips, by September 4. The new shell measured 61 mm.
in length. When captured this specimen was of the second form. After moulting
it was of the first form.

On account of the irregular spawning season it is impossible to trace the life-his-
tory of one and the same individual, and consequently we cannot draw conclusions
as to the duration of life.

The few observations on C. carolinus entirely agree with those made on C.
monongalensis.

Males of the second forms were found in May, June, July, August, and Sep-
tember. For males of the first form I have the following dates: May 17, 1905;
June 24, 1904; August 2, 1905; August II, 1904 ; August 12, 1904 ; September 5,
1905; September 7, 1904 ; October 16, 1905. The smallest male of the first form
measures 56 mm. in length.

Two females with eggs were secured on July 12, 1904. One was 80 mm. long,
and had only three eggs ; the other was 77.5 mm. long, and had seven eggs. These
numbers seem strangely small, and apparently are not normal, for on August 1, 1905,
I found a female 69.5 mm. long, with twenty-two young under the abdomen. But
even this number is below the average of C. monongalensis. While these cases seem
to indicate a spawning season in July, the finding of very young ones in the hole of
the mother at other dates considerably extends this period. I have found such on
May 17, 1905 (17 to 21 mm. long) ; June 13, 1905 (14.5 to 21 mm. long) ; August
1, 1905 (18 to 23 mm. long) ; August 2, 1905 (28 mm. long) ; August 9, 1904 (17
to 25 mm. long) ; August 11, 1904 (19 to 29 mm. long).

The largest young remaining with the mother were 29 mm. long, while the
smallest in a hole by itself was 30.5 mm. long. Thus the time of leaving the hole
of the mother is about the same as in C. monongalensis.

The above observations are not at all sufficient to show that G. carolinus agrees
entirely with G. monongalensis, but sinre both species are alike in so many particu-
lars, morphological and ecological, and since the above dales do not show any differ-
ences, we may safely assume that the life-history of both species is similar.

The seasonal history is rather well known in four of the species above discussed.

492 MEMOIRS OF THE CARNEGIE MUSEUM

G obscurus, G bartoni, G monongalensis, and G 'diogenes. These are the species
found in Allegheny County, and they are most complete, since I had the best chance
to study them, three of them being found in the immediate vicinity of my residence
and the fourth (obscurus) within a few miles and within easy reach.

We are able to distinguish two main types of life-history, which I should like to
call for convenience the warm water and the cool water types. G obscurus and
diogenes represent the first, and agree with each other in having well marked mating-
and spawning-seasons, and in early summer a period when no males of the first form
are found. They differ, however, in the fact that in G obscurus sexual maturity is
reached, as a rule, at the end of the first summer, which does not seem to be the
case in G. diogenes. Of the other species, of which no complete series of dates are at
hand, the river-species, C. limosus, C. propinquus, and G. propinquus sanborni, very
likely agree with G. obscurus, for the comparison of the dates does not reveal any
differences.

The cool water type is represented by C. bartoni and C. monongalensis. Both are
characterized by the absence of well marked mating- and spawning-seasons. They
may be expected in any stage of development at any part of the year, even winter
making no exception. C. carolinus probably belongs also to this type, although the
observations are too scanty to positively establish the fact.

One thing in conclusion should be especially emphasized. The life-history and
the habits of different species of the genus Cambarus are by no means similar. On
the contrary they differ considerably, and the differences may be accounted for pri-
marily by the different ecological conditions under which they live. Consequently
it is inadmissible to generalize from facts observed in one species only, and further
it is to be expected, if other species are studied, that additional types of life-history
will be discovered.

VI. ECONOMIC VALUE.

1. Popular knowledge of Crawfishes.
The crawfishes of this state are generally well known to the population. They
are abundant and large enough to attract the attention even of the casual observer.
But it is chiefly the small boy who is interested in them. Three popular names are
employed for them, crab, crayfish, and crawfish. " Crab " obviously is a misnomer,
belonging originally to the marine Brachyura, but it is largely in use all over the
state, and chiefly so in the cities. The word " crayfish " is used the least. In my
experience I heard it mostly in the mouths of such people as had a certain amount
of schooling and had acquired some knowledge of natural history. This word is

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 493

preferred by teachers generally, very likely in consequence of its use in one of the
standard works on these creatures (Huxley, "The Crayfish"). The third word,
" crawfish," is the proper American name. I found it commonly in use in the rural
communities where "crayfish" and "crab" were often entirely unknown. This is
chiefly the case in the southwestern section of the state and in West Virginia. In one
or two cases in Fayette and Somerset Counties I heard a distinction made between
"crab" and "crawfish." The former name was used for the river and brook
forms, C. obsourus and C. bartoni, the other for the chimney-builders. All three
words go back to the same root, Old German krebis, from which is derived on the
one hand the modern German Krebs, and the English crab; on the other hand the
French ccrevisse, the English crayfish, and the American crawfish. The latter form,
being typically American, and being exclusively known to the natives of a large
part of the country (the farmers), I have decided to use it in preference to the other
two forms.

In literature "crawfish" was used by Say (1817), Harlan (1835), Hagen (1870),
Abbott (1873), Hay (1896). "Crayfish" was used by Abbott (1K84 and 1885),
Faxon (1885, 1890, 1898), Hay (1893, 1899), Andrews (1895, 1904), Shufeldt (1896),
Osburn and Williamson (1898), Harris (1900), Williamson (1901, 1905). Thus
"crawfish " has the priority.

Other names have been given incidentally. Rafinesque (1817) calls C. Hmosus
"mud lobster," (I heard this name once in Delaware County). Say (1817) and
Harlan (1835) call G. bartoni "freshwater lobster," and Williamson (1899) uses the
abbreviation "cray."

2. TJie use of crawfishes asjood and bait.

Although well known, crawfishes are not much used as food by the population
of Pennsylvania ; but this is generally the case in the United States. In some of
the larger cities of the United States they are found more or less regularly on the
market (see Ortmann, 1900, p. 1260), C. Hmosns being one of the species which is
principally used for food. I have, however, never heard that this is the case in our
own state, but it may be found in the markets of Philadelphia. B

Nevertheless crawfishes are eaten in this state, but not regularly. I have heard
sometimes from boys that they had tried them, but only in " sport," and only excep-
tionally have 1 met persons who had eaten them repeatedly and were fond of them.
Generally, this source of food is unknown to the masses in this state. Yet a dish of
crawfishes is not to be despised. It is true, our species never attain the size of the

"Rafinesque (1817, p. 42) says of C. Hmoaus at Philadelphia, that it is "good to eat."

494 MEMOIRS OF THE CARNEGIE MUSEUM

highly esteemed European forms, hut I know from my own experience that, as
regards qualit}^ the former are not inferior to the latter. Young specimens (and
chiefly soft shells) may be fried in butter and eaten shell and all, while the abdom-
inal muscles of older ones, when boiled in water, are very good.

Of course, it is hard to create a taste for crawfishes among the masses, but I do
believe that it would be worth while to try. Crawfishes are so abundant in certain
parts of the larger rivers, G. lirnosus in the Delaware, and C. obscurus in the Ohio
drainage, that it is easy to get any amount of them. It also would not be difficult
to raise them, for instance in ponds, and to supply the market regularly and judi-
ciously. And further, I do not see, why the "tails" (abdomen) could not be used
for canning, exactly like the tails of shrimps and prawns.

Beyond this, crawfishes are used only as bait by fishermen. This use is quite
general, and crawfishes form an important part of the fisherman's outfit especially in
western Pennsylvania. They are most valuable in fishing for Black Bass (Mierop*
term), since these fishes seem to be very partial to this bait.

3. Crawfishes as scavengers. Their food. Their enemies.
The indirect economic value of crawfishes is best expressed by saying that they are
scavengers, as decapod crustaceans in general. They dispose effectively and quickly
of any decayed matter, animal or vegetable, coming within their reach. They also
eat living creatures. This was known previously. Abbott (1873, p. 83) calls them
(C. lirnosus and bartoni) "omnivorous," and "scavengers," and says that they eat
water- weeds, and seize young Cyprinoid fishes. Andrews (1904, p. 175) fed C.
lirnosus in the laboratory on raw and cooked meat, raw eggs, pieces of earthworms,
and on Chara and Hydrodictyon. Williamson (1901, p. 12) reports that G. monon-
galensis was caught in traps baited with raisin and oatmeal. I used for my speci-
mens in the laboratory all kinds of meat, and since I am especially fond of smoked
sausage, I let them often partake of it when I had it for lunch. They also eat earth-
worms and green vegetable matter, for instance seedlings of several weeds (Galin-
soga, and Rumex acetosella), grass, and water-weeds ( Vallisneria). In nature they are
often found at carcasses and other animal refuse lying in the water. They eat in-
sects. For instance I have seen G. bartoni taking grasshoppers used as bait while
fishing for trout (Tub Mill Run, Ross Furnace, Westmoreland County). In the
case of the chimney-builders vegetable matter seems to be largely resorted to, not
only fresh plants, but also decaying vegetation being used. In digging them out of
their holes I repeatedly found masses of decaying leaves and the like lodged in some
side branch of the hole in such a position that they could not have fallen in acci-

ORTMANN: THE£CRAWFISHES OF THE STATE OF PENNSYLVANIA 495

dentally, but must have been brought in by the crawfish. In one case (October 9,
1905, Nine-Mile Run), I found in a side-pocket of a hole of a female G. monongaXenr
sis a number of ripe fruits of Crataegus, about a handful, which under no circum-
stances could have fallen into the position where they were found. The hole was
under a large Crataegus bush.

Thus it seems that any vegetable or animal matter, either fresh, or decaying,
serves as food for crawfishes, and although some species may prefer certain classes of
food on account of taste or necessity, they all take readily to any kind, as is seen by
the fact that in captivity they eat everything that is offered to them without dis-
crimination. If nothing is given, they eat one another.

Crawfishes in turn serve as food for many animals, chiefly those which are
aquatic. Among mammals we know that raccoons hunt for them. As has been
mentioned above, birds eat them, and the kingfisher and other equatic birds do so
quite regularby. The report of Audubon, (see Ortman, 1900, p. 1250), that the White
Ibis captures the chimney-builders by throwing fragments of the chimney into the
hole, and watching for the crawfish to come up, does not seem strange to me. At
Ohiopyle I was told that a domesticated turkey kept upon the grounds of the hotel
had the habit of watching the holes of C. earoliutis, and that frequently he captured
this species. I have myself seen this turkey standing motionless before a hole, but
I did not observe the actual capture. I do not entertain the slightest doubt that
this and other birds are able to catch crawfishes in this way, and do not think that
it is necessary to drop dirt into the hole, since the crawfish comes up frequently on
its own account, when it may be seized.

Crawfishes constitute an important part of the diet of certain snakes, more par-
ticularly of the water-snakes, Natrix sipedon and leberis. I have seen the latter dis-
gorge C. obscurus when captured. (See also Atkinson, Ann. Cam. Mus., I. 1901, p.
149, 150.) On two occasions I have found garter snakes, Eutsenia sirtaHs, in holes of
C. monongalensis ; two specimens of this snake in one hole on October 18, 1904,
(Fern Hollow), and one snake in a hole on October 28, 1905, (Edge wood Park).
However, whether the snakes were after the crawfishes, or whether they simply were
using the holes for winter quarters, remains doubtful.

Professor H. A. Surface writes to me that Cryptohranchus allegherriensis and Nec-
turus maculosus are among the chief enemies of the crawfishes, and, indeed, these
two salamanders are generally found at places where crawfishes abound. (Compare
Eydeshymer, American Naturalist, XL, 1906, p. 128.)

They are, however, most valuable as food for the fish-fauna of our waters. As
has been mentioned above, crawfishes are good bait for certain fishes, and it is very

496 MEMOIRS OF THE CARNEGIE MUSEUM

likely that many of our freshwater fishes depend largely upon crawfishes for nutri-
ment. It would be interesting to investigate how far this mutual correlation
between fishes and crawfishes holds good in our state. The presence of a river-
species in our western streams, and its absence in any drainage systems in the cen-
tral parts is very remarkable. Indeed C. bartoni is found in rivers, but only occa-
sionally, and in small numbers. My own observations are not sufficient to give an
approximate idea as to these relations, since I did not pay much attention to the
fish-fauna, and the latter has decidedly deteriorated, at least in quantity, and the
fish have become rather scarce in most of our streams. Possibly the decrease in the
number of fishes has caused an increase in the number of crawfishes.

4. Crawfishes as obnoxious creatures.

For the river-species hardhy a point can be mentioned which would tend to show
that they are obnoxious to human interests, except the fact that they occasion-
ally capture young fishes. It is different with the burrowing species, which often
become troublesome. In regions where chimney-builders are abundant I have
repeatedly heard complaints about the chimneys, and chiefly so in the case of C.
carolinus in Somerset County, Pennsylvania, Garrett County, Maryland, and Preston
County, West Virginia. Here the mud-piles may hamper farming operations by
interfering with the harvesting machines, clogging and ruining them. At Selbys-
port, Maryland, I was told that conditions were so bad that the farmers tried to
exterminate the crawfishes by throwing unslacked lime broadcast over the fields,
which operation was partly successful, the crawfishes coming out of their holes by
hundreds in a dying condition. I was told that this treatment, repeated several
times, had considerably reduced the numbers of the red crawfish in this neighbor-
hood. At no other place did I hear of attempts made to kill these crawfishes,
although farmers were unanimous in denouncing them as a nuisance.

At a few places another complaint was made, namely, that the chimney-builders
were cutting off and eating up sprouting crops. This was affirmed with reference
to G. caroliims at Reedsville, Preston County, West Virginia, where a farmer told
me that this species had cut off the largest part of a crop of buckwheat, so that
practically nothing was harvested. At Parson, Tucker County, West Virginia,
complaints were made that the same species had damaged sprouting corn ; and at
New Martinsville, Wetzel County, West Virginia, I heard that C. diogenes was
charged with eating up all kinds of sprouting crops, corn and beans being especially
named.

I do not doubt that these complaints are justified, and that the burrowing species

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 497

actually eat and damage crops to a considerable degree. As I have observed, in
captivity (J. monongalensis and C. diogenes eat young plants, and they surely do so
when not in captivity, young sprouting corn, buckwheat, etc., being rather succulent
and attractive to them. If sown in a place where crawfishes abound these crops will
surely be attacked.

This being the case, and besides the chimneys being also a nuisance, it might
be desirable to exterminate the crawfishes in a given locality, or at any rate to
reduce their numbers. For this purpose unslacked lime, the means employed by
the farmers at Selbysport, might be used. But I am in no position to vouch for
the efficiency of this remedy, having no personal experience (with the exception of
the one case mentioned above, p. 346). I simply report what was told me.

Another way might be to drain the places where crawfishes are plentiful. But
this hardly will be as efficient a means as desired. Drainage only lowers the level
of the groundwater, and in the case of C. carolinus, which is the chief offender, we
know that it digs down sometimes over three feet to reach the groundwater. In
Rainier Park at Ohiopyle this species used to be very abundant, but the draining
of the park has reduced its numbers. Still it is present there, and the chimneys
are thrown up all over the lawn, where the holes must in places go down at least
three feet before reaching water. Thus, although a decrease in numbers may be
brought about by drainage, a complete extermination by this method must not be
expected.

Another form of damage done by chimney-builders is known. They are reported
to burrow into and to do damage to the dams on ponds, reservoirs, and rivers.
(The levees of the Mississippi. See Ortmann, 1900, p. 1262.) No instances of this
kind are known to me in Pennsylvania. In one case, at the reservoir of McGee
Run, at Deny, Westmoreland County, I saw holes of G. diogenes not only along the
banks, but also in the dam. The specimens were all young, and the holes small,
since this reservoir has existed only for a few yt-ars. But it would not be astonish-
ing if the crawfishes should gradually work deeper into the dam, finally causing
serious damage.

VII. BEARING OF THE ABOVE STUDIES ON THE THEORY OF

EVOLUTION.
Our observations on the Pennsylvania crawfishes, morphological, ecological, and
geographical, serve to illustrate certain phases of the process of evolution, and certain
theories propounded in connection with them. Naturally they do not elucidate
this process in its fullest scope. Thus I shall only pick out a few points upon which
my observations may have some bearing.

498 MEMOIRS OF THE CARNEGIE MUSEUM

1. Tfie Mutation Theory of De Vries.

The latest fashion in evolution theories is the so-called " mutation theory " of
De Vries (De Vries, 1905). It is much discussed at present, and the general trend
of opinion is that, although De Vries' idea of the origin of species may not hold
good in all cases, he certainly has demonstrated at least one way by which species
may be formed. It is generally maintained with emphasis that his experiments
are beyond doubt and that the facts demonstrated by him cannot be denied.

This indeed is the case, and it would be lamentable if any of the statements pre-
sented by De Vries as facts should prove to be unreliable. I am decidedly of the
opinion that the statements are correct, but I also hold that De Vries was not the
first to bring the facts forward. They belong to a class that was known long ago.
But furthermore, I believe that the conclusions drawn by De Vries from these
facts, are entirely wrong.

I recently have devoted several articles to demonstrate this, and shall not again
go into detail here (see Science, May 11, August 17, and November 30, 1906).

However, I shall discuss here a special part of De Vries' theory, which concerns
the distinction he makes between " fluctuating variation " and " mutation." The
latter is said to be characterized by "sudden leaps," while the former is said to be
by " small steps." Although De Vries sometimes does not lay much stress upon
this distinction (see Copeland, 1904, p. 421), this difference is often regarded as
paramount in his theory (see MacDougal, in Popular Science Monthly, vol. 39,
1906, p. 207). And since De Vries believes that species are formed only by muta-
tion, it should be expected that the morphological differences between existing
species should at least frequently exhibit signs of "sudden leaps." If such leaps
are observed in our species of Cambarns, this would tend to support this part of De
Vries' theory ; if not, the theory that mutations are always or generally marked by
discontinuity of variation, should be dropped.

2. Species, Varieties, and Variations among the Pennsylvania Crawfishes.

I have distinguished in the systematic part of this monograph seven species and
one variety among the Pennsylvania crawfishes. Besides I have discussed another
extralimital variety. This means that the characters distinguishing these forms are
different in their taxonomic value, and the reasons for thus estimating them should
be given.

The seven species of Pennsylvania belong to two subgenera, Faxonius and
Bartonius, which are distinguished by very sharp differences in the male copulatory
organs.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 499

The subgenus Faxoniua is represented in our state, by three species : C. limosus,
C. propinquus, and G. obscurus. The first is geographically, as well as morpho-
logically, separated from the other two ; and here again it is the shape of the male
organs which serves as the chief distinguishing feature. Besides there are other
characters, such as the shape and the spinosity of the carapace, which make it pos-
sible to recognize G. limosus at a glance. No transitional forms being present, the
standing of G. limosus as a " good species " is beyond doubt.

It is different with C. propinquus and obscurus, and the extralimital form G.
scmborrti. These three resemble each other very closely, and it is hard, indeed
impossible, without close examination to distinguish them. They also live under
similar ecological conditions, and their ranges together form a unit, so that it is evi-
dent that they are closely allied genetically. The differences of C. obscurus from the
other two forms are furnished by the " shoulder " of the male organ and the tubercles
of the annulus of the female, together with the complete lack of the median keel of
the rostrum. Other differences, such as sculpture and spinosity of the chelipeds,
are of secondary value and not entirely reliable. But it must be emphasized that
within the established range of C. obscurus, from Fish Creek in the southern part of
the Fanhandle of West Virginia to the upper Alleghany and the Genessee Riven
in McKean and Potter counties, and from Cheat River at the West Virginia state-
line, to the upper Shenango River in Crawford County this species is remarkably
uniform in the characters mentioned. No specimens have been found within this
area which show the slightest tendency toward C. propinquus.

Thus, with reference to this form, the postulate that a species should be sharply
and constantly separated from the coexisting allied forms is fulfilled (see Ortmann,
1896, p. 191) and accordingly I regard C. obscurus as a good species.

As regards C. scmborni, matters seem to be slightly different. It agrees in the
shape of the sexual organs with G. propinquus, and differs only from the latter in
the lack of a rostral keel and some minor features in the armature of the chelipeds.
In the lack of a rostral keel it approaches G. obscurus, but always may he distin-
guished by the shape of the sexual organs. Its relation to C. propinquus remains
doubtful. My observations do not cover the region in which possible transitions
might be expected (northern and western Ohio), and thus I must leave this question
open, and I follow Faxon in regarding C. samborni as a variety of G. propinquus.
But it should be possible to settle this question by proper investigation, and I would
not be astonished if it should be finally discovered that G. sanborni actually is a
good species, sharply and constantly separated from C. propinquus.

The subgenus Bartonius contains four species in Pennsylvania. One of them,

500 MEMOIRS OF THE CARNEGIE MUSEUM

G. bartoni, differs from the rest ecologically as well as morphologically. It is dis-
tinguished by a number of characters, and there is no possibility of morphologically
intermediate forms, so that G. bartoni not only is a good species, but also belongs to
a different section of the subgenus.

C. bartoni possesses in Pennsylvania a variety, G. bartoni robustus, which, accord-
ing to my experience, is constant, and never runs into the typical form. It also
seems to occupy a slightly different territory, although often found associated with
the latter. These facts would justify us in regarding it as a good species. I have
not done so in the systematic part, since the facts at hand are too meagre to finally
decide this question. The range of C. robustus in Pennsylvania is only a small part
of the area occupied by this form, and in the states of Ohio, New York, and in
Canada, the conditions are entirely unknown. Furthermore a form similar to our
robustus, although, as it seems to me, not entirely agreeing with it, has been reported
from Virginia, Maryland, and Kentucky, and before particulars about the relation
of this form to G. bartoiii and to our robustus are known, we cannot judge as to the
taxonomic position of C. robustus. Therefore I have refrained from modifying the
position hitherto assumed, that this form is a variety of C. bartoni.

The other species of the subgenus Bartonius in Pennsylvania are G. carolinus, C.
monongalensis, and C. diogenes. They belong to the diogenes-section, and all three
are closely allied. C. carolinus and monongalensis are more nearly related to one
another than to G. diogenes. The latter apparently is a more advanced form.

C. carolinus and C. monongalensis are distinguished by rather insignificant mor-
phological characters, discovered in the shape of the rostrum and the armature of
the chelipeds. But the difference in color is so striking that it is impossible to con-
found them in the field. Other characters also, although slight, hold good accord-
ing to my experience, and I never have seen intermediate specimens. Moreover
the distribution of these two forms is very characteristic, they being sharply sep-
arated topographically, and never being found associated at the same locality. Thus
all requirements leading us to pronounce them good species are met. Of course this
applies only to conditions in Pennsylvania, Maryland, and northern West Virginia ;
whether they are the same or different farther south remains to be seen.

G. diogenes is more sharply separated from the species just discussed, and there
is no possibility of mistaking this species, more particularly as the color is markedly
different. But the morphological characters are also very nicely expressed, so that
in a case of a red (albinistic) specimen of this species I was not a moment in doubt
that I had to deal with C. diogenes, and not with G. carolinus, although the latter
was found associated with this form at this particular locality (Dunbar). There is

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 501

no doubt that (J. diogenes is a good species, and even when discovered in company
with C. monongalensis or C. carolinus we found no intermediate forma which might
render the identification uncertain.

As has been demonstrated above, there are two races of G. diogenes in Pennsyl-
vania, an eastern and a western. They never have been distinguished before, and
indeed are very similar, so that it is hard to tell them apart. But I think I am
able to do so. The differences are very slight, but I never observed intermediate
forms, and their existence is improbable, the ranges of the two races being widely
distant from each other. The constancy of the differential characters being the
only criterion of specific difference, while the amount of difference is of no con-
sequence at all,83 we might regard the eastern form as a different species from
the western. This may prove to be the correct view, and then the eastern form
should be called G. diogenes < rirard, and the western possibly C. obcsus Hagen.
I have not taken this course in the systematic part, since our knowledge of C. dio-
genes is by no means complete. I know only the conditions in this state, but the
eastern range of this species extends over large parts of the coastal plain, while the
western occupies a vast territory reaching to the Rocky Mountains and the Gulf.
It is also not impossible (although improbable) that the eastern and western areas
are connected somewhere, (in Virginia?). Before this question is finally settled, and
before we know more about the conditions under which C. diogenes occurs in the
extralimital parts, it is best to refrain from expressing a positive opinion. Never-
theless it is quite possible that there is a tendency in G. diogenes to split into varie-
ties and species. A variety has been distinguished in Louisiana.

We see that in certain forms my studies have led to a positive decision as to their
taxonomic position. In other cases my observations must be completed and sup-
plemented by additional evidence to be gathered in other parts of this country before
a final opinion can be reached. The fault is not with the material at hand, but
with the insufficiency of our knowledge of the extralimital parts.

As to variations, that is to say, occasional aberrations from the typical form, we
have seen that such are extremely rare among the Pennsylvania crawfishes, and
have in most cases the character of freaks. Some of them, however, arc interesting
from certain points of view.

No variations were discovered among one hundred and nineteen individuals of
C. limosus. With reference to G. propmqwus in Erie and Crawford Counties, 1 nave
pointed out that there is a certain amount of variation in the development of the

83 De Vries (1905, p. 127) talks of "an old rule in systematic botany, that no form is to be constituted a species
upon the basis of a single character. " This rule is entirely unknown to me in botany as well as in zoology.

502 MEMOIRS OF THE CARNEGIE MUSEUM

keel of the rostrum, and in that of the spines and the carpopodite and meropodite
of the chelipeds, and we have also seen that there is sometimes a notch on the
anterior margin of the male organ. All these characters mark a certain inclination
toward G. obscurus which will be discussed below.

The six hundred and eighty-seven specimens of G. obscurus at hand are, as has
been seen above, very uniform in their characters. A tendency has been observed
toward an increase of the spines of the meropodite of the cheliped in a direction
from the northeast toward the southwest within the range of this species. Here we
have apparently the first step toward the formation of a variety : a variation
becomes more frequent in the southwestern part of the range, possibly in conse-
quence of hereditary transmission, and begins to "breed true." But it is only the
beginning of it, the varying form not being found to the exclusion of the original,
and thus it remains " variation " only.

Other variations (mentioned p. 375 and 376) are very likely due to injuries re-
ceived during life84, and again others are of the character of freaks, namely the two
cases of apparent hermaphroditism. One of these is rather interesting (PL XXXIX,
Fig. 7d and 7e). Here the male sexual organs do not at all correspond to the typical
form of this species, but approach in shape to that known in the limosus-section. It
seems to me that we have to deal here with a case of atavism. The //mosws-section has
been regarded as the most primitive type of the subgenus Faxouius, on account of the
very slightly separated tips of the copulatory organs. The propinquu s-section is next
to it, but here the tips are separated for a greater distance. It is quite probable
that the latter section descended directly from the former, and it seems that in the
instance discussed the sexual organs have reverted to the original limosus-type, and
thus the assumption that the projiiiujiius-section is a descendant of the limusus-
section gains additional strength.

In the seven hundred and twenty-five specimens of C. bartoni we again have to
emphasize the great uniformity of the characters. The variations discussed are
rather insignificant, and consist chiefly in the shape of the rostrum and the size.
A single individual has been observed in which one lateral spine of the carapace
was present, apparently an atavistic feature. Other variations are of the hermaph-
roditic type.

No remarkable variations have been found in G. carolinus, and a few insignifi-

84 Variations due to injuries are most frequently observed in the case of regeneration of the chelipeds. I did
not mention them in the systematic part, since they are very common. If the claws are lost they are replaced by new
claws, which differ from the old ones not only in size, but also in shape. The fingers are proportionally longer, and the
palm proportionally shorter than in normal claws. This difference in shape remains even if the claws, after repeated
monlts, again attain a good size. Regenerated claws may always be recognized by the short palm and long fingers.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 503

cant ones in C. monongalensis. The same is true of <\ dwgenea, leaving out of
account the differences between the eastern and western forms. The most important
variation is that of the width of the areola. In this there is a tendency toward
regional restriction, but it is not complete. The wide areola, being a more primi-
tive character, does not represent the variation, but the original condition, which is
retained only in a small part of the range and is even there not general. It is a
character that has a tendency to disappear and may be classed under atavism. A
case of albinism has been observed in C. diogmes.

The conclusions from the above observations are that in the Cambanis forms of
Pennsylvania the morphological characters are very constant, and that the varia-
tions observed are generally only slight, diverging very little from the typical condi-
tions. Anything that looks like a " mutation " in De Vries' sense is entirely un-
known, for the cases of hermaphroditism cannot be regarded as such, and the cases
of atavism and albinism do not fall under it, being clearly of a " retrograde " char-
acter (De Vries, 1905, p. 121 el aeq.).

Further, even between most of our well established species differences are so
slight that they cannot be regarded as representing " mutations," that is to say,
sudden leaps in a progressive direction (De Vries, /. c, p. 141). This is most evident
in the projiinquus-section, where the three forms, two of which at least must he re-
garded as species, are distinguished by such insignificant characters that it is impos-
sible to talk of "leaps" or of " sudden changes." The same is true of the differ-
ences of C. ccirolinus, C. moibongalensis, and C. diogmes, the amount of the differences,
although well marked, being very small, and the " gaps " between these species be-
ing infinitesimal. The only striking difference is in color, but before we know what
causes the appearance of various colors we cannot express any judgment on this
point.

Even in those species which are more isolated from the rest, the differences do
not amount to much. In C. bartoni the depression of the carapace and the width
of the areola differ only in the degree of the development from the same characters
in the burrowing species. G. limosus is the most strongly marked species, but should
not be compared with the other river-species of Pennsylvania, but with its nearest
relations in southern Indiana (C. indianensis Hay), but then again the difference
is small and consists only of quantitative changes in the same features.

Thus the assumption of De Vries, that species have originated by sudden
leaps, does not find any support whatever in the conditions seen among the
crawfishes of Pennsylvania. On the contrary the close affinity of most of them.
and the comparative insignificance of the specific characters, supports the view

504 MEMOIRS OF THE CARNEGIE MUSEUM

that these species have originated out of rather slight variations from the original
forms. If this is evident in so small a territory, further investigations only can em-
phasize this, for additional material can only bring these forms closer together. (See
Merriam, 1906, p. 257.)

3. Formation, of Species by Isolation, as Exemplified by the
Pcnnsi/lra n ia C 'ra wfiskes

I have repeatedly maintained that the whole process of evolution in nature which
ends in the formation of " species," and which, consequently, may be called "origin
of species," is not subject to one single factor alone, such as " natural selection " or
" isolation " or " mutation " but that it is absolutely necessary that several factors
work together. (See Ortmann, 1896, p. 188 et seq.) Indeed none of these factors
is new, and they have been discussed by various writers, but generally too much
value has been attributed to one or the other of them to the detriment of the rest.
I have insisted, on the contrary, that four factors are equally necessary to form
species, namely : 1, variation ; 2, inheritance ; 3, natural selection ; 4, separation
(I. c, p. 190).

Of these the last one, S&paraition or Isolation, is the one which forms species. To
this is due the fact that the whole mass of organic beings to-day is divided up into
a large number of units, which we call " species." If it had never existed or acted
the process of evolution would have gone on nevertheless, but the organic world
would not consist of sjiecies; but since separation always has acted, species are pres-
ent. This does not imply that species should be everywhere well-marked. This
process is going on all the time, and in many cases it is not yet finished, and thus
it may be difficult sometimes to say whether a particular form is to be regarded as
a species or not ; but, as a rule, our inability to declare positively that a certain form
is a species is only due to the insufficiency of our knowledge.

Separation (or isolation) should not only be conceived of in its broad topograph-
ical and climatic aspect, but is, as I have always maintained, largely also ecological.
(See " bionomic separation," /. c, p. 190.)85 That it may occur under several forms
is amply demonstrated by the Pennsylvania crawfishes, and some form or other of
isolation is evident in every case without exception. Both topographical and eco-
logical separation are recognizable in our material, while climatic separation is
not observed on account of the insignificant differences of climate in the region
investigated.

85 " Barriers" are not necessary. Merriam, 1906, p. 249, thinks that the existence of sharp barriers is necessary for
isolation ; where such are absent he prefers to use the term " divarication." Possibly the term "habitudiual segrega-
tion," introduced by Gulick, 1905, p. 49, and 53 et seq. would be appropriate.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 505

(a) G. liniOSllS.

It has been repeatedly emphasized above, as well as in a previous paper, that
this species is well isolated morphologically and geographically. I have introduced
it as one of the examples for the rule that " morphologically isolated species occupy
isolated stations" (Ortman, 19056, p. 127), and also for the rule that "discon-
tinuity of distribution is a proof of antiquity " (ibid.). Both rules are beautifully illus-
trated by this species, particularly in contrast to the other rule that "closely allied-
species occupy neighboring areas." Thus not only the effect of isolation, as produc-
ing species, is evident in G. limosus, but it is'also seen that the degree of isolation
is in direct proportion to the sharpness of the expression of the specific characters.
C. limosus is geographically the most sharply isolated species of our crawfishes, its
area being several hundred miles distant from that of the most closely allied forms in
Indiana and Kentucky. Correspondingly it is also morphologically well marked,
being sharply distinguished from the other Pennsylvanian species, as well as from
species in the west which are closely related to it. Isolation in this case is purely
topographical, since the ecological habits of G. mdianensis seem to be similar, (Hay.
1896, p. 495) ; though another allied form, G. sloani, differs slightly ecologically,
(Faxon, 1885, p. 90).

(b) C. propinquus, C. jyropinquus sanborni, C. obscurus.

These three forms, as far as our present knowledge goes, are sharply separated
topographically, while they agree with each other ecologically, but the topograph-
ical boundaries between them are not everywhere uniformly sharp. In fact, the
ranges of these three forms are connected on the one side by the Ohio River, on the
other side by the basin of the Great Lakes.

As we have seen above, the present connection of these forms is a secondary
feature developed during the latter part of the Glacial Period, while anterior to this,
at the beginning of the Glacial epoch, different conditions prevailed, which were
different in turn from those of still earlier times. The history of these tonus was
probably as follows. At the end of the Tertiary a form corresponding to these three
crawfishes existed in the drainage of the Erigan River. Probably there was only a
single species resembling the present C. propinquus. This Bpecies lived m the Eri-
gan River, as well as in its southern tributaries, and there was do chance for it to
split up into different species, although variations may have occurred. When the
advancing ice of the Glacial Period covered the Erigan River and thus separated
the southern tributaries from each other, the latter formed lakes, and later, by over-
flow (or other means) they were connected again. Thus the present Ohio was
created. The temporary isolation of these rivers at the beginning of the Glacial

506 MEMOIRS OF THE CARNEGIE MUSEUM

epocli had its effect upon the crawfishes living in them. They developed into as
many species as there were rivers (three). Probably there was already" in Tertiary
times a tendency within the Erigan drainage to form variations and even geograph-
ical varieties, but the fact that these forms~(at least two of them) assumed the char-
acter of species is due to the pl^siographical features of the earlier Glacial Period.

After the Ohio was formed, and the connection between the areas of these
species was reestablished, there must have been a tendency among them to mix
along the course of the Ohio River. How it was in the case of G. propinquus and
C. sanbomi we do not know. But I have investigated the facts in the case of G.
sanborni and G. obscurus. Where they come together in the neighborhood of New
Martinsville, West Virginia, C. sanborni shows no tendency at all to go up the river,
no trace of it being found above New Martinsville. This apparently is due to the
greater difficulty of ascending the river and to contend with a species which is firmly
established there. On the other hand C. obscurus apparently has gone down the
river, and has invaded the original territory of G. sanborni, but it has done so only
to a small extent. For, although it is easier to descend a river, the fact that the
region invaded is occupied by another species with the same ecological habits must
make it rather difficult to oust the latter. Thus, although C. obscurus has the advan-
tage over C. saiihiirni, being favored in its migration by the fact that it is here down-
stream, this advantage is only a slight one, and did not enable C. obscurus to occupy
any of the territory of C. sanborni to the exclusion of the latter. It is found here
associated with it, but its numbers are small, and the original form still prevails.

A curious fact, however, has been observed. I have pointed out (p. 367 and p.
434) that the specimens of G. sanborni captured in Fishing Creek at New Martinsville
showed in certain characters an inclination toward G. obscurus. This suggests hybrid-
ization. Of course it is impossible to ascertain this positively without experiments,
but it seems that a crossing between these two forms is not altogether impossible,
for the shape of the sexual organs is very similar in both. They are generally very
closely allied, and further, their breeding seasons are identical, so that kyesame-
chania86 probably does not exist. This is further suggested by the conditions ob-
served in the Lake Erie drainage in Pennsylvania. Here C. propinquus and C.
obscurus come together, and again we pointed out (p. 365) that G. propinquus in this
region has a tendency towards C. obscurus. In both cases hybridization would
easily explain matters.

86 Impossibility of crossing, due to any cause, mechanical, physiological, or ecological, see Eimer, 1895, p. 14.
Gulick (1905, p. 95) calls this " Impregnational Isolation." Under this head falls also Romanes' "physiological
isolation."

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 507

Even if it should be the case that C. obscurm may cross with <'. propinquua and
C. propinquus sanborni, this does not invalidate its standing as a species, for we have
numerous examples in nature in which true species form hybrids.

Thus we see that these three species, the origin of which as species belongs to
the beginning of the Glacial time, have come together again ; but each seems to
remain in its original area, and where they come into actual contact the one species
is hardly able to oust the other. To a small degree hybridization seems to be pos-
sible at the points of contact. The assumption that C. sanborni might be a hybrid
between C. propinquus and C. obscfwrus is rendered impossible by the exclusive
presence of C. sanborni all over its range (excepting Fishing Creek), without any
trace of the two other species.

It remains to consider the question what the relation of the specific characters
to isolation may be. We see that in the case of C. obscurus it is chiefly the "shoulder"
of the male sexual organ which distinguishes this species. This shoulder is found
at a place where an external stimulus acts upon this organ, namely, just where it
is touched by the fifth pereiopod in the act of copulation. A similar shoulder is
found in many other species of Cambarus of different groups and even subgenera,
and thus it is highly probable that it is this external stimulus which induces the
development of this feature. But this does not afford us an explanation why this
shoulder did not develop in other species, especially in C. propmqum. At present
I am unable to answer this question. The fact remains that we have to deal with
a specific character, which is clearly due lo an external stimulus/7 and I have
always held the opinion that every variation is invariably caused by a reaction of
the organism to some external influence. (See Ortmann, 1896, p. 188, and 1898,
p. 157.) But the view that acquired characters are transmissible is not fashionable,
although now admitted by its chief adversary, Weismann. In consequence of the
modern tendency to deny the effect of external causes upon variation, at any rate
to deny the possibility of the hereditary transmission of such variations, not much
attention has been paid to the mutual relations between external stimuli and the
reaction of the organism upon them. But here I think much room for investigation
is left. In the present case the reaction of the organism upon the external stimulus
caused by the contact of the fifth pereiopod with the sexual organ is to form at the
point of contact a notch or angle (shoulder) on the sexual organ.

This reaction may be slightly advantageous, but it is not absolutely necessary,
for we see that there are many other species in which this reaction has not taken
place, even among the most closely allied forms, which are nevertheless well off and

S7Under "pressure of the environment," as Merriam puts it (1906, p. 24-4).

508 MEMOIRS OF THE CARNEGIE MUSEUM

flourishing. In other words, the "selectional value" of this character is practically
at the zero-mark. This demonstrates again that the conception of " natural selec-
tion " as " selection of the fittest " is incorrect. With regard to fitness there are many
characters which are entirely indifferent, and this is one of them. The absence or
presence of a rostral keel, and of tubercles in the case of the female annulus, the
other specific differences of these forms, belong to the same class. We thus see that
natural selection has played no part in the development of these characters of these
species. But this does not imply that selection has had nothing to do with the evo-
lution of these species, on the contrary this factor has always acted, and if these
characters had not been fit to survive, the species would not have been able to sur-
vive. Natural selection (in the modified sense, according to Pfeffer, see Ortmann,
1896, p. 176), resulted in the fact that the propinquus- group, such as it actually is,
is able to live and to flourish, but it is not responsible for the splitting up of this
group into two or three species.

The latter fact is entirely due to isolation. In the present case the isolation was
in effect only during a short period in the past, but it was enough to differentiate
several species. At the present time there is a tendency to undo this effect. These
species are beginning to mingle again. But this process has not yet progressed far,
and for several reasons will very likely be slow in future. It is hard to say what
the outcome will be, whether we shall have a hybrid form, or whether one will sup-
press the others. C. obscurus is the most advanced form, and also seems to be
slightly more vigorous than the others. Thus it may finally overrun them and
crowd them out, unless it is in turn conquered by a still more vigorous from, C.
rusticus, advancing from the southwest.

From the above discussion we see that whatever may have been the processes
of variation and of natural selection, or independently of what we may think
of the possibility of the inheritance of acquired characters, the fact that the
propinqu us-group has split up into species is solely due to isolation, which in this
case is strictly topographical. We have here three forms with identical ecological
habits, in which topographical isolation is evident, illustrating the rule that " closely
allied species occupy neighboring areas." (See Ortmann, 19056, p. 127, Jordan,
Science, Nov. 3, 1905, p. 546, and Merriam, 1906, p. 248, et seq.)

(c) C. bartoni.
This species is morphologically well isolated from the other Pennsylvanian
species, and also has peculiar ecological habits. Being found all over the state it
necessarily comes into contact with all the other species and is often found associated

OKTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 509

with them. This is preeminently the case with the river forms, C. limosus, C.
obscarus, and G. propvnquus.

Here we have an instance in which at a given locality two species may he found
side by side. This, however, is due to secondary processes. Originally each of the
two species had a different center of radiation, and thus we again see the action of
isolation. The center of G. bartoni lies in the mountains of the Appalachian sys-
tem; the common center of G limosus and the propinquus-grovip is in the central
basin of the Mississippi, and the special center of G. limosus in the coastal plain,
and that of the propi/nquvs-gvoxiTp in the Erigan and Lower Ohio drainage.

Nevertheless these species came together (see Ortmann, 1896, p. 186), but the
migration was in different directions, the river species coming up the rivers, while
C. bartoni migrated down stream. Although living side by side there is no danger
of hybridisation, since their morphological differences are such that kyesame-
chania exists. The different shape of the sexual organs of C. bartoni from that in
the subgenus Fazonius precludes any idea of their being able to cross. Such cases
do not offer anything remarkable, since the occupation of and the association at the
same locality of different forms coming from different directions, and not being
closely allied, is the general rule in any ecological community (bioc&nosis).

Conditions are slightly different in the cases where C. bartoni is found in close
proximity to the chimney-builders. Here there is closer affinity, but also it seems
here that these species are so far separated morphologically that kyesamechania
exists, although the shape of the copulatory organs is similar. Moreover, wherever
C. bartoni comes into contact with the burrowing species it generally occupies situa-
tions slightly different from those preferred by the chimney-builders. It favors
running water in open streams, while the burro wers are found in holes at a certain
distance from the streams. Nevertheless, G. bartoni is sometimes found in burrows
and in springs close to the one or the other of the burrowers (it is even found in the
holes of the latter, see p. 414), but in such cases we have again the same conditions
as above : different species coming from different centers occupy the same
locality.

Yet as a rule G. bartoni occupies a different habitat from the burrowers, even if
found close to the latter. A fine illustration of this is in Nine-Mile Run, near
Pittsburgh. Here three species, C. bartom, G. mortongalenais, and G diogmes are
found together upon a space hardly more than twenty feet square. The locality is
a pile of talus swept down into the valley of Nine-Mile Run by a small stream.
The stream comes through an insignificant ravine, and spreads out over the talus,
forming a kind of a delta, rendering the lower parts of the pile of talus rather

510 MEMOIRS OF THE CARNEGIE MUSEUM

swampy. At the upper end of the talus, in the outcrop of sandstone rock, and not
far (about fifteen feet) from the bed of the spring, is a copious spring, the water of
which runs directly into the clay and humus of the pile of talus, in a large part
underground. C. bartoni is found in the small stream under stones ; G. monon-
galensis is found at and immediately below the spring referred to ; and C. diogcnes
is abundant all over the pile of talus down to the bottom of the valley. At the
upper end of the pile of talus is the place where all three species come close to-
gether, but each is subject to different ecological conditions.

Similar conditions have been frequently observed, and we thus have here the
occupation of the same localities by closely allied species, which differ ecologically,
that is to say, topjographical isolation is not observed here, but the isolation is eco-
logical, and the differentiation of the chimney- builders from G. bartoni very likely
is connected with and largely due to the latter.

(d) C. carolinus and G. monongalensis.

We have seen that these two species are very closely allied, but that the distin-
guishing characters are constant. Ecologically they are similar, so that hybridisa-
tion might occur when they come together. The latter case, however, has never
been observed, at least in Pennsylvania, Maryland, and northern West Virginia.
The western escarpment of the Chestnut Ridge forms a sharp boundary between
them. This case corresponds to that observed in the western river-species (prop-
inqaus-group). Two species identical in their ecological habits are separated topo-
graphically. But in this case the barrier separating them is of a different character.
What the essential feature of this barrier is, is hard to say. Chestnut Ridge in many
respects forms a boundary. Altitude seems to play a part, but whether it is para-
mount is doubtful. Absence of extensive deposits of clay on the western side of
this ridge on account of the destruction of the Old Tertiary base-level by subsequent
erosion, may also be of importance. Further studies in West Virginia surely will
lead to a solution of the question, but this much is certain, that these two species
again illustrate the rule that " closely allied species occupy neighboring areas," and
further they illustrate the fact that specific differentiation is due to isolation, which
is topographical in this case.

What are the actual causes of the difference of the specific characters (color,
shape of rostrum, and sculpture of chelipeds), that is to say, what external influ-
ences are responsible for them is even more obscure, as it is in the case of the pn>-
pinquus-grou\).

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 511

(e) C. diogenes.

C. diogenes is sharply separated from the other chimney-builders, but resembles
them ecologically to a certain degree. In Pennsylvania it comes into contact with
them, but in the case of C. caroHnus this has been observed only once, while it is
more frequent in the case of C. monongalensis. However, intermediate forms have
never been observed, so that we must assume that kyesamechania prevents crossing.

In both cases, with reference to G. caroliwus as well as C. monongalensis, it is to
be remarked that whenever one of these is found associated with G. diogenes it
is always only a contact, not a real mixing of both forms. This is best observed in
the case of G. monongalensis and C. diogenes. All over the range of C. monongalensis
in southwestern Pennsylvania C. diogenes is also found. But as has been stated
(p. 417 and 458), although they frequently dwell at the same localities they do not
occupy the identical locations, C. diogenes belonging to a lower level than G. mononga-
lensis. Thus we see again a separation, which is primarily expressed in the difference of
altitude. Whether the latter is most important seems doubtful. It has been stated
that C. monongalensis prefers spring-water, while C. diogenes lives mostly in swamps,
where the water is more or less stagnant and not so cool in summer. (Compare the
instance from Nine-Mile Run given above.) But, whatever may be the essential
feature which separates both species, it is clear that it is an ecological factor, and,
when these two species are found together, it is at a place where the ecological con-
ditions favorable to them come together.

That C. diogenes depends on different ecological laws from G. monongalensis is
also evident from the fact that the former has, outside of Pennsylvania, an entirely
different range.

Thus we have here a case similar to that of C. bartoni when it associates with the
burrowing forms. Two allied species occupy (in Pennsylvania) almost the same ter-
ritory, and are not separated topographically, but their ecological separation is evi-
dent, and very likely is connected with their specific differentiation.

In the two races C. diogenes, the eastern and western, we again see the influence
of separation. According to our theory that the area of G. diogenes was a unit in
Preglacial times, and that it was separated by the advancing ice into an eastern and a
western section, which subsequently remained separate, we must expect, if isolation
effects specific differentiation, that the eastern and western form of G. diogenes should
show at least a tendency to develop differential characters of specific value. This is
indeed the case, as we have seen above (p. 401 et seq).

Isolation, or Hahitudinal Segregation, as the factor forming species, is thus
clearly seen in every case discussed. We may condense the results obtained in the
following sentences.

512 MEMOIRS OF THE CARNEGIE MUSEUM

1. The normal case is when two closely allied species, possessing identical or
nearly identical ecological habits occupy separated areas, which lie close together but
do not overlap. (Examples: propinquus-groxip ; G. carolinus and G. monongalensis.)

2. Whenever allied species are found in one and the same locality (overlapping),
isolation becomes apparent in the following forms.

(a) The two species have different centers of origin, that is to say, they were
separated formerly, but occupied the same territory subsequently. In this case, if
very closely allied, hybridization maybe possible (C. obscurus and G. sanbomi at
New Martinsville, and C. obscurus and C. propinqmis in the Lake Erie drainage),
if no kyesamechania exists. If the latter is present, which always means that the
two species in question are less closely allied, the two species may actually live
side by side under identical conditions (C. bartoni and the river-species), or one may
conquer and suppress the other. No instances of the latter kind are known in
Pennsylvania, but may possibly occur in southwestern Ohio and in Indiana, between
G. rusticus and G. propinquus.

(b) If the centers of origin are more or less identical (absolute identity is hardly
possible), the two species always differ ecologically, and although living at the
same localities, prefer different surroundings. In this case they are not so closely
associated, and they generally remain at a certain distance from one another,
although their general areas are overlapping. Under such conditions hybridisation
might occur, but it has not been observed in Pennsylvania, and the species existing
under such conditions are probably separated by kyesamechania. (Example : C.
diogenes and monongalensis.)

Case (a) and (b) may be combined, that is to say, two species living together may
have different centers of origin and may be ecologically different. This is seen in
the example of C. bartoni and the burrowing species.

I believe that in every case where closely allied species overlap in parts of their
ranges a close investigation will reveal that one or the other of the above cases is
realized. Isolation is, in my opinion, a necessary factor in the differentiation of
species, and I do not think that a case ever will be discovered where two closely
allied species possess precisely the same distribution. But in order to ascertain this
a mere superficial knowledge of the species in question and their range is insuffi-
cient, and every case should be investigated as exactly as possible, in a manner
similar to the above studies.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA

513

1902.

Abbe, C.

1873.

Abbott, C. C.

1884.

Abbott, C. C.

1885.

Abbott, C. C.

1902.

Adams, C. C.

1905.

Adams, C. C.

1895.

Andrews, E. A.

1904.

Andrews, E. A,

1859.

Bell, H.

1877.

Bundy, W. F.

1882.

Bundy, W. F.

1883.

Bundy, W. F.

1903. Campbell, M. R.

1880.
1904.

Carll, J. F.

COPELAND, E. B.

1889. Davis, W. M.

1844.
1905.

De Kay, J. E.
De Vries, H.

1895. Eimer, G. H. T.

BIBLIOGRAPHY.

The Physiography of Garrett County. Maryland Geological Sur-
vey, Garrett County.

Notes on the Habits of Certain Crawfish. American Naturalist,
VII, p. 80.

Are the "Chimneys" of Burrowing Crayfish Designed? Ameri-
can Naturalist, XVIII, p. 1,157.

How the Burrowing Crayfish Works. The Inland Monthly, I,
p. 31.

The Southern United States as a Center of Geographic Distribu-
tion of Flora and Fauna. Biol. Bull., Ill, p. 115.

The Postglacial Dispersal of the North American Biota. Biol.
Bull,, IX, p. 53.

Conjugation in an American Crayfish. American Naturalist,
XXIX, p. 867.

Breeding Habits of Crayfish. American Naturalist, XXXVIII,
p. 165.

On the Natural History of the Gulf of St. Lawrence, and the
Distribution of the Mollusca of Eastern Canada. Canadian
Naturalist and Geologist, IV, p. 210.

On the Cambari of Northern Indiana. Proc. Acad. Nat. Sciences
of Pli i/ade/ph ia, p. 171.

A List of the Crustacea of Wisconsin, with Notes on Some New or
Little Known Species. Trans. Wisconsin Academy, V, p. 177.

The Crustacean Fauna of Wisconsin, with Descriptions of Little
Known Species of Cambarus. Qeal. Surv. Wine, 1873-79,
I, p. 402.

Geographic Development of Northern Pennsylvania and Southern
New York. Bull. Geol. 8oc. Amer., XIV, p. 277.

The Geology of the Oil Region. Secoml deal. 8urv. /'"., 13.

The Variation of some California Plants. Bot<in. Gazette,
XXXVIII, p. 401.

The Rivers and Valleys of Pennsylvania. National Qeogr. .!/".</.,

1, p. 183.

Zoology of New York, Part 6, Crustacea, p. 22.

Species and Varieties, Their Origin by Mutation. Edited by D.

T. MacDougal.
Die Artbildung und Verwandtschaft bei deu Sclimetterlingen.

2. Teil.

514

MEMOIRS OF THE CARNEGIE MUSEUM

1846. Erichson, W. F.

1798.
1896.
1884a.

Fabricius, J. C.
Fairchild, H. L.
Faxon, W.

18846. Faxon, W.

1885a.

Faxon, W.

18856.

Faxon, W.

1885c.

Faxon, W.

1890.

Faxon, W.

1898.

Faxon, W.

1876.

Forbes, S. A.

1890.

Foshay, P. M.

1887.

Ganong, W. F.

1850.

Gibbes, L. R.

1852.

GlRARD, C.

1833. Goodman, J. D.

1841. Gould, A. A.

1901. Grabau, A. W.

1905. Gulick, J. T.

1870. Hagen, H. A.

Uebersicht der Arten der Gattung Astacus. Arch. f. Naturg

Jahrg., XII, p. 86.
Sapplementum Entomologise Systematica?.
Glacial Genessee Lakes. Bull. Geol. Soc. Amer., VII, p. 423.
On the So-called Dimorphism in the Genus Cambarus. Amer.

Journ. Sci,, XXVII, p. 42.
Descriptions of New Species of Cambarus. Proc. Amer. Acad.

Science, XX, p. 107.
A Revision of the Astacidse. Mem. Mus. Harvard, X.
A List of the Astacidse in the U. S. National Museum. Pr. U.

S. Nat. Mus., 1885, p. 356.
Preliminary Catalogue of the Crayfishes of Kansas. Bull. Wash-
burn Laboratory, Topeka, I, p. 140.
Notes on North American Crayfishes, Family Astacidse. Pr. U.

S. Nat, Mus., XII, p. 619.
Observations on the Astacidse in the LT. S. National Museum and

in the Museum of Comparative Zoology, with Descriptions of

New Species. Pr. U. S. Nat. Mus., XX, p. 643.
List of Illinois Crustacea, with Descriptions of New Species. Bull.

Illin. Mus., I, p. 3.
Preglacial Drainage and Recent Geological History of Western

Pennsylvania. Amer. Journ. Sci., XL, p. 397.
The Crayfish in New Brunswick. Bull. Nat. Hist. Soc. N Brunsw.,

VI, p. 74.
On the Carcinological Collections of the United States. Pr. Amer.

Assoc, III, p. 167.
A Revision of the North American Astaci, with Observations on

their Habits and Geographical Distribution. Pr. Acad. Nat.

Sciences, Philadelphia, VI, p. 87.
Rambles of a Naturalist.88

Report on the Invertebrata of Massachusetts, p. 330.
Guide to the Geology and Paleontology of Niagara Falls and

Vicinity. Bid!. N. Y. State Museum, XLV.
Evolution, Racial and Habitudinal. Carnegie Institution of

Washington,
Monograph of the North American Astaeidae. Illustr. Catal.

Mus. Harvard, III.

98 1 have only seen a copy of a later edition, in "American Natural History," Vol. II, 3d ed., 1842, p. 293 (Library
of Acad. Nat. So., Phila. ). Girard and Hagen quote this work as of 1833 ; Faxon (1885a, p. 63) as of 1859.

ORTMANN: THE CRAWFISHES OF THE STATE OF PENNSYLVANIA

515

1835.

Hari.an, R.

1900.

Harris, J. A.

1903.

Harris, J. A.

1896.

Hay, W. P.

1899.

Hay, W. P.

1902a.

Hay, W. P.

19026.

Hay, W. P.

1905.

Hay, W. P.

1903.

Hice, R. R.

1904.

HOLLISTER.

1905.

Hoyt, J. C. and

1904. Hulbert, A. B.

1903.

Jenkins, H. M.

1903.

Jillson, B. C.

1900.

Klein, T. B.

1903.

Leighton, M. O.

1904.

Leighton, M. C.

1902.

Lever ett, P.

1884.

Lewis, H. C.

1888.

McGee, W J

1906. Merriam, C. H,

Description of Three Species of the Genus Astacus inhabiting the
United States. Medio, and Physio. Research, p. 229.

Annotated Catalogue of the Crayfishes of Kansas. Kama. Univ.
Quart, IX, p. 263.

The Habits of Cambarus. Amer, Natural., XXXVII, p. 601.

The Crawfishes of the State of Indiana. SOth Rep. Indiana Geol.
Sun:, p. 475.

Synopsis of North American Invertebrates. 6. The Astacidse of
North America. Americ. Natural., XXXIII, p. 957.

Observations on the Crustacean Fauna of the Region about Mam-
moth Cave, Kentucky. Pr. U. 8. Nat. Mus., XXV, p. 223.

On the Proper Application of the Name ( ambarus earolimu
Erichson. Pr. Biol. Soc. Washington, X"V, p. 38.

Instances of Hermaphroditism in Crayfishes. Smithson. Misc. < oil.,
XLVIII, p. 222.

Northward Flow of Ancient Beaver River. Butt. Geol. Soc.
Amer., XIV, p. 297.

See Leighton and Hollister.
Anderson R. H. Hydrography of the Susquehanna River Drainage
Basin. U. 8. Geol. Sun-., Water 8uppL and Irrig. Papers, 109.

The Great American Canals, Vol. II. (Historic Highways of

America, Vol. XIII.)
Pennsylvania, Colonial and Federal. Vol. III.

River Terraces in and near Pittsburgh. TV. Acad. 8ei. and Art
of Pittsburg,

The Canals of Pennsylvania, and the Systems of Internal Improve-
ments of the Commonwealth. Ann. Rep. Secret. Intern, Aff.
of Pa., 28, Part 4, p. LXIV.

Normal and Polluted Waters in Northeastern United States. U.
S. Geol. Sun:, Water Suppl. and Irrig. Papers, 79.
and Hollister, G. B. Quality of Water in the Susquehanna Drainage
Basin, witli an Introductory Chapter on Physiographic Features.
U. S. Geol. 8urv., Water Suppl. and Irrig. Papers, 108.

Glacial Formations and Drainage Features of the Erie and Ohio
Basins. M<>nt></r. I'. 8. Geol. 8urv., XLI.

Report on the Terminal Moraine. Second! Geol. Sun: Pa., '/■■

Three Formations of the Middle Atlantic Slope. Amer. Journ.
8oi., XXXV, ]>. 1 20, 328, 367, 448.

Is Mutation a Factor in the Evolution of the Higher Verte-
brates? Science, N. 8., XXIII, Febr. 16, p. 241.

516

MEMOIRS OF THE CARNEGIE MUSEUM

1837.
1902.
1891.

1896.

1898.
1899-

1905a

19056

Milne-Edward, H. Histoire Natnrelle des Crustaces. II, p. 331.

Newsom, J. F.
Ortmann, A. E.

Ortmann, A. E.

Ortmann, A. E.
900. Ortmann,

Ortmann, A. E.

Ortmann, A. E.

Zool.

Soc,

Drainage of Southern Indiana. Journ. Geol., X, p. 1 66.
Die Decapoden Krebse des Strassburger Museums. III.

Jahrb., 8yd. VI, p. 12.
On Natural Selection and Separation. Pr. Amer. Philos

XXXV, p. 175.
Ueber Keini variation. Biolog. Cen&ralbl., XVIII, p. 139.
1900. Ortmann, A. E. Bronn's Klassen und Ordnungen des Tierreicbs, V, Crust. 2, p.

1210, 1242.
The Crawfishes of Western Pennsylvania. Ann. Carnegie Mus.,

Ill, p. 387.
The Mutual Affinities of the Species of the Genus Cambarus and

their Dispersal over the United States. Pr. Amer. Philos. Soc,

XLIV, p. 91.
Osborn, R. C. and Williamson, E. B. The Crayfish of Ohio. Sixth Ann. Pep. Ohio

State Acad. Sci., p. 21.
Higher Crustacea of New York City. Bull. N. Y. State Mus., 91,

p. 117.
Physiographic Regions of the United States. (The Physiography

of the United States. Nat. Geogr. Mag., p. 65.)
Remarks on the Mammoth Cave and Some of Its Animals. Bull.

Essex Inst., VI, p. 191.
Synopsis of Four New Genera and Ten New Species of Crustacea

Found in the United States. Americ. Monthly Magazine, II, p. 40.
Fauna of New England. 5. List of the Crustacea. Occas. Papas.

Boston Soc. Nat. Hist., VII, p. 18.
Nomenclature of Colors.
An Account of the Crustacea of the United States. Journ. Acad.

Nat. Sci. Philadelphia, I, p. 167.
The Chimneys of Burrowing Crayfish. The Observer, VII, No.

3, p. 85.
The Crustacea of the Freshwaters of the Northern United States,

Rep. U. S. Fish Comm, for 1872 and 187S, p. 637.
Discovery of the Preglacial Outlet of the Basin of Lake Erie into

that of Lake Ontario ; with Notes on the Origin of our Lower

Great Lakes. Pr. Amer. Philos. Soc, XIX, p. 300.
A Review of the History of the Great Lakes. Amer. Geologist,

XIV, p. 289.
The Ligonier Valley. Second Geol. Surv. Pa., K3.

1898.

Osborn, R. C. anc

1905.

Paulmier, F. C.

1896.

Powell, J. W.

1874.

Putnam, F. W.

1817.

Rafinesque, C. S.

1905.

Rathbun, M. J.

1886.

Ridgway, R.

1617.

Say, T.

1896.

Shufeldt, R. W.

1874.

Smith, S. I.

1881.

Spencer, J. W.

1894.

Spencer, J. W.

1878.

Stevenson, J. J.

ORTMANN

T

1884.

Tarr, R. S.

1842.

Thompson, Z.

1903.

Tight, W. G.

1886.

Underwood,

L.

M.

1896.

Ward, H. B.

1896.

White, I. C.

1899.

Williamson,

E.

]}.

1901.

Williamson,

E.

B.

1905.

Williamson,

E.

B,

1896.

Willis, B.

THE CRAWFISHES OF THE STATE OF PENNSYLVANIA

517

Habits of Burrowing Crayfishes in the United States. Nature,

XXX, p. 127.
Natural History of Vermont, I, p. 1 70.
Drainage Modifications in Southeastern Ohio and Adjacent Parts

of West Virginia and Kentucky. U. S. Geo/. Sure., Profess.

Papers XIII.
List of the Described Species of Freshwater < Iruatacea from

America North of Mexico. Bull. III. Stale Laboratory, II, p. 365.
A Biological Examination of Lake Michigan in the Traverse Bay

Region. B\dl. Mich. Fish. Comm., VI, p. 15.
Origin of the High Terrace Deposits of the Monongahola River.

Amer. Geologist, XVIII, p. 368.
Notes on Ohio Astacidre. Seventh Ann. Rep. Ohio Stale Acad.,

p. 47.
The Crayfish of Allegheny County, Pa. Ann. Cam. Mus.,Ifp. 8.
Crayfishes from Kentucky. Ohio Natural., V, p. 310.
The Northern Appalachians. (The Physiography of the United

States. Nat. Geogr. Mag., p. 169.)

518 MEMOIRS OF THE CARNEGIE MUSEUM

EXPLANATION OF PLATES.

Plate A.

Fig. 1. Cambarus obscurus Hagen. Male of the first form, natural size. Collected by the

writer, Sept. 7, 1905, in the Alleghany River, Sandy Creek, Allegheny County.
Fig. 2. Cambarus obscurus Hagen. Female, natural size. From same locality.
Fig. 3. Cambarus diogenes Girard. Male of the first form, natural size. Collected by the

writer, Aug. 26, 1905, at Baden, Beaver County.
Fig. 4. Cambarus carolinus Erichson. Female, natural size. Collected by the writer, Sept. 5,
1905, at Rainier Park, Ohiopyle, Fayette County.

Plate B.
Fig. 1. Cambarus bartoni (Fabricius). Male of first form, natural size. Collected by the

writer, Aug. 7, 1905, in Fern Hollow, Pittsburgh.
Fig. 2. Cambarus bartoni robustus (Girard). Female, natural size. Collected by Miss G.

Kinzer, Aug. 27, 1905, at Sixteen Mile Creek, Northeast, Erie County.
Fig. 3. Cambarus limosus (Rafinesque). Female, natural size. Collected by the writer, Sept.

10, 1905, in the Schuylkill Canal, Manayunk, Philadelphia County.
Fig. 4. Cambarus monongalensis Ortmann. Female, natural size. Collected by the writer,

Aug. 18, 1905, at Edgewood Park, Allegheny County.

Plate XXXIX.
Fig. 1. Cambarus bartoni (Fabricius). Rostrum. All figures 4-

la. Female, 70 mm. long. Collected by the writer, June 3, 1904, in North Versailles
Township, Allegheny County, opposite Stewart. Catalogue number 74.327. Shape
very broad, margins parallel. Not rare in western Pennsylvania.

16. Female, 52 mm. long. Collected by the writer, Aug. 22, 1905, at Squaw Run, Alle-
gheny County. Catalogue number 74.626. Shape typical ; very frequent.

lc. Male, first form, 63 mm. long. Collected by the writer, Sept. 16, 1904, at Valley
Forge, Chester County. Catalogue number 74.413. Shape typical, and character-
istic of eastern specimens, but also found in the west.

Id. Young male, second form, 34 mm. long. Collected by the writer, Aug. 22, 1905,
in Squaw Run, Allegheny County. Catalogue number 74.626. Usual shape in
young specimens.

le. Young female, 21 mm. long. Collected by the writer, June 25, 1904, in Jacob's
Creek, Laurelville, Fayette County. Catalogue number 74.356. Slightly longer
than usual, but not rare in young specimens.

If. Male, first form, 78 mm. long. Collected by the writer, May 27, 1904, in Squaw Run,
Allegheny County. Catalogue number 74.320. Unusually short and strongly taper-
ing, with exceptionally thick margin.

OKTMANN: THE CRAWFISHES 0V THE STATE OF PENNSYLVANIA 519

Fig. 2. Cambarus baHom robustus (Girard). Rostrum. *.

2a. Female, 89 mm. long. Collected by the writer ( tot. 4, 1904, in Temple Creek, Albion,
Erie County. Catalogue number 74.4.!o. Normal shape.

26. Young female, 18 mm. long. Collected by Miss G. Kinzer, Aug. 27, 1905, in Sixteen-
Mile Creek, Northeast, Erie County. Catalogue number 74.030. Normal shape in
young specimens.
Fro. 3. Cambarus caroHnus Erichson, Rostrum. -L

3«. Male, first form, 00 mm. long. Collected by the writer, Sept. 5, 1905, at Ohiopvle,
Fayette County. Catalogue number 74.640. Normal shape.

36. Male, second form, 30 mm. long. Collected by the writer, June 11, 1904, at Indian
Creek, Fayette County. Catalogue number 74.30"). Shape exceptional Most ex-
treme case as regards convergence of margins.
Fig. 4. Cambarus monongcUensia Ortman. Rostrum, i.

4a. Male, first form, 63.5 mm. long. (Type.) Collected by the writer, May 21, 1905, a
Edgewood Park, Allegheny County. Catalogue number 74.316. Normal shape.

46. Female, 73 mm. long. Collected by the writer, Oct. 12, 1904, at Hill, Westmoreland
County. Catalogue number, 74.440. Shape unusually broad, and margins almost
parallel. Most extreme ease in this direction, standing rather isolated.
Fig. 5. Cambarus Umosm (Rafinesque). Left first pleopod of male, first form. ».

5a. Inner view. Collected by the writer, Sept. 19, 1904, in Marcus Hook Creek, Mar-
cus Hook, Delaware County. Catalogue number 74.423.

56. Posterior view of same.

Fig. 6. Cambarus prtpmqvus Girard. Left first pleopod of male. f.

6. Inner view, male, first form. Collected by the writer, Oct. 4, 1904, in Temple (reek,

Albion, Erie County. Catalogue number 74.4.'.!).
66. Inner view, male, second form. Collected by the writer, June 7, 1904, in a tributary

of Conueaut Creek, Conneautville Station, Crawford County. Catalogue number

74.336.

Fig. 7. Cambarus obseurus Hagen. Left first pleopod of male. f.

7a. Inner view, male, first form. Collected by the writer, Aug. 24, 19(14, in the Ohio
River, Ambridge, Beaver County. Catalogue number 74.401.

76. Posterior view of same (horny tip of outer part hidden behind inner pari).

7c. Inner view, male, second form. Collected by the writer, June 24, 1904, in the Loval-
hanna River, Crisp, Westmoreland County. Catalogue number 74.352.

Id. Inner view, hermaphroditic specimen (type of male, first form). Collected by Atkin-
son, Graf, and Williamson, May 14, 1899, in the Ohio River, Neville Island, Alle-
gheny County. Catalogue number 7.436. (See text, p. 376.)

7e. Posterior view of same.

520 MEMOIRS OF THE CARNEGIE MUSEUM

Fig. 8. Cambarus bartoni (Fabricius). Inner view of left first pleopod of male, first form. ■£,
Collected by the writer, Oct. 6, 1905, at Weskit, near Kittanning, Armstrong
County. Catalogue number 74.665.

Fig. 9. Cambarus oarolinus Erichsou. Inner view of left first pleopod of male, first form. ^.
Collected by the writer, Sept. 7, 1904, at Dunbar, Fayette County. Catalogue
number 74.410.

Fig. 10. Cambarus monongalensis Ortmann. Inner view of left first pleopod of male, first form
(cotype), |. Collected by the writer, Oct. 10, 1903, at Edgewood Park, Allegheny
Couuty. Catalogue number 74.182.

Fig. 11. Cambarus diogenes Girard. Inner view of left first pleopod of male, first form. ^.

Collected by the writer, September 5, 1904, at Smithfield, Fayette County. Cata-
logue number 74.406.

Plate XL.

Fig. 1. Cambarus obscurus Hagen. Upper view of right chela of a male, first form, 77 mm.

long, natural size. Collected by the writer, Sept. 30, 1905, in the Alleghany River,

Twelve-Mile Island, Allegheny County. Catalogue number 74.663.
Fig. 2. Cambarus bartoni (Fabricius). Upper view of right chela of a male, first form, 82 mm.

long, natural size. Collected by the writer, Nov. 22, 1905, in Fern Hollow, Pitts-
burgh, Allegheny County. Catalogue number 74.681.
Fig. 3. Cambarus bartoni robustus (Girard). Upper view of right chela of a male, first form

98 mm. long, natural size. Collected by the writer, July 11, 1905, at Spartansburg

Crawford County. Catalogue number 74.596.
Fig. 4. Cambarus carolinus Erichsou. Upper view of right chela of a female, 77 mm. long,

natural size. Collected by the writer, Oct. 16, 1905, at Dunbar, Fayette County.

Catalogue number 74.669.
Fig. 5. Cambarus monongalensis Ortmann. Upper view of right chela of female, 71 mm. long,

natural size. Collected by the writer at Edgewood Park, Allegheny County, April

4, 1905. Catalogue number 74.495.
Fig. 6. Cambarus diogenes Girard. (Eastern form.) Upper view of right chela of a male, first

form, 83 mm. long, natural size. Collected by the writer, Sept. 21, 1905, at Ridley

Park, Delaware County. Catalogue number 74.654.
Fig. 7. Cambarus diogenes Girard. (Western form.) Upper view of right chela of a male,

second form, 93 mm. long, natural size. Collected by the writer, April 15, 1905,

at Millvalle, Allegheny County. Catalogue number 74.507.
Fig. 8. Burrow of Cambarus bartoni (Fabricius). In spring on hillside, west of Spruce Run,

Avalon, Allegheny Couuty, opened by the writer, July 2, 1904.
8a. Side view (section) ; 86. Upper view, mp, pile of mud consisting of clay, sand, and

small stones; d, ditch; x, place where crawfish, female, 52 mm. long, was found.
-At x springwater was running into the hole in a strong flow, and running out through

the hole, over and past the pile of mud into a ditch.

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 521

Fig. 9. Burrow of Cambarus oarolinus Ericfason, located in a swampy place in stiff yellow clay,
at Listie, Somerset County. Opened by the writer Aug. 12, 1904.
9«. Diagram of disposition of piles of mud seen from above. 96. Section of hole along line
A-D. 9c. Section of hole along line A-B-C. A. Open chimney. B. Closed
chimney. D. Closed chimney, hole filled up a good distance below surface of ground.
C. Open hole, without pile of mud, situated under the edge of a large flat stone (*).
wl, water level ; x, place where the crawfish (male, first form, 61 mm. long) was found.

Plate XLI.

Fig. 1. Burrow of CambarvA bartmi (Fabricius). Located in the sand and gravel of the dry
bed of a small stream, Edgewood Park, Allegheny County. Opened by the writer,
Oct. 10, 1903. vip, pile of mud, consisting of mud, sand, and gravel ; 8, large slab
of stone, lying imbedded in sand and gravel ; W, water level (the stream was dry for
long stretches, only here and there pools of water were left); X, place where crawfish
(female, 03.5 mm. long) was taken.

FlG. 2. Burrow of Cambarus monongal.ends Ortmann. Located in yellow clay (mixed with
humus), at a springy place on the bank of small stream, near Monongahela City,
Washington County. Dug out by the writer, June 16, 1904.
2a. Diagram of burrow and chimneys, seen from above ; 26, section of hole along line
A-B-C; 2c, section of hole along line C-D-E. A, hole opening laterally, with
one-sided pile of mud in front, keeping up the level of water ; B and D, closed
chimneys ; C,open, large, and regular chimney ; «'/, water level ; at, stream ; x, places
where the old female (mother, 65 mm. long), and ten young (20.5 to 32.5 mm. long)
were found. Water, in a weak flow, was running in at E, and was running out at A.

Fig. 3. Burrow of Cambarus monongalensh Ortmann. Located in yellow clay, at a springy
place on the bank of a small stream, Edgewood Park, Allegheny < '.unity. Dug out
by the writer, May 9, 1904. The burrow is of a type similar to the one figured in
Fig. 2, but less complex, a, hole opening laterally, with one-sided pile of mud keep-
ing up the level of the water ; 6, closed chimney ; wl, water level ; .«t, stream ; X, plaoe
where the crawfish (female, 63 mm. long) was taken.

Fig. 4. Burrow of Cambarus monmffalenns Ortmann. Located in black muck, at a springy
and swampy place at the bottom of the upper part of Fern Hollow, Pittsburgh, Alle-
gheny County, opened by the writer, Oct. 18, 1903. Type of a hole in level ground,
with the water near the surface. No adults and only four young were found in this
hole, but possibly the hole had additional branches, which were not discovered, the
high stage of the water and its icy coldness rendering investigation difficult. About
1.50 m. from this hole another was opened, which contained a female C. avtgena.
a, closed chimney ; 6, one-sided chimney in front of hole opening obliquely ; ,/•/, water
level ; x, places where young specimens (11.5 to 16.5 mm. long) were found.

522 MEMOIRS OF THE CARNEGIE MUSEUM

Fig. 5. Burrow of Oambarus dint/cues Girard. Located in stiff blue clay, in a ditch on a road-
side, Nine-Mile Run, Pittsburgh, Allegheny County. Opened by the writer, Nov.
5, 1904. The season had been very dry, and not much water was in the hole.
Pebbles were lying on the bottom of the hole, a, old chimney, leveled down by rain,
probably built in spring; b, fresh mud, brought up recently (beginning of fall
activity) ; wl, water level ; x, place where the specimen (female, 77 mm. long) was
taken.

Fig. 6. Burrow of Oambarua diogenes Girard. Located in yellow clay and humus, at a springy
and swampy place in woods on the side of a wagon road, upon which water was
standing (after a heavy thunder-shower on the previous day), at Squaw Run, Alle-
gheny County. Pug out by the writer, May 27, 1904. (/, chimney, consisting of
yellow clay ; b, " stopper " in the mouth of the chimney, distinctly differing from the
chimney, the material being yellow clay mixed with blackish mud and leaf-mould ;
wl, water level ; r, road, with mud-puddle upon it ; x, place where the crawfish (male,
first form, 70 mm. long) was found.

Fig. 7. Burrow of Oambarus diogenes Girard. Located in yellow and blue clay, on the border
of a swampy place, Schenley Farm, Pittsburgh, Allegheny County. After a sketch
furnished by Mr. F. E. Kelly, Nov. 14, 1904. sw, swamp; be, blue clay ; yc, yel-
low clay; a, one-sided chimney, consisting of yellow clay (probably made in spring
and summer) ; b, new chimney, consisting of blue clay (fall activity, reclaiming of
old burrow at c) ; e, old burrow, filled in (during summer) with blue clav, taken or
washed in from near the mouth of the lower entrance (a) of burrow ; wl, water level ;
x, place where the crawfish was taken.

Fig. 8. Burrow of Oambarua diogenes Girard. Located in blue and yellow clay on the bank
of a small stream, Schenley Farm, Pittsburg, Allegheny County. After a sketch
drawn by Mr. F. E. Kelly, Nov. 15, 1904. s, stream ; be, blue clay ; ye, yellow
clay ; a, new chimney, consisting of yellow clay, evidently coming from the newly dug
shaft going down vertically ; b, upper end of ascending branch of hole, without open-
ing (possibly originally open, but sealed up, and the pile of mud overgrown and
obliterated by vegetation) ; wl, water level ; x, place where the crawfish was taken.
The chimney at a shows fall activity, and the vertical shaft is being built by the
crawfish in order to get deeper down into the ground.

Plate XLII.
Fig. 1. Preglacial Monongahela River, after Leverett (1902, p. 89, fig. 1).
Fig. 2. Present range of Oambarus obscurus Hagen and C. propinquus Girard. (Including

variety sanborni (Faxon)).
Fig. 3. Distribution of Oambarus propinquus Girard, propinquus sanborni (Faxon), and C.
obscurus Hagen.
(For further explanation, see legend on map, and text, p. 433-446).

ORTMANN : THE CRAWFISHES OF THE STATE OF PENNSYLVANIA 523

Plate XLIII.
Map of Pennsylvania, showing distributional areas of crawfishes. (See Legend on
map, and text, p. 46 ~> — 1<>6).

CONTENTS.

Paoe.

I. Introduction : SCOPE of Work ; Methods ; Acknowledgments 343

II. Historical Review of oub Systematic Knowledge of the Crawfishes of

Pennsylvania 348

III. Morphology and Chorolocjv of the Pennsylvania Species 350

A. General Remarks 350

B. Key for the Pennsylvania species 351

C. Description and Distribution of the species 352

1. Cambarus (Faxonius) limosus (Raf.) 352

2. Cambarus (Faxonius) propinquus Gir 358

2a. Cambarus (Faxonius) propinquus sanhorni (Fax.) (extralimital) 365

3. Cambarus (Faxonius) obscurus Hag 369

4. Cambarus (Bartonius) bartoni (F.) 377

4(i. Cambarus (Bartonius) bartoni robustus (Gir.) 388

5. Cambarus (Bartonius) carolinus Er 394

6. Cambarus (Bartonius) mouongalensis Ortm 398

7. Cambarus (Bartonius) diogenes Gir 402

IV. Ecology and Geographical Distribution 410

A. Ecology 41°

1. The River-Species : C. limosus, propinquus, obscurus 410

2. The Mountain-Stream Species : C. bartoni 413

3. The Burrowing Species : C. carolinus, mouongalensis, diogenes 41(i

a. General Habitat 41b

b. Shape of Burrows 418

c. Construction of Burrows 419

B. Geographical Distribution '-•'

1. Cambarus limosus ' '-'•'

a. Summary of facts ' -•'

b. Origin of distribution 426

2. Cambarus propinquus, sanhorni, and obscurus 433

a. Summary of facts 433

b. Origin of distribution '" '

3. Cambarus bartoni 44 /

a. Summary of facts '

524 MEMOIRS OF THE CARNEGIE MUSEUM

Page.

b. Origin of distribution 447

4. Cambarus bartoni robustus 449

a. Summary of facts 449

b. Origin of distribution 449

5. Cambarus carolinus 451

a. Summary of facts 451

b. Origin of distribution 452

6. Cambarus monongalensis 453

a. Summary of facts 453

b. Origin of distribution 454

7. Cambarus diogenes 457

a. Summary of facts 457

b. Origin of distribution 459

8. Summary of studies on distribution 462

V. Life History 469

1. Cambarus obscurus 470

2. Cambarus propinquus and C. propinquus sanborni 476

3. Cambarus limosus 477

4. Cambarus diogenes 480

5. Cambarus bartoni and C. bartoni robustus 486

6. Cambarus carolinus and monongalensis 489

VI. Economic Value 492

1. Popular knowledge of crawfishes 492

2. The use of crawfishes as food and bait 493

3. Crawfishes as scavengers ; their food ; their enemies 494

4. Crawfishes as obnoxious creatures 496

VII. Bearing of the aboa^e Studies on the Theory of Evolution 497

1. The Mutation-Theory of De Vries 498

2. Species, Varieties, and Variations among the Pennsylvania Crawfishes 498

3. Formation of Species by Isolation, as exemplified by Pennsylvania Crawfishes. 504

Bibliography 513

Explanation of Plates 518

MUSEUM. V

BSCURUS Hi G1RARD.

4. CAMBAR'

MEMOIRS CF THE CARNEGIE MUSEUM.

J.MBARUS BARTONI II
\MBARUS BARTONI ROBUSTU

Memoirs Carnegie Museum, vol. 1 1.

Plate XXXIX.

Mobpeolooii it Details of Genus Cambanu.

Memoirs Carnegie Museum, vol. II.

Plate XL.

scale: SO om.

Cnfx.k of Cambarus, Figb. 1-7. Btjbbowb 01 Cambanu, Kigb. 8-9.

[emoirs Carnegie Museum, Vol. II.

PLATE XLI.

scale: SO cm.

BURBOWB OF ( 'iiiiihiii-us.

MEMOIRS OF THE CARNEGIE MUSEUM. VOL. a

Fig 1 Preglacial Monongakela Fig. 2- Range of C. obscurus&propin^uus

Fig. 3. Distribution of C propincyuus, propinffii"* .anborm, opscnrus