z

o

in

X

z

I ^ I i y i ^

j‘/>N0linIlJ.SN|ZNVIN0SHllWSWS3 I HVB 8 nZLI B RAR I ES^SMITHSONIAN INSTIT

(/) _ Z \ ^ Z ^ ay Z

c/J Jfe, 2 A \

o 2 5 2* y/*' - V

l^L! B RAR I ES^SMITHSONIAN^SNSTITUTION NOlinilJLSNI^NVINOSHIIWS^SB I 9'

^SS\ § ./ v A m /^kwi&s. c w r /0\

t f\

S %

rrs .... w _

NOliniUSNrNVINOSHilWS^Sa I HVH aifllB RAR 1 ES^SMiTHSONIAN INST!*
</* z g 5 .<v- g ___ ^ *N

2 .< S . **— S'

'5 /« 5 s -3Sn£

-f LIBRARIES SMITHSONIAN INSTITUTION NOlifUllSNi NVIN0SH1IWS S3 ! H

sK

c .^c7 y''MK 5 S

H VemP?oV

O X^AfOC^ __ N^UIJLV^ Q

NZ NOliniliSNS^NVINOSHlII/yS^SB I &VH 9 IT LIB RAR I ES^ SMITHSONIAN INST!
z r- » z ._ r* z

ro

33 /ii

'< > tel

2 \L

g m z m ___ ^ ^

1 L. I 8 RAR I ESVnSMITHSONIAN_!NSTITUTION' NOlifUllSNi “NVINOSH1I INS S3 I H

z •» <g __ Z , g

< s» 3 ><Svwc^N. «€ .a. 2

5 A 3 MW'Az*. z H

S: o

CO

X

Q£3

^ xjwjjx- ^ > ' s> X£v osy^X >

N“/N0liniliSN|ZNVIN0SHlllNs'/>S3 I aVdSn^LIB RAR I ES^SMITHSONIAN JNSTI
(/> — ”3* \ 2 JML, Z

o w 2 5 2 r 5 v>'

n"JLIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IINS S3 1 i
1 V z r- z r* ._, Z

03

DN

*v 2 s

s) > 1

5/ 33

P&mM&r&y n \iM4Mw — " vo;

m Ng.D-vr 2; w''' ^ ^

W NOIlfliliSNI- NVINOSHillNS^SS I aVaan~UBRAR I ES ^SMITHSONIAN JNST

m

b KAK I to v>(Vl( I noUi^jMIN IliO I I I U I Iwli IMUIXI (MWUHWOnAI^O » ova a 8

CO _ Z CO Z W

OlinillSM NVINOSHilWS S3IBV88n LIBRARIES SMITHSONIAN INSTITUTIO

“ <o _ — CO

IBRARIES SMITHSONIAN INSTITUTION NOJlfUIJLSM NVIN0SH1IWS S3IHV88I

SSS!3\ 2 . . _■£ /$2Sni&\ 2 XSiX 5

50
>

-

m ^ m . _

_ co' £ co X 5

JOliniUSNI NV1N0SH1IINS S3lUVd8l1 LIBRARIES SMITHSONIAN INSTITUTIC

Z £ 5 ,v<’' £ 2 ,v.

•1 I vyy-t I 1

s \aszy 1 1 Xf l

.1 B R AR I ES^SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHilWS^SS I « VB 9

<o X co . CO

°

JOliniliSNI^NVINOSHJLIINS^SS ! 8 VH 8 II B RAR I ES^SMITHSONIAN^INSTITUTIC
r- . z r5 z f”

ro '■ 2: rn _

.IBRARI ES^SMITHSONIAN~INSTITUTION%OlinillSM~NVINOSHJLIWS S3! HV88I
<0 ^ z co z CO

CO
O
z

> s *' \^V > ' 2 X^osv^vx >’

JOlinillSNI^NVINOSHlIWS^SB I HV8 8 !l\l B RAR S ES^SMITHSONIAN JNSTITUTIO

32ESSS. w V w v> xfSESx “ w

h mv,4 ~ h ~ Wm£%. h .

,^c

_ '' R

2 *J Z _

JBRARIES SMITHSONIAN INSTITUTION NOIifUIiSNI NVIN0SH1MS S3I8YB8
2 z r- z

lOIlfUIiSNI NVINOSHIIINS^SS I HV8 8 I1~L I B RAR I ES ^SMITHSONIAN INSTITUTE
z CO z g z

vX

5! J&'

,<ro7»? 3

, f Waffn,

I

ART. 1. NEW RECORDS OF FRESH WATER GASTROPODS
FROM THE BAHAMA ISLANDS

By Juan J. Parodiz
Associate Curator of Invertebrates
Carnegie Museum

In March 1956 Dr. M. Graham Netting and Neil D. Richmond of
Carnegie Museum spent two weeks collecting in the Bimini Islands*.
Incidental to their other work they made a small but interesting collection
of fresh- water snails. This collection contains five species, three of which
had not previously been reported from the Bahama Islands. Including these,
there are now 11 species of fresh-water snails known to occur in the Ba-
hamas. Although obviously an; impoverished fauna it is still remarkable in
view of the scarcity and impermanence of bodies of fresh water in these
islands.

The following list shows the known distribution of the 11 species:

Littoridina (Littoridinops) tenuipes (Couper). South Bimini, Grand Ba-
hama and Andros islands.

Drepanotrema lucidum (Pfeiffer). Grand Bahama and Cat islands.

Drepanotrema (Fossulorbis) cimex (Moricand). Grand Bahama and Cat
islands.

Drepanotrema ( Fossulorbis ) lanierianus (d’Orbigny). South Bimini (new
for Bahamas).

Planorbula albicans (Pfeiffer). Cat Island.

Tropicorbis stagnicola (Moricand). Grand Bahama.

Tropicorbis havanensis (Pfeiffer). Cat Island.

Physella (Phy sella ) cubensis (Pfeiffer). South Bimini, Grand Bahama, Cat,
Andros, New Providence, and Eleuthera islands.

Lymnaea cubensis (Pfeiffer). South Bimini (new for Bahamas).

Laevapex bahamensis (Clench). Grand Bahama.

Gundlachia bakeri Pilsbry. South Bimini (new for Bahamas).

With the exception of Laevapex bahamensis , all of the other species
have a wide distribution in the West Indies, and most of them occur also
in the Bermudas, Florida Keys and southern United States. Since the
Bahamas were submerged during Pliocene and early Pleistocene, the present
fauna is not the result of land connections, but of very new elements
which have arrived by accidental means agreeing with what Clench has
suggested for the land shell fauna. Glacial control during Pleistocene,
lowering the sea-level on the Great Bahamas Bank formed a single or
a few larger islands, where proximity with other near-by West Indies
favored dispersion.

♦During this time they were guests at the Lerner Marine Laboratory through
the courtesy of Dr. C. M. Breder, Jr., Chairman and Curator, Department of
Fishes and Aquatic Biology of the American Museum of Natural History.

1

Issued April 29, 1957. II MAY 28 1957 1

2

Annals of Carnegie Museum

vol. 35

The Bimini group of islands is on the western side of Great Bahama
Bank, on the edge of the Florida Strait. The highest point is about 30
feet above mean tide. The fresh-water shells herewith reported come from
a pond on the northeastern corner of South Bimini, near the shore, and
surrounded by a mangrove community. An ecological description, repro-
ducing a photograph of this pond, was given by R. A. Howard in 1950
(Ecological Monographs, v. 20, no. 4, p. 334). This body of water is semi-
permanent and, at the time when the shells were collected, following a
very dry season, no water remained except for a wet muddy area in the
lowest part of the basin. This wettest part of the pond was covered by
a dense stand of cattails ( Typha dominguensis). All of the shells collected
were recovered from a sample of the mud and sand from the bottom.
Associated with them were numerous fossil valves of Glycimeris, Lucina
and other bivalves weathered from the coquina rock that forms the sides
and bottom of the pond. The presence of Glycimeris ( Tureta ) arata (Conrad)
suggests that this coquina may be of Pliocene rather than Pleistocene age.
Glycimeris ( Glycimella ) decussata (Linnaeus) is also abundant.

Littoridina (Littoridinops) tenuipes (Couper)

Amnicola tenuipes Couper, in Haldeman, Monograph of the Limniadae
of North America, 1844, pt. 7, p. 23, plate 1, fig. 14-15 and last page
of cover.

Bythinella tenuipes , Binney, Land and fresh-water shells of North Amer-
ica, 1865, pt. 3, p. 69.

Hydrobia blacki Pilsbry, Proceedings Academy of Natural Sciences of
Philadelphia, 1930, v. 82, p. 301, plate 22, fig. 4-6.

Hydrobia tenuipes, Clench, Memorias Sociedad Cubana de Historia
Natural, 1938, v. 12, p. 314.

Littoridina (Littoridinops) tenuipes, Pilsbry, Nautilus, 1951, v. 66, no. 2,
p. 50.

Type locality: Hopeton, Georgia. Hydrobia blacki Pilsbry is from Lake
Forsyth, Andros Island, and Clench recorded it from Grand Bahama.
The species is also well distributed in eastern United States from New
York to Florida. This is the first record for South Bimini, where 700
specimens were obtained.

Drepanotrema (. Fossulorbis ) lanierianus (d’Orbigny)

Planorbis lanierianus d’Orbigny, Moluscos, in La Sagra, “Historia fisica,
politica y natural de la isla de Cuba,” 1845, v. 5, p. 104, plate 14, fig. 1-4.

Planorbis sumichrasti Crosse and Fischer, Journal de conchyliologie, 1879,
v. 27, p. 342. Crosse and Fischer, Ltudes sur les mollusques terrestres et
fluviatiles, in “Recherches zoologiques de l’Am^rique Centrale,” 1900, pt.
7, p. 69, plate 23, fig. 6-6d.

Drepanotrema (Fossulorbis) sumichrasti, F. C. Baker, The molluscan
family Planorbidae, 1945, p. 118.

Type locality: near Havana, Cuba. For P. sumichrasti, Cacoprieto,
Tehuantepec, Mexico.

Fourteen adult and young specimens. First record for Bahamas, and
not known from other Antilles except Cuba. Our specimens agree very

1957

Parodiz: Fresh* water Gastropods, Bahama Islands

3

well with d’Orbigny's figures and description; D. lanierianus has been
a rather neglected species and F. C. Baker did not mention it in his
monograph of 1945. It does not seem to the writer that sumichrasti can
be separated from lanierianus. Other species of Drepanotrema in the Ba-
hamas are D. cimex (Moricand) and-: D. lucidum (Pfeiffer).

Lymnaea cubensis Pfeiffer

Limnaeus cubensis Pfeiffer, Archiv fur Naturgeschichte, 1839, v. 5, p. 354.

Limnaea cubensis , Arango, Contribucion a la fauna malacoldgica cubana,
1879, p. 134.

Limnaea umbilicata C. B. Adams, Contributions to conchology, 1849, p. 45.

Lymnaea lecontii Lea, Proceedings of the Academy of Natural Sciences
of Philadelphia, 1864, p. 113.

Galba cubensis, F. C. Baker, Lymnaeidae of North America, 1911, p. 204,
plate 27, fig. 9-16.

Lymnaea (Nasonia) cubensis , F. C. Baker, Bulletin of the Wisconsin
Geological and Natural History Survey, v. 70, pt. 1, p. 264. Aguayo,
Memorias Sociedad Cubana de Historia Natural, v. 12, p. 274, plate 18,
fig. 6-9.

Lymnaea cubensis, Vigueras and Moreno, Memorias Sociedad Cubana de
Historia Natural, v. 12, p. 74.

A single specimen was obtained in South Bimini which measures 11.5
by 6.5 mm., a little larger than the average size of the species. The type
is from Cuba but it is widely distributed from Florida to Texas and northern
South America, west to Lower California and most of the Antilles. Not
recorded before for the Bahamas. F. C. Baker in 1928 created the sub-
genus Nasonia with this species as the type, but according to Hubendick
1951 (Recent Lymnaeidae, p. 117) the characteristics of the shell, as well
as the genitalia, on which this separation was based, are only of specific
value.

Physa (Physella) cubensis (Pfeiffer)

Physa cubensis Pfeiffer, Archiv fur Naturgeschichte, 1839, v. 1, p. 354.

Physa acuta d’Orbigny, (not Draparnaud 1805; not Sowerby 1840),
Moluscos, in La Sagra, “Historia fisica, politica y natural de la isla de
Cuba,” 1841, v. 1, p. 191.

Physa sowerby ana C. B. Adams, (not sowerby ana d’Orbigny 1841), Con-
tributions to conchology, 1849, no. 3, p. 45.

Physa cubensis, Arango, Fauna malacoldgica cubana, 1879, p. 135.

Physa orbignyi Maz£, Journal de conchyliologie, v. 31, p. 30.

Physa jamaicensis “Mouson,” Clessin, Conchylien cabinet, 1885, v. 1,
no. 17, p. 291, plate 42, fig. 7 (not jamaicensis C. B. Adams).

Physa guadeloupensis “Grateloup,” Clessin, Conchylien cabinet, 1885,
v. 1, no. 17, p. 291, plate 42, fig. 12.

Physa peninsulae Pilsbry, Nautilus, 1899, v. 13, p. 48.

Physa calaban Vanatta, Proceedings of the Academy of Natural Sciences
of Philadelphia, 1910, v. 62, p. 668.

Physa cubensis, Aguayo, Memorias Sociedad Cubana de Historia Natural,
v. 21, p. 267, plate 18, fig. 1-3, 1938.

4

Annals of Carnegie Museum

vol. 35

Physa cubensis. Clench, Memorias Sociedad Cubana de Historia Natural,
1936, v. 10, p. 339, plate 25, fig. 2. Bulletin Museum of Comparative Zoology,
v. 80, no. 14, p. 517. Revista Sociedad malacoldgica “Carlos de la Torre/'
v. 8, p. 103, 1952.

Of 140 specimens, adult and young, the average specimen is 5 by 3 mm.,
the largest 9 by 5 mm. It is common in all the Antilles and also Florida
and northern South America; probably it occurs in; all the Bahamas where
permanent or semi-permanent bodies of water permit its development.
This is the first record for South Bimini. According to Aguayo^ and Clench,
it is a very variable species, but our specimens, except for their sizes, are
very uniform.

Gundlachia bakeri Pilsbry (Fig. 1, A-H)

Gundlachia bakeri Pilsbry, Proceedings of the Academy of Natural Sciences
of Philadelphia, 1913, v. 65, p. 670, plate 26, fig. 1-3.

Original description: “In the second, or septate stage the shell is oblong,
the width contained twice in the length; sides slightly convex. The obtuse,
rounded summit is very close to the posterior end and overhangs, the right
margin. Back and left slope, are strongly convex; posterior and right slopes
short and steep. There are a few low, wave-like concentric wrinkles near
the embryonic shell and some faint radial lines on the anterior part. The
sides curve into the deck, except posteriorly where there is an angle or
keel at junction of upper surface and deck. The deck or septum is convex
and covers about two thirds of the lower face. The aperture is oval, its
margins elevated, blackish, and a little reflexed. Length 2.8, width 1.4,
alt. 1 mm."

The type was obtained in 1911 by Fred Baker in an artificial pond in
the city of Para, Brazil.

The descriptions and dimensions, as well as the figures of Gundlachia
bakeri agree perfectly with the 12 specimens collected at South Bimini.
Most of them represent the septate stage, ranging from 1 to 2.5 mm. long,
(Fig. 1, B and E). The complete septate stage is rather tubular inferiorly,
with a circular opening (Fig. 1, F). One specimen is of ancyloid form,
corresponding to that found by Pilsbry with the type “which may be a
dimorphic form, in which the septate stage is omitted". There were no
specimens in the type series, according to Pilsbry, of the third or complete
Gundlachia stage, here called post-septate, but some of the specimens of
the Bimini lot represent the initiation of this post-septate stage (Fig. 1, A),
and also the full development of this stage (Fig. 1, G and H); this supports
Dr. Dali’s (1904) opinion that “while retaining the shell of the first year
(of a septate stage), an enlarged and somewhat discrepant shell is secreted
during the second summer".

A very closely related species is Gundlachia textilis (Guppy) from the
Island of Trinidad, although it is more than twice as large. The shell
of textilis when not septate is (after Guppy), of approximate' shape between
Uncancylus concentricus (d’Orbigny) and Hebetancylus culicoides (d’Or-
bigny), but “small specimens are liable to develop a peculiar monstrosity
consisting of a plate joining more or less extensively the margin of the
aperture". By keeping them alive in aquaria Guppy (1870) found that

1957

Parodiz: Fresh water Gastropods, Bahama Islands

5

Fig. 1. Gundlachi bakeri Pilsbry. A, Initial stage of the post-septate form.
B, C, Septate stage, ventral and dorsal view of same specimen. D, E, another
septate individual, lateral and dorsal view. F, Complete septate stage. H, G,
Dorsal and ventral view of an individual of advanced third, or post-

septate stage

6

Annals of Carnegie Museum

vol. 35

septate and non-septate forms develop in the same species. These observa-
tions were confirmed by Nordenskjold (1903), and Dali (1904); the former
author had the opinion that some Gundlachia from the Chaco region in
South America are transitional stages in the development of the species
known as Hebetancylus moricandi (d’Orbigny).

Pilsbry (1913) admitted that Gundlachia is dimorphic and sometimes
without any septate stage, although different from the true Ancylus which
never develops a septum; he said, however, that the status of “Gundlachia
as a genus depends whether we consider as a taxonomic value the ability
to produce a septate form” and (1925), “the status of Hebetancylus whether
it assumes the Gundlachia shell form”. In the first place, the production
of a septum is, indeed, of taxonomic value, but not restricted to Gundlachia ,
being rather a characteristic of the subfamily, because other genera in
the Ferrissinae are septa producers. In the second case, the recent observa-
tions made by Aguayo (1946) and Hylton Scott (1953), proved that Hebetan-
cylus are necessarily non-septate forms of Gundlachia , and the systematics
of the group are still very obscure.

According to Aguayo, Gundlachia ancyliformis Pfeiffer, from Havana,
has different stages linked with Hebetancylus havanensis (Pfeiffer), and
Hylton Scott synonymized Gundlachia nordenskjoldi Pilsbry with Hebetan-
cylus moricandi (d’Orbigny). The last author made careful comparative
studies of the anatomy in both supposedly different species, and concluded
that the foot, which in ancyloid individuals is elongated, becomes smaller,
rounded, and reduced to an anterior position, and the posterior edge of
the mantle is expanded forward and finally forms a ring around the foot,
with the visceral sac extended dorsally. In such condition the mantle is
able| to secrete that portion of the shell which corresponds to the septum.
As the shells of Hebetancylus and Gundlachia differ by their, respectively,
radially striated and smooth apices, it is inferable that the passing from
one stage to another diminishes the taxonomic value of such sculpture.
Hannibal (1912) and other authors state that none of the North American
Ferrissia are known to develop a septum, but I have observed Gundlachia-
like forms, from Ohio and Indiana with radially striated apices. The
taxonomic arrangement of the Ancylidae is still very unsatisfactory.

REFERENCES

Adams, C. B.

1849. Catalogue of freshwater shells which inhabit Jamaica. Con
tributions to conchology, no. 3, p. 45.

Aguayo, C. G.

1938. Los moluscos fluviatiles cubanos, pt. 2. Memorias Sociedad
Cubana de Historia Natural, v. 12, no. 4, p. 253-276.

1946. Gundlachia y Hebetancylus en Cuba. Revista sociedad mala-
coldgica “Carlos de la Torre,” v. 4, no. 2, p. 55-62.

Arango y Molina, R.

1879. Contribucidn a la fauna malacologica cubana.

1957

Parodiz: Fresh-water Gastropods, Bahama Islands

7

82° 80° 78°

0

Fig. 2. Location of the South Bimini Island and the Cattail Pond

8

Annals of Carnegie Museum

vol. 35

Baker, F. C.

1911. Lymnaeidae of North America. Chicago Academy of Sciences,
Special publication no. 3.

1928. The freshwater mollusca of Wisconsin. Bulletin of the Wis-
consin Geological and Natural History Survey, v. 70, pt. 1,
p. 1-494.

1945. The molluscan family Planorbidae. University of Illinois Press.

Binney, W. C.

1865. Land and fresh-water shells of North America. Smithsonian
Miscellaneous Collection, no. 144.

Clench, W. J.

1936. The Physidae of the West Indies. Memorias Sociedad Cubana
de Historia Natural, v. 10, no. 5, p. 335-342.

1938. Origin of the land and freshwater mollusks of the Bahamas.

Bulletin of the Museum of Comparative Zoology, v. 80, no. 14,
p. 481-541.

1938. Land and freshwater mollusks of Grand Bahama and Abaco
islands. Memorias Sociedad Cubana de Historia Natural, v. 12,

no. 4, p. 303-333.

1952. Land and freshwater mollusks of Eleuthera Island, Bahama
islands. Revista Sociedad malacoldgica “Carlos de la Torre,”
v. 8, no. 3, p. 96-116.

Clessin, S.

1885. Conchylien cabinet, v. 1, no. 17, p. 291.

Dali, W. H.

1904. Gundlachia and Ancylus. Nautilus, v. 17, p. 97-98.

d’Orbigny, A. D.

1845. Moluscos, in La Sagra, “Historia fi'sica, politica y natural de
la isla de Cuba,” v. 5.

Fischer, P., and H. Crosse

1900. Etudes sur les mollusques terrestres et fluviatiles. Recherches
zoologiques de l’Am&rique Centrale et du Mexique, pt. 7, v. 2.

Guppy, R. J. L.

1870. Notes on some new forms of terrestrial and fluviatile mollusca
found in Trinidad. American Journal of Conchology, v. 6,
p. 306-311.

Haldeman, S. S.

1844. Monograph of the Limniadae of North America, pt. 7, p. 23,
and last page of cover.

Hannibal, H.

1912. Synopsis of recent and tertiary mollusks of the Californian
province. Proceedings of the Malacological Society of London,
v. 10, p. 112-211.

Howard, R. A.

1950. Vegetation of the Bimini islands group, Bahamas, B.W.I.
Ecological Monographs, v. 20, no. 4, p. 317-349.

1957

Parodiz: Fresh-water Gastropods, Bahama Islands

9

Hubendick, B.

1951. Recent Lymnaeidae. Kungliga Svenska Vetenskapsakademiens
handlingar, series 4, v. 3, no. 1, p. 1-223.

Hylton Scott, M. I.

1954. Notas sobre la morfologia de Gundlachia Pfeiffer. Physis, v.
20, no. 59, p. 467-473.

Lea, I.

1864. Description of five new species of Lymneae of North America.

Proceedings of the Academy of Natural Sciences of Phila-
delphia, p. 113.

Nordenskjold, E.

1903. Ueber die Trockenzeitanpassung eines Ancylus von Sudamerika.
Zoologische Anzeiger, v. 26, no. 574, p. 590-593.

Olson, A., and A. Harbison

1953. Pliocene mollusca of southern Florida. Academy of Natural
Sciences of Philadelphia, Monograph no. 8.

Pilsbry, H. A.

1899. Annotated list of land and fresh-water shells collected in the
vicinity of Miami. Nautilus, v. 13, no. 4, p. 43-48.

1913. Appendix on Gundlachia, in Fred Baker, “Land and fresh-
water mollusks of the Stanford expedition to Brazil.” Pro-
ceedings of the Academy of Natural Sciences of Philadelphia,
v. 65, p. 668-671.

1930. List of land and freshwater mollusks collected on Andros,
Bahamas. Proceedings of the Academy of Natural Sciences of
Philadelphia, v. 82, p. 297-302.

1951. Littoridina tenuipes (Couper). Nautilus, v. 66, no. 2, p. 50-54.

Vanatta, E. C.

1910. Bermuda shells. Proceedings of the Academy of Natural Sci-
ences of Philadelphia, v. 62, p. 664-668.

Vigueras, I., and A. Moreno

1938. Physa cubensis, nuevo hospedero intermediario de Fasciola
hepatica. Memorias Sociedad Cubana de Historia Natural,
v. 12, p. 74.

Wurtz, C. B.

1951. Catalogue of the Ancylidae of South and Central America
and the West Indies, with description of a new species. Nau-
tilus, v. 64, no. 4, p. 123-131.

ART. 2. A SACRED BUNDLE OF THE ARAPAHO INDIANS

By Marianne L. Stoller
Assistant in Ethnology, Carnegie Museum

A •> \

INTRODUCTION

Sacred bundles were a part of the religious and ceremonial regalia of
the Arapaho Indian religion. They usually contained objects empowered
with spiritual and ritual significance. One such bundle is in the Arapaho
collection of Carnegie Museum (Number 3179A/147). This collection,
acquired by the Museum from the Fred Harvey Company, was assembled
by Dr. George A. Dorsey, presumably in 1901 or 1902, when he was study-
ing the Arapaho Indians of Oklahoma. The purpose of this paper is to
describe the bundle. This will be accomplished by (1) general remarks on
sacred bundles and their place in the religion of the Plains Indians;
(2) a complete quotation from the catalogue of the collection compiled
by Dorsey; and (3) a description of the bundle and its contents in order
as they were unwrapped by the writer. In a concluding section the bundle
will be compared with one described by A. L. Kroeber, the only other
known description of an Arapaho bundle.

SACRED BUNDLES IN THE RELIGION OF THE PLAINS INDIANS

Sacred bundles were peculiar to the religion of many of the Plains
Indians of the United States, as well as to some of the tribes living along
the Mississippi River and the western Great Lakes. These bundles usually
contained a variety of objects believed to be endowed with spiritual powers
and to have originated in a supernatural experience. As such, the bundles
were regarded as extremely sacred (unlike the paraphernalia of the age
societies which was more or less secular). They were opened only at certain
religious ceremonies, and in a hallowed atmosphere.*

The Arapaho Indians did not have an extreme development of the
bundle complex as did, for example, the Pawnee and the Blackfoot Indi-
ans. Further, it appears that bundles were not regarded by the Arapaho
to be as sacred as their sacred pipe and wheel (Kroeber, p. 309-311). Neverthe-
less, the bundles were of considerable importance in Arapaho religion, and the
opening of them involved much ritual. These bundles were kept in the
tribe and passed from one generation to another in accordance with the
“keeper-owner” concept of ownership. Among some tribes, for example
the Blackfoot, the bundles could be sold or transferred by one owner
to another (Wissler, 1912, p. 272-278). Exact Arapaho customs as regards own-
ership are not known; Kroeber reported the bundle which he described to

^Present day Plains Indians still regard these bundles with sincere respect and
religious awe. Mr. Hassrick states that he has seen Indians refuse to be present
when a bundle has been opened in a museum. Their feeling and respect for
these sacred objects are very real.

11

Issued June 20, 1957.

12

Annals of Carnegie Museum

vol. 35

have been acquired through inheritance (Kroeber, p. 310). This was prob-
ably the normal method of passing on bundles, with the owner serving
as a keeper or custodian of a bundle for the other tribal members.

DORSEY’S INFORMATION ON THE BUNDLE

An account of this bundle was given by Dorsey with the collection as
it came to Carnegie Museum; apparently he simply recorded the informa-
tion given by his Indian informant. Some of this information is not too
clear, and there are some statements which appear inaccurate according
to the physical facts of the bundle. It is possible that the informant was
speaking of sacred bundles in general, or did not remember the contents
of this bundle too clearly.

Dorsey’s description of the bundles is quoted in full below, with in-
formation in brackets supplied by the present writer.

WAHSAWK (Bear’s Bag), or SACRED BUNDLE. The whole oblong bag is a
chain of mountains, with a man sitting on it, watching constantly, and at the same
time feeding the inhabitants of the earth. Any person, especially the owner of it,
stores away in it goods, provisions, valuable articles, for the future. Inside, next
to the designs, thus making it complete, are found good dried buffalo meat, elk
meat, thick pieces of fat, pemmican, mixed berries, strouding, calico, coffee,
sugar, etc. Since these Indians are poor and can not afford to put away such
things, this Wahsawk has no contents, excepting a Dog-soldier’s paint string.
There were some spruce needles in the bag, but an old man bought them some-
time ago, for his future use. This sacred bag, as also others, was made after the
buffalow [sic.] cow, which afterwards pitied him, appeared to the hunter several
times [in a vision.] This man and his wife had one sore-backed pony. The camp
moved quite a distance, which caused them to camp near the creek, which had
plenty of timber. During their camp a change of life took place with them. On
their return to the tribe, the man erected an Old Men’s Lodge and his wife
erected the Buffalo Women’s Lodge. Thus the Wahsawk bags were made with
society rules and regulations. After the men have reached the old men’s lodge,
they have walked to the top of the mountain, i.e., attained old age. At the bottom
of the “Image” on each side, are bear’s claws, typifying the person begging the
“god” for mercy and continuous blessing. The foundations of the image are
a leg bone of the bear, several bladder bags, a serpent's body in hide and several
bunches of human hair. Inside of the image there is found buffalo wallow hair.
The sticks are used for the braces of a dragging-cart for it [the bundle]. The
feather, loose in the bag, is from a whip-poor-will. In all the designs the beads and
porcupine quills are used to signify persons and animals. The sticks are from a bush
called “basanechi” (big garter-snake), not used for a dragging cart; but they are
all utilized for various purposes — canes, pokers, and inside braces for beds. The
feather of the whip-poor-will dramatizes praying all night, or continuous worship,
without fatigue. The porcupine loops denote tipis and furniture. On these designs
are represented the Four Old Men. Both of the coverings are from “yellow” calves, —
meaning bright days, prosperous years, blessings upon the people. The calf’s hide
next to the image is from an extraordinary [unusual, perhaps abnormal] calf,
thus making it charitable, generous and humane to the image and more natural
to the people. The buffalo wallow hairs represent society lodges; the image,
a creator and the earth.

Dorsey’s information will be amplified and perhaps clarified by the
following description and by comparing this bundle with the one described
by Kroeber.

1957 Stoller: A Sacred Bundle of the Arapaho Indians 13

A DESCRIPTION OF THE SACRED BUNDLE

The part$ of the bundle are described and illustrated in the order in
which they occurred when the bundle was unwrapped. The various cover-
ings or bags, and all of the thongs and ties had been thoroughly impreg-
nated with red ocher. This had also rubbed off on most of the contents
of the bags giving everything a reddish-brown color and perhaps con-
tributing to the pungent odor of the bundle.

The outer layers of the bundle. The second bag was loosely wrapped in
a piece of figured calico material and tied with a rawhide thong. The
material was a floral print in blue, black, and yellow. This in turn was
wrapped in a second piece of calico which partly protruded from the
bundle and was quite dirty. This material was a brown, blue, and white
plaid.

The outer covering was a rather roughly dressed deer or elk skin. It
was much worn and weathered. Few traces of red paint remained on the
exterior. This bag is shown in Fig. 1. It was roughly constructed, apparently
from as whole skin. The skin; was slit where a seam was necessary and then
sewn with large stitches of sinew. One pair of ties over the top remained,
and there was evidence of another pair. A long, thick thong tied the whole
together.

Complete dimensions of the bag were: length, about 26 1/2 inches; width,
14 inches; and height, 10 inches.

The second bag. The second bag contained the innermost bag. When
this most sacred part of the bundle was removed there was found at the
bottom of the second bag a large undecorated “envelope,” made of roughly
dressed hide, partially filled with dried wild plums ( Prunus sp.). The
envelope is shown in Fig. 2. It was rectangular, with a triangular flap
fastened by two short thongs. There were two sets, each of two thongs,
on the upper back of the bag, apparently to hold a suspension cord. The
envelope was 15 inches long, 12i/4 inches wide at the bottom, and IO14
at the top.

The second bag is shown in Fig. 3. It was lined with canvas or duck
around the sides. Stippling in the illustration indicates this canvas lining.
The sewing was done with large stitches of sinew.

The bag itself was made of well dressed buffalo hide (probably from
a young animal) and was quite worn. It had been torn and patched many
times, especially around the top where the ties were. Two sets of two
thongs each were placed below the edge, about five inches apart, on each
of the long sides. Apparently these ties were fastened first, and then the
long thongs, one at each end, were tied over them pulling the large ends
over the opening of the bag for further protection of the contents. These
thongs were then tied around the bag. A looped strap for carrying or
suspending the bag was fastened to the back.

The ends of the bag were decorated with a beaded strip of black and
white “pony” beads. The complete design is shown, reconstructed, at the
left in Fig. 3. Dorsey’s explanation stated that the beads, like the quills.

14

Annals of Carnegie Museum

vol. 35

Fig. 1. Exterior of sacred bundle

1957

Stoller: A Sacred Bundle of the Arapaho Indians

15

Fig. 2. Envelope containing dried food

16

Annals of Carnegie Museum

vol. 35

1957 Stoller: A Sacred Bundle of the Arapaho Indians 17

signified persons or animals. The design may represent the Four Old Men
of Arapaho mythology.

This bag was approximately 25 inches long, 12 14 inches wide, and 9 14
inches high when filled with its contents.

The inner bag and its contents. The contents of this bundle were the
most important of the whole bundle; that is, the remainder consisted
mostly of coverings. Although the entire contents of the innermost bundle
were completely removed for examination, the bundle or bag itself— folded
and laced together with heavy thongs— was not completely unfolded. Fig. 4
shows its appearance and gives some idea of the way in which it was
constructed. It would seem that the piece of buffalo calf skin was first
doubled, and then at either end the separate folds were folded inward
and laced with heavy thongs. The method of folding the skin explains
why the fur appears both on the inside and outside (shown in the illus-
tration by hatching). Since the bag was made of buffalo skin it was quite
bulky and fairly heavy. It measured 21 inches in length and nine inches
in width (at its widest part) and was approximately six inches high.

The thong, shown partly under the bag in the figure, was loosely wrapped
around it. The length of the looped thong was approximately 31 inches,
while it was five-eighths of an inch wide.

In the bottom of the bag, partially buried by the folds, was a polished
stick placed lengthwise; a second such stick was loose inside the bag. It
is shown in Fig. 5, at A. The ends are carved smooth, and the whole is
polished and has a slight sheen. One end was slightly larger than the
other. The stick illustrated was 21 inches long, and had a diameter of
one-half inch at the largest end. Dorsey reported that the sticks were
from a bush called “basanechi” (big garter-snake). He first noted that the
sticks were used for a dragging cart for the bundle (a kind of travois),
but, further on, his information was contradictory for he said that they
were not used for a dragging cart but for canes, pokers, bed braces, etc.
In Kroeber’s description of a sacred bag he mentioned “four small sticks
of unequal length, which are leaned up together like tentpoles, and
which represent a tent” (Kroeber, p. 311). This may be some form of an
altar. He also referred to a fifth stick which was kept inside the case and
“represents the gift of a horse.” This bag had a total of five sticks, two
of which were incorporated in the base of the “image” and one in the
vertical part of it (See below). They were several inches shorter than the
two described above.

A knotted thong was lying in a heap in the bottom of the bag. A por-
tion of it is shown at B in Fig. 5; it was 53 inches long. Like all the
other articles of hide, it had been covered with a reddish paint, presumably
red ocher. It was a dog-soldier’s paint string, according to Dorsey, and
contained 89 knots. The Dog Lodge is the fourth of the Arapaho age
societies (open to men around fifty years of age), and, according to Kroeber,
painting was an important part of the initiation ceremony (Kroeber,
p. 199-200).

18

Annals of Carnegie Museum

vol. 35

Fig. 4. The inner bag

1957

Stoller: A Sacred Bundle of the Arapaho Indians

19

Fig. 5. A, One of four sticks of various lengths. B, Knotted thong.

C, Piece of skin

20

Annals of Carnegie Museum

vol. 35

Every morning each dancer, accompanied by his wife, repairs to his grandfather.
His grandfather has a string with a number of knots tied in it, representing the
number of times he was painted when he went through the ceremony. This number
is often about forty, but sometimes runs as high as ninety. Too great a number
is thought to be undesirable. Every time the grandfather paints the dancer he
unties one knot in his string. At the same time the younger man makes a knot
in a string which he brings. At the end of the painting the grandfather has thus
untied all the knots in his string, and his grandson’s string now contains the
same number, which, when he in turn becomes the grandfather of a later dancer,
will be untied in the same way.

A fragile, dried piece of skin was folded in the bottom of the bag. It is
shown at C in Fig. 5, and was identified as pigskin or peccary skin.

Bits and pieces of fibers and a seed were embedded in the fur of the
bag. The remnants of two feathers— mostly the quills— were also in the
bottom. A third feather was found between the two folds of one side
of the bag. This apparently was the feather referred to by Dorsey as being
from a whip-poor-will and represented continuous worship. The feather
is sketched at B in Fig. 6. It was b]/2 inches long. It was reddish in color
as a result of some of the bag’s paint rubbing on it.

Together with the items discussed above were found three bladder bags.
These bags were in poor repair. The bottom bag contained no opening
except for a few torn holes. It was empty but may once have contained
the spruce needles referred to by Dorsey. This bag was about 12 inches long.
One of the other bags is shown at A in Fig. 6. It was filled with porcupine

quills, mostly white ones, but a few were dyed yellow or red. The bag

was split lengthwise and laced with sinew at both ends. The third bag

contained mostly yellow quills with a few white ones; it was made like

the one just described but was laced with commercial thread. The two
bags were of equal length, 10 i/2 inches. The quills had not been flattened
as was the custom of the Arapaho women in preparing them for use
in decoration.

In the top of this bag, loosely tied in by the aforementioned thong, was
the “image.” It is shown in Fig. 7 (with the base slightly foreshortened).
The “image” was not entirely disassembled in this examination, since it
was so thoroughly bound and fastened together. Dorsey mentioned a
serpent’s skin, but this was not in evidence in the “image”; possibly the
skin in question was the one described above and found loose in the bag.

The foundation, or cross-bar of this inverted T-shaped object, consisted
of several things all bound together with thongs wound around them.
There were two polished sticks, one of which protruded at either end.
These sticks were surrounded by bladder bags. Another stiff element ran
lengthwise of the base, but it was wrapped with quill work and was not
otherwise visible. It might have been the bear’s leg bone referred to by
Dorsey; it was about the same size as the sticks, with diameters not ex-
ceeding one-half inch. Large and small loops of white porcupine quills
were incorporated in the base. At one end was a doubled piece of hide,
partly protruding, the edges of which were stitched with quills. Wisps of
hair (Dorsey said it was human hair and this seems likely) stuck out in
various places.

1957

Stoller: A Sacred Bundle of the Arapaho Indians

21

Fig. 6. A, Bladder bag with porcupine quills. B, One of three feathers;
this one separate from others

Annals of Carnegie Museum

vol. 35

1957 Stoller: A Sacred Bundle of the Arapaho Indians 23

The vertical part of the “image” was encased in a cone-shaped, thick,
and well dressed piece of hide laced up one side. It was held, none too
securely, to the foundation by a few thongs which passed around and
were incorporated in the base. At the bottom of this piece were several
vertical elements wedged between the horizontal ones of the base. Among
these were three fairly short objects wrapped in quills and narrow thongs.
These were the bear claws referred to by Dorsey except that instead of
two there were three, and they were in the middle and not at the ends
of the “image.” The fifth polished stick stood inside the case; it was the
shortest of all, being only 7 14 inches long. The remainder of the contents
of this vertical section consisted only of buffalo hair stuffed inside. The
dimensions of the “image” were, length, 15 inches; height, 10i/2 inches;
height of the vertical piece only, approximately 8 inches; approximate
diameter of the bottom on the vertical piece, 4i/2 inches.

Dorsey gave the symbolism of the various parts of the “image.” The
bear claws represented a person begging the “god” for mercy and blessing.
The quills were supposed to represent persons and animals, while the
quilled loops stood for tipis and furniture; designs on them were supposed
to represent the “Four Old Men,” but since they were devoid of design
perhaps this remark referred to the beadwork on the outer cover. The
buffalo hairs represented society lodges (or degrees), while the “image”
itself represented a creator and the earth.

A COMPARISON WITH THE BUNDLE DESCRIBED BY KROEBER

Krober’s remarks on a sacred bundle are here quoted in full (Kroeber,
p. 310-311).

There are a number of sacred bags in the tribe, and one was obtained, through
inheritance, by a young woman shortly before the summer of 1899. In Plate LXIII
the bag is shown unpacked, and its contents in the position in which they are
set in the tent. The upright cylindrical leather object is a person, apparently the
spirit of the bag. Food and gifts were given to the bag on the occasions on which
it was opened. The food, or part of it, was then kept in the bag. The object
of these gifts was to secure prosperity and an abundance of food to the giver.

The disposition of the contents of the bag during the ceremony is as follows: —

The owner of the bag sits at the rear of the tent. In the middle is a small
fire of coals. Directly before the owner is the outer bag, now holding only the
various gifts of food contained in small bags and pieces of cloth. The outer bag
itself is still covered, to a large extent, with the calico with which it is ordinarily
kept wrapped, and which presumably has been given to it. The smaller longi-
tudinal bag is removed from the inside of the large bag, and put in front of it,
nearer the fire. In front of this are placed grizzly-bear claws. To the right of
these claws is a small bag of incense called “man,” and to the left a similar bag
called “woman.” Immediately in front of the grizzly-bear claws are four small
sticks of unequal length, which are leaned up together like tentpoles, and which
represent a tent. In front of these sticks, and not far from the fire, is the cylindri-
cal leather case representing the “owner” of the bag. By the side of this object
and the four sticks, another stick, ordinarily kept inside of the cylindrical case,
is laid on the ground. This represents the gift of a horse.

This is all the exact information Kroeber gave on the subject of sacred
bundles. He did not specify on what occasions the bundle was opened.

24

Annals of Carnegie Museum

vol. 35

whose tent was used, who opened the bundle, and who could or was
expected to give it gifts.

There is little reason to doubt that the bundle described by Kroeber
and the one described in this paper are of the same type. Some details
differ, however. The bundle in Carnegie Museum had three cases or
bags instead of two; this bundle contained no bags of incense unless they
were embedded in the foundation of the “image" (which was not taken
apart); the “image” in Kroeber’s bundle did not have a foundation or
base attached as this one did; the feather, bladder cases of quills, and
knotted thong were unique to this bundle. Perhaps they had been gifts to it.

Kroeber’s description made clear some of the remarks in Dorsey’s cata-
logue notes. For instance, Kroeber explained some of the contents of the
bundle as being gifts to it, apparently as a form of supplication or offering.
The “image” represented a person and the spirit of the bag. This is a
less esoteric explanation than Dorsey’s that it was “a creator and the
earth.” Kroeber’s remarks suggested something of the ceremony involved
in opening such a bundle. Apparently only the owner could open it; there
was an established order for removing the contents, and they were placed
in pre-determined positions. Each object had symbolic significance. Dorsey’s
information was concerned mostly with this symbolism.

Dorsey’s references to the Four Old Men were apparently from the lore
surrounding the Sun Dance; neither Kroeber’s account of the bundle nor
his discussions of Arapaho religion and ceremonies contained such an
allusion. The Sun Dance was one of the biggest and most important cere-
monies in Arapaho life. Dorsey identified the Four Old Men as “the gods
of the four world quarters . . . the Four Old Men, are considered as ever-
present, ever-watching sentinels, always alert to guard the people from
harm and injury” (Dorsey, p. 14). It is possible that this bundle may have
been used in the Sun Dance ceremonies.

ACKNOWLEDGMENTS

The author is grateful to Carnegie Museum, and especially to James L.
Swauger, Assistant Director, for permission to describe this bundle. She
would also like to acknowledge the assistance of Mr. Swauger and Miss
Kay Reichert in reading and criticizing the manuscript. Royal Hassrick,
of the Denver Art Museum, also read the paper and gave it the benefit
of his wide knowledge of the Plains Indians. Dr. J. K. Doutt, Curator of
Mammals, identified all the skins in the bundle, and Dr. L. K. Henry and
Mrs. Dorothy L. Pearth of the Section of Plants identified the plums in
the second bag.

J957

Stoller: A Sacred Bundle of the Arapaho Indians

25

REFERENCES

Dorsey, George A.

1903. The Arapaho sun dance; the ceremony of the offerings lodge.

Field Columbian Museum Anthropological Series, v. 4, 228 p.
Chicago.

Kroeber, A. L.

1907. The Arapaho. Bulletin of the American Museum of Natural
History, v. 18, pt. 4, p. 1-229, 279-454. New York.

Lowie, Robert H.

1916. Plains Indian age-societies; historical and comparative sum-
mary. In “Societies of the Plains Indians,” edited by Clark
Wissler. American Museum of Natural History Anthropo-
logical Papers, v. 11, p. 877-984. New York.

Wissler, Clark

1912. Ceremonial bundles of the Blackfoot Indians. American Mu-
seum of Natural History Anthropological Papers, v. 7, pt. 2,
p. 65-289. New York.

1920. North American Indians of the Plains. American Museum of
Natural History Handbook Series, no. 1, 172 p. New York.

m.-u

t5 HP\A

ART. 3. REMARKS ON THE BEHAVIOR OF THE SQUIRREL
TREEFROG, HYLA SQUIRELLA

By Coleman J. Goin and Olive B. Goin
Carnegie Museum and University of Florida

INTRODUCTION

Several years ago we became} interested in the day-by-day activities of the
local treefrogs and began keeping detailed notes on individuals observed
during a continuing study of our back-yard fauna. Although much of this
information has been published elsewhere (Goin and Goin, 1953; Goin,
1955) we have a residue of data on the non-breeding behavior of these
frogs which we believe should be made available. The present paper deals
with the habits and habitat of the Squirrel Treefrog, Hyla squirella Sonnini
and Latreille.

The observations on which this paper is based were made in our back
yard near Gainesville, Alachua County, Florida. The yard is a plot nearly
rectangular in shape and approximately a third of an acre in extent,
bordering on a typical North Florida mesic hammock. The most conspicu
ous and abundant trees are water oaks ( Quercus nigra), but in addition
there are black gums ( Nyssa sylvatica biflora), sweetgums (Liquidambar
styraciflua), hickories ( Carya sp.), blue beeches ( Carpinus caroliniana),
several live oaks ( Quercus- virginiana), one small holly ( Ilex sp.), a magnolia
sapling ( Magnolia grandiflora), and a large loblolly pine ( Pinus taeda). The
undergrowth in the yard has been cleared out but the leaf mold has been
left undisturbed. Bordering the yard on three sides, however, is a dense
thicket of seedling oaks, palmettos ( Sabal minor) and shrubs and vines.

Observations in previous years on Hyla squirella had indicated that these
frogs are present in considerable numbers in the yard during the fall,
winter and spring months. Further, we had observed that they are noc-
turnal, stirring with the approach of darkness and settling down just after
daylight. They show a strong tendency to return time after time to the
same spot to rest during the day. We also noted that when the temperature
drops below a certain level the frogs do not stir from their hiding places
at night. It was in order to determine, if possible, the temperature at
which activity ceases that we initiated an intensive study of the movements
of these frogs in the fall of 1953. Late in August we set six stakes about
four feet high across the eastern end of the lot. On top of each stake
we inverted an empty tin can. This row of stakes with their cans was
about fifty feet east of the house, parallel to and just at the edge of the
woods bordering the yard. From October 22, 1953, until May 31, 1954, these
cans were examined morning and evening for the presence of Hylas, as
were the undergrowth bordering the yard and various other habitats, both
natural and edificarian, in which we knew treefrogs were apt, to occur.

27

JWG 1 2 m

Issued July 2, 1957.

28

Annals of Carnegie Museum

vol. 35

FALL RETURN

In Florida Hyla squirella breeds in the spring and summer, utilizing
temporary ponds and flooded ditches as well as more permanent bodies
of water. Not all of the frogs leave the area around our house at this time,
for we hear the rain call throughout the summer months. This rain call
should not be confused with the call given in the breeding choruses; in
warm, humid weather it is uttered sporadically by resting individuals at
any time during the day. Apparently the frogs that remain in the yard
during the summer move high up in the tops of the oak trees for we
almost never see them.

The summer of 1953 was one of the wettest ever recorded for northern
Florida. The yard was flooded much of the time during late August and
early September and although we observed no breeding activity, we did
record treefrogs more frequently than normal for this time of year. Usually
the reappearance of adult frogs in the hammock occurs quite abruptly
toward the end of September.

In the fall, young squirella begin to invade the hammock in increasing
numbers. We never saw more than one immature a day from the time the
cans were put up until October 30 when we saw three. On October 31
we recorded 10. The young frogs seem to move into the hammock in
waves. For the next 13 days we saw an average of four a day, then on
November 14 we counted 15. There was apparently another wave on
November 21, when we recorded 10, and another on November 24 when
17 were seen. This was the greatest number observed on any one day during
the course of the study. On December 8, we recorded 16. Thereafter the
fluctuations in number from day to day seemed more to reflect variations
in temperature than to indicate fresh invasion.

NUMBERS IN YARD

Since the preceding summer had been an unusually favorable one for
breeding, the number of young squirella in the yard during the course of
thel study probably represents a high point in the population cycle for the
area. It is difficult to form an accurate estimate of the actual number
of frogs resident in the yard. It seems probable that not all of those which
were noted when one of the “waves” moved in, found suitable niches
and remained in the yard for the winter. The average number of young
seen per day for each month was as follows: October (last week), 2.38;
November, 5.97; December, 7.65; January, 4.81; February, 3.40; March, 3.29;
April, 2.23; May, 2.43.

This decrease from midwinter to late spring was not due simply to the
growth of the individual frogs to a size at which they were no longer
recorded as immatures. Even in April and May the young of the previous
summer are recognizable as not full grown. Although we madei no measure-
ments, our observations would tend to agree with those of Wright (1932,
p. 322-323) that the one year olds are smaller than the two year old
(and older?) ones. He lists 13 mm. as transformation size, 19-23 mm. as
one year old and 26 mm. -f- for twoi year olds.

1957 Goin and Goin: Behavior of the Squirrel Treefrog 29

The decline in number of young recorded in the spring was undoubtedly
due in part to mortality. We observed no actual instances of predation,
but on one occasion a Rough Green Snake ( Opheodrys aestivus) was found
on two consecutive days coiled on the branches of a young water oak in
the thicket just back of can no. 3. Several times Yellow Rat Snakes ( Elaphe
obsoleta quadrivittata), which we have previously reported (Goin; 1955,
p. 101) preying on Hylas, were found in the vicinity. At 9:15 on the evening
of February 4, a small Yellow Rat Snake was seen on a bar of the rack
for the fuel oil tank against the back of the house. The temperature was
then 59°. Two adult H. squirella and one H. cinerea had been on the
bar during the day. Now one squirella remained on a leg of the rack below
the Elaphe. At 11:00 P.M. the snake was still there and the frog jumped
to the ground. The next morning, at 9:00 o’clock with the temperature
at 46°, the Elaphe wasi coiled on the bar, a squirella was sleeping on a
brace bar about an inch below the snake, and an Anolis, which had not
been present the night before, was resting about 6 inches farther along
on the same bar as the snake. At 10:30 the temperature had warmed up
to 56° and the snake was beginning to stir a little. By 11:45 it was crawling
slowly along the bar. The Anolis and squirella were still quiescent. Half
an hour later, the snake was where the lizard had been, the lizard was
gone, and the frog still slept on the brace bar below. The next time the
rack wa4 inspected the snake had, left.

Adult squirella were much less numerous than were the young. We
never saw more than three on any one day. The total number of records
for each month were as follows: October (last week only), 7; November, 22;
December, 10; January, 28; February, 44; March, 26; April, 0; May, 10.
Many of these records represent the same frog seen day after day.

DEPARTURE IN SPRING

Carr (1940, p. 61) records breeding choruses of squirella in Florida
from April 2 to August 20. No adult squirella were observed in the yard
during April. The 10 records listed for May represent what was probably
a single individual (possibly two). In marked contrast, the young were
nearly as numerous in April (61 records) and May (66 records) as they
were in March (78 records), and were still present when the study was
terminated in June. This suggests that yearling squirella do not leave
the winter habitat to move to the breeding ponds with the adults. During
the humid rainy season in the summer they move higher up into the
oak trees where they escape observation. Probably it is these yearlings that
are responsible for the rain calls heard in the non-breeding habitats in
the summer months.

SELECTION OF HABITAT-YOUNG

Table 1 gives the sites in the study area selected by immature squirella
for daytime resting places from October 22, 1953 to May 31, 1954. Of the
edificarian situations, the “rack” is the metal stand for the fuel oil tank
which is located against the back of the house. The “pipe” is an iron
pipe an inch in diameter, with one end angled upward to a height of

50

Annals of Carnegie Museum

vol. 35

fcj

ft*

5

O'

EC!

w

D

h

go

5 c/5

5<
. w

>i C/5

« k

. o

Q g

HQ
u w
w w
J pa
w a

So

$o

pa s

§g

Q Q

O

><

C*

<

120

27

42

O

o>

150

26

89

in

550

52

156

rO

00

00

338

00

00

00

ro

—

—

**

CD

—

99

O

66

ro

—

30

CO

40

O

00

ro

OJ

CD

ro

00

o>

m

CO

37

34

Is-

<vT

78

Is-

CD

26

CO

in

42

—

47

66

29

CD

sfr

00

25

CM

CM

—

105

CO

—

43

148

36

CM

CM

CO

47

44

CD

CO

176

—

in

in

00

CO

ro

62

00

ro

CO

20

CO

CO

25

-

rO

—

64

36

00

Is-

46

115

179

G)

—

O

in

—

CO

—

<T>

CD

Z

<

O

CAN 2

CAN 3

CAN 4

JO

z

<

Q

z

«£

O

RACK

PIPE

OTHER

TOTAL

EDiFSCARIAN

NATURAL

OAKS

HOLLY

MAGNOLIA

in

U.I

O

H-

1—

UJ

2

_J

i

OTHERS

TOTAL NATURAL

TOTAL

1957 Goin and Coin: Behavior of the Squirrel Treefrog 31

about 15 inches, lying on the ground behind can no. 3. Under natural
situations, the records for the oaks and palmettos represent a number
of stations in the edge of the thicket surrounding the yard. The holly,
on the other hand, is a single small tree, standing in a rather isolated
position in the yard about five feet west of can no. 2. The magnolia is
an isolated sapling about five feet tall. A number of these records repre-
sent a single frog returning day after day to the same resting place. On the
other hand, not all the records from one niche represent a single frog;
for example, at least four frogs are known to have been involved in the
records for cans no. 1 and no. 6.

It is apparent that there is a shift during the winter in the niches chosen
by young squirella. When they first move into the yard, they do not
select diurnal resting places that will serve throughout the winter. They
rather seem to occupy the first available site they find and are much
less apt to return to it the next morning. Apparently neither the young
oaks nor the palmettos provide niches sufficiently sheltered to be utilized
throughout the winter. All of the 52 records for young squirella on oaks
and all but four of the 81 records on palmettos were made during October,
November and December. Only the holly, and to a lesser extent the
magnolia, were utilized consistently until late spring when some of the
leaves begin to drop from these broad-leaved evergreens. The suitability
of the holly as a winter niche is, we believe, a reflection of the structure
of the leaves. They are leathery and tend to curl over at the edges.
A resting squirella clings to the under side of one of these leaves, fre-
quently choosing one that is lying on top of another leaf so that the
frog can find better protection on the holly than on oak or palmetto leaves.

Edificarian situations were preferred to the majority of natural ones
as permanent niches. During November, when the young frogs were just
moving into the yard, 115 were found in natural situations and 64 in
edificarian ones. In December there was a marked reversal with 62 frogs
found in natural situations and 176 in edificarian. During February and
March both types of habitats were utilized in approximately equal num-
bers, largely because of the availability of the holly. During May only
edificarian sites were occupied.

One noticeable phenomenon was the failure of the frogs to utilize can
no. 4 although they were observed in the other cans a total of 423 times.
We suspected at first that there might be something about the can itself
that repelled the frogs. On December 11 the can was exchanged for a; fresh
one and the old can placed on the outside kitchen window sill. The next
morning) a squirella was resting on the can on the sill. We suspect now
that the absence of frogs in no. 4 was due to its position in the middle
of the row. There was simply a much greater probability that a frog, as
it moved along the edge of the wood, would reach a can at the end of
the line and stop than that it would continue until it came to no. 4.

32

vol. 35

Annals of Carnegie Museum

SELECTION OF HABITAT-ADULTS
Table 2 shows the niches selected by adult squirella as diurnal resting
places. In contrast to the young, adult squirella are much more inclined
to select edificarian situations as soon as they arrive in the yard. Of 147
records, only five are from natural situations, four from the holly in
October and November and one from a palmetto in November. In further
contrast to the young, however, the adults did not utilize the cans to any
extent (four records during October and November). The most favored
site was the metal rack for the fuel oil tank. We recorded adult squirella
from this rack 94 times from October through March. The other 44 records
were all from situations around or close to the windows of the house— a
jar on the kitchen window sill, a crack between the screen and window
frame, the back porch, etc., so that all but nine of the records for adults
were of frogs on or about the house. Probably this preference of the adults
for niches close to the* house is correlated with availability of food supply.
The lighted windows attract insects in the evenings and the insects in
turn attract the frogs. When these frogs seek daytime resting places, they
simply go to the first sheltered niches they find. In marked contrast, only
31 of the 888 observations on immatures were made on or about the house.
This suggests that there may be a difference in diet between the adults
and young. Perhaps the latter feed more on ground-dwelling forms such
as ants rather than on the flying insects which are attracted to the lighted
windows.

REACTIONS TO TEMPERATURE

Under appropriate weather conditions, squirella leaves its place of diurnal
abode about dark and returns again about dawn the following morning.
It was in order to determine the effect of temperature on these nightly
wanderings that we began to make a check for the presence of Hylas after
dark as well as in the mornings, recording whether or not the frog had
stirred from its niche, and noting the temperature at the time the observa-
tion was made. We also recorded the low temperature for the diel
period and the temperature at which the niches were inspected the fol-
lowing morning. The temperature readings were taken from a Taylor
maximum-minimum thermometer fastened about five feet from the ground
on an oak tree in the middle of the yard.

A total of 974 observations is summarized by means of the bar diagrams
in Fig. 1. The discrepancy between this figure and the totals in the habitat
tables results from the fact that occasionally a frog was disturbed when
the niches were inspected in the morning or for some other reason left
during the day. 1 In each of these diagrams the vertical line represents the
mean, the horizontal line represents the range, and the bar represents one
standard deviation on each side of the mean to indicate dispersion. It can
be seen that, even though there is overlap, there is a pronounced tendency
for the frogs to remain in on cold nights and to forage on warm nights.
It can further be seen that the young and adults react in essentially the
same manner.

TABLE 2. SUMMARY OF DIURNAL NICHES SELECTED BY ADULT HYLA SQUIRELLA
DURING THE NON-BREEDING SEASON

1957

33

Goin and Goin: Behavior of the Squirrel Treefrog

94

44

142

o

o

o

24

CM

26

rO

rO

44

GO

o

28

CO

CM

o

CM

<D

r-

00

OJ

CD

CANS

RACK

OTHER

TOTAL

EDIFICARIAN

—

to

r^

O

O

26

44

00

CM

O

rO

—

22

—

h-

HOLLY

PALMETTO

TOTAL NATURAL

TOTAL

34

Annals of Carnegie Museum

vol. 35

Sometimes it happens that a drop in temperature causes the frogs to
return to their abodes well before daylight. This is to be expected since
in this part of Florida the lowest temperature of the diel cycle usually
occurs about 7:00 A.M. This means, of course, that it is possible for it
to be warm enough for the frogs to move about early in the evening
but to get too cold for activity long before morning. For example, on
March 14, at 10:45 P.M., when the temperature was 57°, three frogs that
had been out earlier were seen returning to their abodes. The temperature
the following morning was 42°. On another occasion one was observed
climbing the pole to a can at 9:50 P.M. when the temperature was 53°.

The frogs are still capable of activity when the temperature is in the
low fifties or high forties although they do not ordinarily move at such
temperatures. On the night of November 30, seven squirella which were
out when the niches were examined at 8:30 P.M., temperature 52°, had
returned by the next morning when the temperature was 53°. The over-
night low was 48°. Again the next night, four frogs returned between
temperatures of 48° and 51° with an overnight low of 40°. The lowest
temperature at which any of the frogs were observed to be active was 44°.
Twq immature squirella were seen moving into can no. 1 at this tempera-
ture at 10:30 on the night of January; 8.

Our observations suggest that the frogs may actively seek more sheltered
niches, perhaps burrowing under the leaf mold or under rotten bark,
with the approach of a cold front. On the morning of December 14, when
the low temperature for the preceding 24 hours had been 55°, 13 squirella
were recorded. That day the weather turned cold and windy, dropping to
a low of 33° the next morning. Three of the squirella had left by 3:00 P.M.;
when the niches were examined that evening, nine of the frogs had de-
parted although the temperature was down to 49°. One more had left the
following morning. High temperature for the 15th was only 52°, yet at
some time during the day two of the remaining three frogs moved from
the cans. These are temperatures at which squirella does not ordinarily
move out to feed. The weather warmed up somewhat on the 16th, with
a high temperature for the day of 59°. On the morning of the 17th, eight
squirella had returned to their normal niches. That day it turned very
cold again. High for the day was 50° and the low the next morning was
29°. By 3:00 P.M., two of the frogs had moved, and at 8:15, when the
temperature had dropped to 40°, only one frog was found. That one was
recorded as stirring; it had moved down to the bottom of can no. 5 from
the top and was still there the following morning. If this apparent tend-
ency of the frogs to move from their usual niches with the approach of
a cold front is valid, then the bar diagrams in Fig. 1 would include this
as well as feeding activity, and the overlap of the horizontal lines indi-
cating range of temperature is greater than would be the case if only
feeding activity were included.

1957

Coin and Goin: Behavior of the Squirrel Treefrog

55

36 Annals of Carnegie Museum vol. 35

REFERENCES

Carr, Archie

1940. A contribution to the herpetology of Florida. University of
Florida Publications, Biological Science Series, v. 3, no. 1, p.
1-118.

Goin, Coleman Jett, and Olive Bown Goin

1953. Temporal variation in a small community of amphibians and
reptiles. Ecology, v. 34, no. 2, p. 406-408.

Goin, Olive Bown

1955. World outside my door. 184 p. Macmillan, New York.

Wright, Albert Hazen

1932. Life-histories of the frogs of Okefinokee Swamp, Georgia.
497 p. Macmillan, New York.

ART. 4. DESIGNATION OF THE LECTOTYPES AND
LECTOALLOTYPES OF THE NEOTROPICAL TETRIGID
SPECIES IN CARNEGIE MUSEUM AS DESCRIBED BY
LAWRENCE BRUNER (ORTHOPTERA; ACRIDOIDEA)

By Harold J. Grant, Jr.

Assistant Curator of Entomology
Academy of Natural Sciences of Philadelphia

Early in this century, at the request of officials of Carnegie Museum,
Professor Lawrence Bruner of the University of Nebraska agreed to examine
and report upon a collection of neotropical tetrigids then acquired by that
institution. His subsequent research resulted in the description of 22 novelties
and the determination of many others (See Bruner, 1910, Annals of the
Carnegie Museum, v. 7, p. 89-143.) Most of the new species were repre-
sented by small series, but Bruner, following the procedure of many of
his contemporaries, neglected to select single specimens as holotype and
allotype of the species being described. Inasmuch as nearly all modern
systematists are in agreement in recognizing the necessity of designating single
types, I here propose to accomplish this for the species involved.

While some of the names involved are absolute synonyms, it is not my
purpose to expose them as such here, neither is this paper intended to be
revisionary in any way and the sequence of names used is that of Bruner’s
1910 paper. In all cases where it is possible, I attempted to select as lectotype
that specimen which most closely fits the description, type locality, measure-
ments, and date of collection mentioned by Bruner. Where any priority
is indicated (such as several dates being mentioned in the discussion follow-
ing the description) it has been utilized in selecting the lectotype. Where
it has been deemed appropriate, certain annotations regarding condition
of the specimens, size of the series, etc., are appended after the designation
of the single type.

The author wishes to thank the officials of Carnegie Museum for their
co-operation in publishing this account. Especially I wish to thank Dr.
George Wallace, Curator of Insects of that institution, for permitting me
to transport the series to Philadelphia to accomplish this work in my own
laboratory. Finally, thanks are due my colleague, James A. G. Rehn, for
his help and his many kindnesses in the course of the work.

Designation of the lectotypes and lectoallotypes.

Platytettix gibbinotus, p. 99-100.*

Described from two females and one male.

Lectotype designated: $ , Pard, August.j-

*The page number following the specific name refers to that upon which it appears
in Bruner’s 1910 paper.

fOnly the data on the labels are listed.

37

Issued July 2, 1957.

$8

Annals of Carnegie Museum

vol. 35

Lectoallotype designated: $ , same data as lectotype, but taken in July.
The lectotype bears a label in Dr. Hugo Kahl’s hand, “Returned by
Prof. Bruner July 17, 1914 in answer to remarks of Sept. 2, 1913, Prof.
Bruner writes: ‘A $ at hand but marked type, may be it, and not so
labeled by oversight,’ It should read above but not marked type. (Kahl).”
The specimen thought to be “apparently a nymph of the same species”
by Bruner (p. 100) is an adult female of another species ( Platytettix uni -
formis Bruner).

Platytettix uniformis , p. 100-101.

Described from three females and one male.

Lectotype designated: $ , Pard, July.

Lectoallotype designated: $, Benevides, July.

The lectoallotype is the “nymph” referred to by Bruner (p. 101). This speci-
men is actually an adult male (It lacks the median left limb and both an-
tennae).

Gladiotettix hancocki, p. 102-103.

Described from a single male which is now before me and in good
condition.

Sclerotettix abbreviatus , p. 105-106.

Described from two females and one male.

Lectotype designated: $ , Chapada, near Cuyabd, Matto Grosso, Brazil,
April.

Lectoallotype designated: $, Chapada, May.

Sclerotettix tibialis , p. 106-107.

Described from three specimens, apparently two males and one female,
of which the female is missing.

Lectotype designated: $ , Chapada, November.

The lectotype lacks the antennae, apical portion of the pronotum, and
the tarsi of all except the posterior limbs.

Sclerotettix variegatus , p. 107.

Described from an unknown number of specimens (apparently three
or four) of which two, a male and a female, remain.

Lectotype designated: $ , Chapada, April.

Lectoallotype designated: $ , Chapada, near Cuyabd, Matto Grosso, Brazil,

July-

Sclerotettix infuscatus , p. 107-109.

Described from three females and three males.

Lectotype; designated: $ , Corumbd, April.

Lectoalloype designated: $ , same data as lectotype.

The one specimen from Chapada (which I presume to be the female
reported, but which now lacks the entire abdomen) is labeled “April”
and not “June” as recorded by Bruner.

Sclerotettix minor, p. 109.

Described from a “single pair,” but three specimens are present, all
labeled with this name in Bruner’s own hand.

Lectotype designated: $ , Chapada, April.

Lectoallotype designated: $ , same data as lectotype.

The lectotype lacks the right middle and rear limbs, the left anterior
one, and one antenna.

1957 Grant: Neotropical Tetrigids in Carnegie Museum 39

Otumba basalis, p. 112-113.

The single male specimen upon which the species is based lacks the
antennae and all the limbs of the left side.

Allotettix fuscipennis, p. 115-116.

Described from three males.

Lectotype designated: $, Benevides, July.

Only the lectotype bears the word “type" on its label in Bruner's hand-
writing.

Allotettix chapadensis, p. 117-118.

Described from “two male specimens,” but I find a third in the series,
all of which bear the same data except that two were collected in January
and one in April.

Lectotype designated: $ , Chapada, near Cuyabd, Matto Grosso, Brazil,
January.

Tettigidea annulipes , p. 127-128.

The single female upon which the description was based is in good
condition.

Tettigidea gracilicornis, p. 128-129.

Described from five females and nine males.

Lectotype designated: $ , Para, Brazil, September.

Lectoallotype designated: $ , Chapada, August.

The designated single types are the only two of the series which bear the
word “type” upon their labels in Bruner’s handwriting.

The lectotype lacks an antenna, while the right rear limb of the lecto-
allotype is also missing.

Tettigidea chapadensis , p. 130-131.

Described from ten females and 21 males.

Lectotype designated: $, Corumbd, April.

Lectoallotype designated: $ , Chapada near Cuyaba, Matto Grosso, Brazil,
April.

Only the two specimens designated above bear Bruner’s notation “type”
on their labels.

Tettigidea hancocki, p. 131.

The single male upon which the description is based is in the present
series and in good condition except that the right rear leg is missing.
Tettigidea australis, p. 132.

The single specimen is in poor condition, lacking antennae and all
limbs except the first and third of the left side.

Tettigidea intermedia, p. 132-133.

Based on “one male and two female specimens,” but I find one male
and three females from the same locality in this series.

Lectotype designated: $ , Chapada, November.

Lectoallotype designated: $ , same data as lectotype except the date which
is April.

The two specimens designated are the only ones to bear a label with
“type” written upon it in Bruner’s handwriting.

vol. 35

40 Annals of Carnegie Museum

Tettigidea costulis, p. 133-134.

The single male described by Bruner is found to be in good condition.
Tettigidea subaptera, p. 134-135.

Described from nineteen females and fourteen males.

Lectotype designated: $ , Chapada, April.

Lectoallotype designated: $ , Chapada.

The designated types alone bear the word “type” on their labels.

Tettigidea arcuata , p. 135-136.

Described from sixteen females, eight! males and two nymphs.

Lectotype designated: $ , Chapada, April.

Lectoallotype designated: $ , same data as lectotype.

Both type specimens bear “type” on their labels and are the only two
so labeled.

Tettigidea corrugata, p. 136-137.

Described from “three females” of which only two remain.

Lectotype designated: $ , Chapada, near Cuyaba, Matto Grosso, Brazil,
October.

Both specimens bear the word “type” on their labels.

Lophotettix lineatus, p. 137-138.

Described from “four females” of which only three remain.

Lectotype designated: $ , Chapada, May.

The lectotype is in excellent condition, but the remainder of the series
is in very poor condition.

Apteropedon apicale , p. 140.

Described from two females.

Lectotype designated: $ , Rio de Janeiro, October.

The lectotype lacks the right antenna.

ART. 5. INDIVIDUAL AND GEOGRAPHIC VARIATION IN
BLARINA BREV1CAUDA FROM PENNSYLVANIA
By John E. Guilday

Assistant Curator of Comparative Anatomy
Carnegie Museum

Within the Commonwealth of Pennsylvania and in most of the north-
eastern states, the short-tailed shrew, Blarina brevicauda, is one of the
most abundant small mammals encountered by collectors. This semi-fossorial
insectivore is the cosmopolite of Pennsylvania mammals, being found in
practically every terrestrial habitat throughout the area. Local conditions
may affect the population density within a given area, but they present
no major barriers to genetic interchange. In effect, these animals comprise
one continuous interbreeding population throughout the area. This study
is an attempt to analyze the extent of variation present in a dense and
relatively stable population, as seen in Blarina brevicauda from Pennsyl-
vania.

As one of the outgrowths of the Pennsylvania Mammal Survey the
collection of Carnegie Museum contains slightly over 3000 specimens of
the short-tailed shrew, preserved as study skins accompanied by skulls
or as skeletal material without skins. Of this collection, 1193 specimens
were examined during the course of the study. The remainder of the
specimens were either scattered in insufficient numbers throughout the
state, or for one reason or another, primarily age and damage in trapping,
were not used. Of the state’s 67 counties, 23 were represented by a sufficient
amount of material to allow the application of limited statistical analysis.

Data on variation are drawn almost entirely from external and cranial
characters. Variation of pelage in Blarina is too subtle to permit any
objective study of it within the area.

I wish to extend my thanks to Dr. J. Kenneth Doutt, Curator of Mammals,
Carnegie Museum, not only for the unrestricted use of the mammal
collection and its records, but also for his patient and understanding advice
and criticism, without which this project would have suffered greatly.
I also wish to thank Miss Caroline A. Heppenstall, Assistant Curator of
Mammals, Dr. Dana P. Snyder, Research Mammalogist, C. J. Morrow, Jr.,
Staff Illustrator, and the many others who have aided me immeasurably.

The material on which this study is based is a part of the mammal
collection of Carnegie Museum. Most of the specimens were collected by
the Pennsylvania Mammal Survey— Pittman-Robertson Projects conducted
under the Federal Aid to Wildlife Restoration Act of 1937. Final reports
have been published on each project (Richmond and Rosland [sic], 1949;
Grimm and Roberts, 1950; Roslund, 1951; Gifford and Whitebread, 1951;
Grimm and Whitebread, 1952; Roberts and Early, 1952).

The study of subspeciation in any form of animal life can not be ap-

INTRODUCTION

VARIATION WITHIN A SINGLE POPULATION

Issued December 16, 1957

42

Annals of Carnegie Museum

vol. 35

proached intelligently without first gaining an idea of the extent of
individual variation within one interbreeding population.

This section is based on a study of 121 skulls of Blarina brevicauda col-
lected at East Springfield, Erie Co., Pa., from October 24, 1947 to November
15, 1947, by W. W. Goodpaster. They were selected because they com-
prised the largest group available from a single locality and, so far as
collecting techniques were concerned, the group was one of the most
homogeneous. The measuring and sexing of the collection in the field
was handled by one man. Specimens were all taken in “Museum special"
snap traps.

The purpose was to ascertain, as far as possible, the amount of variation
due to age, sex, and individual variation present in the population. Nine
of the specimens could not be aged because damaged by traps, and meas-
urements for these nine were included only in those instances where age
was not a problem. The field measurements taken were the standard-
total length, length of tail, hind foot, and ear measured to the nearest
millimeter. Measurements were taken by the collector with a pair of
calipers, read against a steel rule. The specimens were weighed to the
nearest half gram on Ohaus scales. Skull measurements (Fig. 1), recorded
to the nearest tenth of a millimeter, were taken by myself with the aid
of a binocular dissecting microscope and vernier calipers.

The skulls were aged by a system based on morphological changes in
the skull itself, especially degree of suture closure and relative development
of the sagittal and lambdoid crests. This is a modification of the methods
outlined in Doutt (1942). Tooth-wear was not used as a criterion of age
in this case because it was felt to be too easily influenced by environmental
effects, diet, habitat, etc. Specimens were found which had advanced tooth-
wear, yet which appeared younger than other specimens showing only
moderate wear on the teeth but a greater development of skull crests and
other signs of age.

Pearson (1945), aging samples of Blarina, used tooth-wear as the criterion
and obtained satisfactory results. Pruitt (1954) likewise has found tooth-wear a
satisfactory criterion for aging Sorex cinereus. It is my feeling that both
systems of arriving at a relative age (that is, tooth-wear and suture closure)
would work equally well when applied to a single sample. Plowever, in
comparing samples from widely scattered localities, a system based on a
single criterion, especially one so directly affected by the environment as
tooth-wear, may lead to the establishment of relative age scales which
would not be comparable between samples. Rate of suture closure, being
relatively independent of direct environmental influence, would seem to
be the better criterion of age when geographically scattered samples are
to be compared.

The general pattern of changes due to increasing age is as follows. In
the earliest stage examined, the lambdoid, sagittal and otic sutures appear
to be completely open. (As here used, the term “otic suture" designates
that suture which bounds the otic complex where it contacts the parietal
dorsally and the squamosal anteriorly.) The naso-frontal suture is in the
process of closing. The sagittal suture soon begins to close at its anterior

Guilday: Variation in Blarina from Pennsylvania

43

M5fyf2 M* P4 UNICUJ5PIDS I1
( — ^ \

i B9 1 1 ■ ra c

v j

Fig. 1. Diagrammatic views of skull and upper tooth row of Blarina
brevicauda showing measurements employed. A, dorsal view of skull;
B, posterior view of skull; C, semi-diagrammatic view, left side of upper

tooth row

Definition of measurements: 1. Total length of skull. Anterior surface of in-
cisors to most posterior surface of occipital condyles. 2. Cranial breadth. Greatest
transverse diameter of skull. 3. Interorbital width. Least diameter between orbits.

4. Maxillary breadth. Transverse diameter of skull across maxillary processes.

5. Rostral width. Greatest external diameter, measured across external nares.

6. Length of upper tooth row. From anterior surface of first incisor to posterior
surface of last molar. 7. Length of upper molariform tooth row. From anterior
surface of last premolar to posterior surface of last molar. 8. Width of foramen
magnum. Greatest transverse diameter. 9. Occipital height. Vertical distance from
lambda to external basioccipital fossa.

44

Annals of Carnegie Museum

vol. 35

end and forms almost immediately that part of the sagittal crest from which
M. levator labii superioris proprius (Arnback-Christie-Linde, 1907) originates.
At this stage the posterior half of the sagittal suture remains open and
the remainder of the crest undeveloped.

Next, the portion of the otic suture which runs from the zygomatic
process of the squamosal to the tympanic region on the ventral surface
of the skull begins to fuse. At about the same time the lambdoid suture
begins to fuse sporadically along its lateral aspects, and occasionally fusion
is also found at the lambda. The lambdoid and sagittal crests begin to form
by a bony thickening of the areas involved before the sutures fuse to any
extent. The sagittal crest continues to form while further closing of the
sagittal suture appears to be arrested temporarily. Fusion continues in
an erratic fashion on the lambdoid suture. The lambdoid crest continues
to develop not only by thickening at the suture but also by a bony exten-
sion which appears to grow anteriorly over the parietals for a short distance.
This is especially marked at the lambda, diminishing laterally, and it makes
the exact status of the degree of closure of the lambdoid suture hard to
ascertain in those cases in which it is pronounced. The otic suture, aside
from the seemingly early closure of its ventral one-third, closes erratically,
but lags slightly behind the lambdoid suture. The naso-frontal suture con-
tinues to close at a slow rate during this process and in the later stages
closure is complete. The last suture to close completely is the sagittal. While,
as noted above, its anterior portion closes very early, the posterior half
seems to be the last of the cranial sutures to coalesce.

During the later aspects of suture closure, the muscular attachments,
crests, ridges, etc., become more pronounced and the cranium takes on
a flat appearance. The rostrum appears to widen with age, this impression
being intensified by the rotation of the first incisors in older individuals.
In Fig. 2, A shows a lateral view of the three upper incisors as they appear
in a relatively young animal. The shaded parts of the teeth represent the
pigmented areas. From a dorsal view, Fig. 2, B, the first incisors can
readily be seen extending the profile out to a point. In Fig. 3, A, the
condition in old age is seen. The entire tooth (incisor 1) rotates down to
compensate for increasing wear and to maintain an efficient cutting edge.
As the rotation increases, the bone covering the anterior margin of the
root seemingly erodes back and the anterior portion of the root of the
first incisor is exposed. This condition has been recorded for Sorex by
Warren (1942) and is apparently characteristic of the family. From a dorsal
view. Fig. 3, B, it will be seen that the first incisors, having swung to a
nearly vertical position, no longer enter as conspicuously into the dorsal
profile of the skull. This abruptly truncated appearance tends to give
an illusion of width to the rostrum, which apparently (Table 1) does not
widen appreciably with age.

To establish a basis for classification by age, each suture was assigned
a number which coincided with its degree of closure. The sutures that
were used for determining relative age were the naso-frontal, sagittal,
lambdoid, and the otic. All other cranial sutures closed at a very early

Guilday: Variation in Blarina from Pennsylvania

45

Fig. 2. Lateral view (A) and dorsal view (B) of anterior portion of skull
of a relatively young Blarina brevicauda. Shaded portion indicates pig-
mentation

Fig. S. Lateral view (A) and dorsal view (B) of anterior portion of skull
of a relatively old Blarina brevicauda. Note that most of the pigmented
area has been worn away. Dashed line indicates the original boundary

of incisor

46

Annals of Carnegie Museum

vol. 35

age and were of no help in determination of age. Numbers were assigned
as follows: 1, open; 2, less than half-closed; 3, more than half-closed;
4, closed. Numbers were also given the sagittal and lambdoid crests,
based on their state of development: 1, not present; 2, forming; 3, present;
4, prominent.

The numbers assigned to the four sutures and two crests used for each
skull were then added together, and the resulting number was the assigned
age class of that animal. The skulls could then be arranged in 18 relative
age classes, having a theoretical range from 6 to 24. This was more than
sufficient to include all of the 112 specimens in the sample. The actual
range obtained was from 9 (youngest) to 22 (oldest). None of the age
classes showed any significant difference in sex ratio. This would seem to
indicate that the rate of suture closure and crest formation is the same
in both sexes, assuming that a given relative age class represents the same
chronological age in both sexes.

Aging by suture closure is subject to some error, and one is often in
doubt as to which number to assign a given suture; furthermore, the order
of suture closure is not constant in all cases, so an error of plus or minus
one age class could be expected. The skulls were therefore arbitrarily
lumped into groups having a range of two age classes. Age classes 17
through 22 were lumped into one group, because at 17 or beyond the
animals had obviously reached old age and further development was
practically nil; this afforded a group of sufficient size to be treated
statistically.

In Table 1 the means of the youngest and the oldest groups (sexes
combined) are compared. Length of upper molariform tooth row, length
of upper tooth row, and total length of skull increase up to age group 15-16
and then decrease with age, because of tooth-wear, so in these three dimen-
sions, comparisons were between the youngest and the oldest group in which
an increase was noted. A significant increase from youngest to oldest was
noted in all except three dimensions— tail length, interorbital breadth and
length of upper tooth row. However, in each case the mean of the younger
group was more than 90% of the mean of the older group.

Since the differences in size between the oldest and youngest specimens
were quite small, a further lumping of age classes seemed desirable. It
was found that the means of all the age classes for the measurements, weight
and maxillary breadth (two dimensions showing highly significant dif-
ferences between young and old) could be arranged in three major age
groups. Within each group, the difference between the means of component
age classes was not significant and between each two groups the difference
was either significant or bordering thereon.

Group Age classes

1: 9-10

2: 11-16

3: 17-22

In any comparison between samples collected at different times of the
year, age changes should be taken into account, since the age composition
of a population of Blarina varies markedly from season to season (Pearson,

Guilday: Variation in Blarina from Pennsylvania

47

O

§

hh

ca

2

o

u

CO

w

fa <M

C/3 lO

©

©

o o

©

CO

GO

©

cd

©

©

ifa

©

GO

©

GO

00

»d

©

©

©

©

©

©

©

©

©

©

©

9 •

b O,

o c
£ °
fa

ft!

fa .

r

fa

°o:

au

Cl, «

|2

co

*3

2w

I— I

QO

CO

CO

oo

©

©

©

CO

©

©

©

t-H

©

©

©

l—H

CM

CM

oo

CM

cd

+

od

i-3

od

CM

—I

CM

QJ

(U 73
U

ro

ft.

X

V

c

o

o

£ !>;

C".

©

©

CM

©

CM

CO

©

CM

c/5 -rf

©

r-3

©

©

©

©

©

©

0)

<D

©

s

QJ t

bC +
S3

+

+

+

"T"

+

+

4-

+

o

+

cm irs coto^coooar'it

CrH^t£)^^^^wi>(MC)oqC50-HOoqi')-.; i.-; <

tO^OOtOOOt'iftOOOMOO

^ t o i i^. /*/> — - _f

-H (M (M

fa O 2

fa

o©O)®a)co-H^>^O0ai!oaooo©oo(DN!O

nfa

POh

h

Wfa
w fa

og

c o

Hctf

< fa

c4

<

>

OJ©<M©<MO<£>OC'I©(M©<MO<M©<0©<£>0<M©

$F^©e^©ir^©ds©iA©iA©i>i©i>»©iri©»r3©£'^©

3

©

£

4h

73

b0

03

c

<u

©

C/5

4->

73

©

a

a

a

<u

lh

o

©

o3

CD

u

73

RS

©

©

Jh

<v

S3

<u

Sh

s ©

i

o3

*-

•5

73

03

©

<d

Lh

©

73

c3

o

o

4-*

Cl,

CU

3

a

03

S 2P

cu

C

1

1

bb

S3

<u

<U

Lh

©

*03

4->

©

a

©

LH

o

<4H

O

Lh

l2

©

1

4-

7s

Lh

o

03

73

©

©

©

ccS

O

h

bo

S3

fa

.2?

’5

&

Is

o

H

'2

o3

u

5h

CD

S3

HH

*x

o3

s

5h

C/5

o

bb

S3

<u

©

bb

r+

4J

J

'S

%

Occipital height 17-22 19 6.66 +0.14 2.25* 97.8%

9-10 15 6.52

* Probability between 5% and 1% level f Probability beyond 1% level
( The following abbreviations and symbols have been used in various tables: d —Mean $ minus mean $ .
M= Mean ± <rni. N= Number of specimens in sample. K— Range. V —Coefficient of variation . <j— Standard deviation.)

48

Annals of Carnegie Museum

voL- 35

1945). However, in Blarina brevicauda a detailed breakdown of age does
not appear necessary for studies of geographic variation, and produces
no significant refinement beyond the three major groups outlined above,
composed of relatively young, adult, and relatively old individuals.

There is a rapid growth rate in nestling Blarina (Hamilton, 1943; Pearson,
1945). Shrews attain essentially adult size before they become susceptible
to trapping, and hence the younger animals in collections are, in many
characters, indistinguishable from adults. However, in order to show the
maximum variation to be expected, the three age groups are combined
in computing coefficients of variation. It will be seen that the external
measurements, in most cases exhibit higher coefficients than do the cranial
measurements (Table 1).

Sexual differences in the population studied were not apparent visually,
but could be demonstrated statistically (Table 2). In every case the males
averaged slightly larger than the females.

A small, but significant, degree of sexual difference could be demon-
strated in weight. Weight showed the greatest increase with age, 1.67
grams from the youngest to oldest (Table 1) with a relatively high coeffi-
cient of variation, males, 10.25; females, 11.51. Weight, subject to such
factors as age, sex, reproductive status, daily fluctuations, diet, general
health of the animal, season, and time elapsed since capture, would
appear to be one of the least dependable of all measurements.

The ear measurement, although apparently unaffected by age or sex,
is also an unreliable measurement in Blarina. In 121 specimens the
observed range for this measurement was only two millimeters (6-8 mm.).
It is not an easy measurement to take with any degree of consistency, and
the error of measurement might easily equal the actual range of variation.
In other words, the unit of measurement, to the nearest millimeter, is
too coarse to demonstrate the true extent of variation present.

Length of tail shows no significant increase due to age (Table 1) and
a slight, but significant sexual difference (Table 2). However, since
individual variation is high (V = 6.78 in males; 5.29 in females) and could
easily be affected by injury, length of tail is not a particularly desirable
character for use in the study of geographic variation.

Of the five external measurements, total length had one of the lowest
coefficients of variation (males, 3.66; females, 4.73). It was influenced
only slightly by sex and showed a significant increase of 4.7 millimeters
due to growth. The fact that the measuring was done by only one person
in this case may have been responsible for the relatively low coefficient
of variation.

Length of hind foot would seem to be one of the most dependable
external measurements available. Compared with the other external
measurements it has a low coefficient of variation (males, 4.63; females,
4.61); it is not affected by age, and the sexual difference (Table 1) would
cancel out in comparing samples of approximately equal sex composition.
However, as brought out below, all external measurements may be subject
to bias and should be used with caution; particularly in geographic studies.

TABLE 2. VARIATION DUE TO SEX IN A SAMPLE OF BLARINA BREVICAUDA
FROM EAST SPRINGFIELD, ERIE CO., PA.

Guilday: Variation in Blarina from Pennsylvania

49

QJ

v U

■3*3

G

a

*

CM

*

TjH

©

*

©

©

o

r—

o

*

CM

©

«=

t"

00

r}H

rt-

©

t>:

©

CM

co

co

f-H

©

©

©

1-H

<-H

C/5

Of

05

05

CM

pH

©

©

GO

G 2

c^S ^

G

'G

GO

oo

©

CO

o

CM

CO

©

©

CM

©

o

©

©

©

©

©

©

©

©

©

« a

QJ

a

+

+

+

+

+

+

+

+

+

+

+

+

+

c*1

£j

CM

©

05

o

CO

00

05

I>»

05

©

Tt<

QO

oo

©

co

CO

©

©

CO

©

©

rn

CU

o

CO Tf

©

Th

05

©

CM

CM

CM

CM

CO

CM

CM

co

GO

CM

CM

CM

CM

©

©

OO

CM

(j , , .

© ©

©

©

©

©

©

p4

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

V t+H

fcg o

.2

>

+1 +1 +1

fl

+1

+1

+1

il

fl

+1

fl

+1

+1

+1

+1

fl

+1

+1

+1

+1

+1

+1

+1

+1

+1

+1

QJ

QO CO

00

05

CO

©

CM

©

00

o

©

o

co

©

CM

t"*

CO

CO

©

r-H

©

©

o

a

cq

CM

QO

QO

CM

©

th

00

05

00

©

©

©

©

©

CO

00

©

CM

05

U

>

CO ^

QO

©

©

»-4

CM

CM

CM

CM

CM

CO

CO

CO

TJH

CM

CM

CO

CM

©

©

CO

CM

pj b

53 S

CO Tj

a

« +1
^ s

Jh

QJ

-Q

a .g

s

£

QJ

to

a £

n

CO r-H

©

CM

©

©

©

©

Tfl

©

CO

CO

pH

CM

CM

CM

pH

r-H

CM

CM

pH

r-H

pH

pH

CM

CM

CO ©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

© ©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

CO

©

©

©

©

©

©

©

©

©

+1 +1 +1

+1

+1

+1

+1

+1

fl

+1

+1

+i

fl

fl

fl

fl

fl

fl

fl

fl

fl

fl

fl

fl

fl

fl

© ©

c^«

©

CO

CO

©

©

©

©

CM

CO

CO

OO

00

©

CO

©

©

CM

©

CO

CO

QO

QO

00

©

©

CO

CO

— H

CM

CM

CM

i— J

CM

CM

>-H

l—H

CM

r-H

©

tJh CO

CM

!>•

00

00

CO

00

©

©

CM

CM

CO

CO

GO

GO

CM

CM

CM

CM

CM

CM

© £—

CM

©

©

CM

CM

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

© ©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

CO

©

©

©

©

©

©

©

+1 +1 +1

+1

+1

+1

+1

fl

fl

+1

+1

+1

+1

fl

fl

fl

f!

fl

fl

fl

fl

fl

fl

fl

fl

fi

© ©

©

©

GO

t"-

OO

©

©

©

©

©

OO

CM

©

©

CM

©

pH

TjH

©

CO

©

CM 05

CO

©

CO

00

©

©

CM

©

05

05

00

00

oo

t"-

CM

CM

CM

QO

©

G ©

QO

©

CO

CO

©

CO

CO

CM

CM

©

©

G

c4

GO

CO

©

©

QO

QO

CO

OO

QO

QO

~ £

CM

CM

1—1

1—1

CM

CM

r— 1

©

©

©

©

CM

OO

CO

CM CM

©

©

©

©

CM

CM

©

CO

GO

CO

QO

Th

OO

QO

p4

p4

©

QO

l>-

©

CM

05

<-H r-H

CO

CM

i—l

CM

CM

CM

CM

1—1

f-H

QO

QO

00

GO

CO

CO

i— t

r—t

©

QO

co

CO

tf

QO

co ©

CM

CM

CO

CM

©

©

00

00

G

G

©

©

CM

05

CO

CM

©

CO

©

©

GO

©

CM

© ©

CM

CM

CO

CM

CM

rJ,

•-4

©

©

c4

CM

CM

©

©

©

©

CM

CM

QO

QO

r-H

S— 1

CM

CM

i— i

*— •

r-H

^OOt^QQ^W^OO^O+^O+fOOM^Ot^lOt^oOt^OOt^OO

0105

© © *0

'B

bb

to

G

QJ

T3

P3

QJ

Jh

40

Jh

rt

G ^

’V

£

£

G

bO

Cu

s

5

a

a s

s u<

J |o

mb ^

*53 ««

,0 5»

QJ

2 I

CL '£■}

V G

QJ I— I

o *

50

Annals of Carnegie Museum

vol. 35

Bias in external measurements

In order to demonstrate the extent of variation due to bias, samples
of Blarina brevicauda were taken from six localities in the northwestern
one-sixth of Pennsylvania.

Measurements of total length, length of tail, length of ear and of hind
foot were all taken in the field (using a steel millimeter rule) with the
exception of those from East Springfield, Erie Co., Pa. In this case
the collector used calipers, which were then read against a steel rule.

The following list groups the localities examined according to possible
variations in measuring technique.

Locality

External Skull Number Month

measurements measurements of of

taken by taken by specimens collection

A. East Springfield, Erie Co. W. W. Goodpaster

B. State game lands 154,

Erie Co.

C. Benson Swamp, 5 mi. E.

of Columbus, Warren }»C. L. Gifford
Co.

D. 1.5 mi. N. of Pittsfield,

Warren Co.

G. 2 mi. E. of Mars, Butler

Co.

H. 1.5 mi. N. of Darlington

Beaver Co.

}■

J.J. Christian

■J. E. Guilday

77

Oct.-Nov.

18

Oct.

19

Aug.

29

Sept.-Oct.

23

Sept.-Nov.

25

July-Aug.

In the graphs. Fig. 4, statistics of three external measurements are pre-
sented for each of the six localities.

The shrews from each locality were aged by the method outlined above
and only those which fell into age group 2 were used. These samples,
therefore, represent shrews of as nearly as possible a comparable age
group. Both males and females were included in the samples.

In all the graphs (total length, length of tail, and hind foot) the means
for the East Springfield group (locality A) are seen to be significantly
lower than the means of localities B, C and D, although the populations
are only some fifty miles apart. In each case locality A appears signifi-
cantly smaller. Localities B, C and D show no significant differences

between their respective means, but in length of tail and length of hind
foot appear to be significantly greater again than localities G and H.
These differences between the means, especially in length of tail (graph
2) and length of hind foot (graph 3) are associated with variation in meas-
uring techniques. The measurements from locality A were taken with
calipers and the means of the external measurements were, as a rule,
consistently lower than those of the other five localities examined, which
were measured directly against a steel rule. Localities B, C and D were

measured by a second collector, while localities G and H were measured

by a third collector. Skull measurements taken by a single individual

Guilday: Variation in Blarina from Pennsylvania

51

Fig. 4. Comparison of population samples of Blarina brevicauda from
northwestern one-sixth of Pennsylvania illustrating bias due to collectors’

techniques

52

Annals of Carnegie Museum

vol. S5

and involving only one technique do not show this pattern. This makes
it appear likely that the observed differences are the result of bias intro-
duced by differing measuring techniques. In this case the collection was
accompanied by sufficient data to detect the bias; however, in most cases
the magnitude of this significant source of error would not be known. In
view of this, external measurements of small mammals should be used
with care in comparing populations.

Cranial measurements

The cranial measurements were much more constant than were the
external measurements, the coefficients of variation in the former averaging
much lower. Several measurements, however, are subject to differences
which might affect comparisons of samples, and these should be kept
in mind when they are used.

Highly significant (probability 0.01) sexual differences occurred in the
following measurements— total length of skull, length of upper tooth row,
and occipital height (Table 2). Significant or highly significant age dif-
ferences occurred in the following measurements— total length of skull,
cranial breadth, maxillary breadth, rostral breadth, length of upper molari-
form row, width of foramen magnum and occipital height (Table 1). Even
though these differences are significant, they are slight and might not
affect conclusions concerning geographic variation among large samples.
However, with the small samples generally available, such differences
should be recognized in order to avoid erroneous interpretations as to
the taxonomic status of the populations represented.

GEOGRAPHIC VARIATION

In this survey, 1193 skulls of Blarina brevicauda, from 26 areas in Penn-
sylvania (Fig. 5) were aged and measured, using the techniques outlined
above. All immature, senile, and obviously abnormal individuals were
discarded. In order to have the samples as large as possible the sexes
were combined in all cases except weight. The differences are slight and
the sex ratio in the samples not markedly biased. Weight is treated for
each sex separately because it varies seasonally in response to the repro-
ductive cycle. In southeastern Pennsylvania four of the areas (designated
as V, Fig. 5) were lumped into one sample because of the small number
of specimens from each. Likewise, two areas (T, Fig. 5) in northeastern
Pennsylvania were combined to make one sample. The resulting data are
shown by graphs (Fig. 6-14) and in Tables 3-15.

All the graphs except that of weight agree in a common geographic
placing of localities. The graph for weight deviates from this plan, in
that the localities are placed in a seasonal sequence rather than a geo-
graphical one. Each graph is split into three divisions. Localities A through
J are all located in western Pennsylvania; localities K through P are lo-
cated in central Pennsylvania, and localities Q through V in the eastern
part of the state. As one progresses from left to right within each of the
three divisions one travels from north to south; that is, locality A is in
extreme northwestern Pennsylvania, locality J, extreme southwest, etc.

Guilday: Variation in Blarina from Pennsylvania

55

■§22l^3'S

lN.§zg<3«|

o -« .2 2 oi

25-s a a s- 3 s

"S o “ • « ~ U s S

^ w* -5^ ,oOM ^

2 W G <<0 o C_3 •« S

3.- *3 • u ^ r,3 -S

^ 0 ** O ^ fj* Q*

> O * &.S §2 o
^ § css ^.2

S 8 1 g .a:50! jj

,§|£ l^oeil^

^ r~ • 2* • ^

(U fr, O "* - W

8-u*“

T ^ S d 5 u -~

' . . on O »® R-v "tj

• iJ *”1 <£ O S

£l >^5 * D
v h > c ^

£11

» . § °._--§ .i’s
S1*« as® g §

= oo a d

3 ,-es 8

2 2 3 a ® oA’S I

?•§ a f k § 2 m b
; S £\® a s Ki . o

5 ^ O ■« v . *r3 12

« S O ?> . . *> *>

_ £ ? bd ^ ^o-a t?

ss'sari sli

1 «o|-_-ii|’S "u

ssaBaSsBjf

«a0

«g|02

w* £ O o

S»3

o
1 <J

I . <M r-j

' C 2 O . 2 £ o o

“’aw” S--u°

^a"- diaaa

i saaH •

JS feM

! CM

■ O O C/5

« . .P

o e s T3 £ -

.5 u rt o v. ^ cj

u -rl . « o Prf

Oh g ^ .g .8 ^ So «

~ ^*s ° cq :s o

rdd rf -S O

.o a* -0"

o <j

-* flJ CjJ ®W

3"So

GO

.*> ^ o ^ s;c]^ ^
r4 ho o v ^

^ J=s fl H 3 S

^ W 1^

54

Annals of Carnegie Museum

vol. 35

External measurements

Total length (Fig. 6; Table 3) is the most reliable of the external
measurements of Blarina brevicauda ; it is of large enough magnitude so
that it is not as readily influenced by bias as are the measurements of hind
foot, length of tail and length of ear. The graph, Fig. 6, shows a general
decrease in total length from north to south and from west to east. It
has been demonstrated (p. 50, 52) that the mean of the sample from area A
is low because of bias and it is possible that the same may be true of
areas F and M.

The two dimensions, length of tail and length of hind foot, did not
follow the trend shown by other measurements, possibly because of the
bias introduced by different collectors. The sample statistics are shown
in Fig. 7-8, and Tables 4-5.

A valid comparison of weights, particularly in this case, can not be
made because of variation introduced by such factors as reproductive
status, time of year collected, time of day collected, time elapsed since
death, sex and age. Within Pennsylvania, no significant correlation can
be obtained between weight and geography. Any geographic variation
present is obscured by the operation of extrinsic factors, which can not
be eliminated without a prohibitive amount of labor on the part of
the collector.

The weight of the female Blarina studied can, however, be correlated
with the time of the year the samples were collected (Table 6, Fig. 9).
The weights of the males, although fluctuating greatly, appeared to follow
no consistent pattern. Collections were compared which were made from
April through November. The weights of the females dropped from a
peak in April, May and June, at which time female weight equaled or
exceeded male weight, presumably in response to reproduction, to a low
from August through November, when female weight averaged less than
that of the males. There is one exception, area C (August-October) where
the females slightly exceeded the males in weight.

Evidently the influence of the reproductive cycles has a measurable
effect upon the weight of female Blarina , but produces no effect, or one
that is slight enough to be overshadowed by other effects, in the males.

In the light of the great influence exerted by collecting bias, external
measurements seemingly have slight value in a study of geographic varia-
tion in Blarina , unless the entire series studied is measured by one individual,
using one technique and the same instruments throughout, or unless
allowance can be made for this source of error. Variations due to differing
techniques of mensuration can, and in some cases do, overshadow any
true variation present.

Cranial measurements

The cranial measurements of the specimens examined appear to fall
into a definite pattern. An east-west cline is clearly demonstrable both
statistically and subjectively (Tables 7-15; Fig. 10-14). Blarina crania from
northwestern Pennsylvania were consistently larger in five of the nine
cranial measurements considered. These measurements exhibited some
degree of local, apparently random, variation, but when viewed in their

Guilday: Variation in Blarina from Pennsylvania

55

entirety they decrease in size towards the east and southeast. The crania
diminished in total length of skull, cranial breadth, maxillary breadth,
length of tooth row, and rostral breadth. The length of the upper molar
series did not follow the general trend but appeared to vary at random
throughout the area. The length of the unicuspid series, however, diminished
steadily toward the southeast. Width of foramen magnum, interorbital
breadth and cranial height showed no apparent clinal variation. In addi-

TABLE S. VARIATION IN TOTAL LENGTH OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 6)

Sta-

tion N

M

a

V

R

A

75

116.66+0.49

4.30

3.68

108-125

B

18

123.05+0.72

3.06

2.48

116-129

C

18

121.10+0.89

3.80

3.13

115-127

D

29

119.72+0.95

5.13

4.28

106-131

E

56

120.50+0.73

5.44

4.51

106-132

F

22

114.40+0.86

4.05

3.54

108-123

G

25

118.64+1.12

5.60

4.71

107-129

H

23

119.91+1.11

5.35

4.46

110-130

I

38

117.24+0.84

5.20

4.43

106-132

1

19

118.47+1.08

4.73

3.99

109-127

K

28

121.71+0.92

4.90

4.02

115-131

Sta-

tion

. N

M

a

V

R

L

48

119.29+0.76

5.25

4.40

110-130

M

22

111.81+1.24

5.83

5.21

102-121

N

15

119.66+0.50

1.95

1.63

110-132

O

13

117.30+1.00

3.61

3.08

112-123

P

18

118.50+1.36

5.76

4.86

110-130

Q

26

118.76+0.75

3.87

3.85

112-126

R

22

117.04+1.17

5.47

4.67

104-126

S

31

119.80+1.05

5.87

4.89

107-131

T

29

117.72+0.92

4.94

5.19

110-127

U

54

115.88+0.84

6.15

5.31

100-130

V

39

115.56+0.97

6.09

5.27

103-131

56

Annals of Carnegie Museum

vol. 55

tion, the crania in the eastern portion of the state appeared more delicate
in general appearance, less angular and rugged. The amount of pigment
on the teeth decreased in intensity and the percentage of occurrence of the
prominent hook-shaped maxillary processes became progressively less towards
the east.

Pennsylvania, according to the most recent taxonomic work on Blarina
(Bole and Moulthrop, 1942), falls within the range of Blarina brevicauda
kirtlandi Bole and Moulthrop, with intergradation in the northeastern
portion of the state approaching Blarina brevicauda talpoides (Gapper).
The latter is characterized as being a larger race of Blarina and if this

Fig. 7. Variation in length of tail of Blarina brevicauda

Fig. 8. Variation in hind foot of Blarina brevicauda

Guilday: Variation in Blarina from Pennsylvania

57

is true, any intergradation should be towards increasing size in the north-
eastern portion of the state. This is not the case. The animals become
smaller in the east, reaching the minimum size in southeastern Pennsyl-
vania. This would seem to cast some doubt on the present taxonomic
status of Blarina, so far as Pennsylvania is concerned. Possibly this condi-
tion occurs farther north and east, beyond the limits of the state (my
data furnish no evidence on this point), but not within the state itself.

From an inspection of the graphs for the various measurements (Fig.
10-14), non-clinal variation may be seen superimposed upon the clinal
pattern. Examination of a few scattered localities may, therefore, give
a misleading impression of the true picture. If these random variations
occur in such a ubiquitous form as Blarina which exhibits no pronounced
degree of ecological or geographical isolation within the area studied,
they may become a problem in the taxonomic interpretation of other
species which respond more readily to isolating mechanisms. This demon-
strates the necessity for adequate sampling in studies of geographical
variation. For example, a sample of 34 shrews from Sandy Creek, Venango
County, Pa., locality F on the graphs, appears to be significantly different

TABLE 4. VARIATION IN LENGTH OF TAIL OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 7)

Sta- Sta-

tion

N

M

<r

V

R

tion

N

M

or

V

R

A

77

25.74±0.19

1.70

6.60

22-30

L

48

27.16±0.26

1.82

6.70

24-31

B

18

28.50±:0.25

1.54

5.43

25-30

M

23

24.48±0.30

1.44

5.88

21-26

C

18

28.30rt0.35

1.50

5.30

26-31

N

15

26.66±0.51

1.97

7.36

24-31

D

28

28.10rb0.32

1.74

6.19

25-32

O

13

24.07±0.37

1.33

5.52

22-27

E

55

24.31 ±0.26

1.90

7.81

20-28

P

18

24.88±0.48

2.05

8.24

22-29

F

22

24.68±0.44

2.09

8.47

20-28

Q

25

27.20±0.34

1.74

6.39

25-32

G

24

25.12±0.52

2.58

10.27

19-29

R

22

24.54±0.30

1.41

5.74

23-28

H

22

25.86±0.47

2.22

8.58

21-30

S

31

27.32±0.40

2.23

8.16

22-31

I

38

26.31 ±0.35

2.22

8.43

20-30

T

29

26.58±0.44

2.38

8.95

22-30

I

19

27.21 ±0.35

1.52

5.59

24-30

U

54

23.88±0.28

2.05

8.58

19-30

K

28

26.71 ±0.47

2.49

9.32

18-32

V

39

23.48±0.31

1.96

8.35

18-26

TABLE 5. VARIATION IN HIND FOOT OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 8)

Sta- Sta-

tion

N

M

<7

V

R

tion

N

M

a

V

R

A

77

13.85±0.07

0.67

4.83

12-15

L

48

14.58±0.09

0.65

4.46

13-16

B

18

15.16±0.12

0.51

3.36

14-16

M

24

14.54±0.13

0.68

4.68

14-16

C

19

15.05±0.05

0.22

1.46

15-16

N

15

14.66±0.16

0.64

4.36

14-15

D

29

15.13±0.09

0.52

3.43

14-16

O

13

15.07±0.15

0.53

3.52

15-16

E

57

14.40±0.11

0.85

5.90

12-16

P

18

14.94±0.15

0.64

4.28

14-16

F

22

13.90±0.16

0.75

5.39

13-15

Q

26

14.88±0.15

0.79

5.30

14-17

G

25

14.68±0.15

0.73

4.90

14-16

R

22

14.40±0.16

0.77

5.35

13-15

H

22

14.54±0.13

0.65

4.47

14-16

S

31

15.38±0.16

0.92

5.98

14-17

I

38

14.39±0.04

0.26

1.80

13-15

T

29

15.07±0.09

0.48

3.18

14-16

J

19

14.84±0.12

0.54

3.64

14-16

U

56

14.91±0.07

0.53

3.55

14-16

K

28

14.57±0.12

0.65

4.46

13-15

V

39

14.85±0.07

0.47

3.16

14-16

58

Annals of Carnegie Museum

vol, 35

TABLE 6. VARIATION IN WEIGHT OF BLARINA BREVICAUDA FROM
VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 9)

Loc. Sex

N

M

cr

V

Range

Season

R

8

13

17.61+0.58

2.10+0.41

11.92+2.34

14.0-22.0

Apr.-May

2

9

17.33+1.08

3.25+0.76

18.75±4.42

12.0-21.0

P

8

8

17.69+1.11

3.15+0.79

17.81+4.45

13.5-20.5

May

2

10

18.25+0.79

2.51+0.56

13.75+3.07

15.5-23.5

N

8

7

15.85+1.08

2.88+0.76

18.17+4.86

12.0-21.0

May-June

2

8

17.87+1.04

2.94+0.73

16.45+4.11

13.0-22.0

O

8

4

17.00+0.81

1.27+0.45

7.47+2.64

15.5-18.5

June

2

9

16.93+0.47

1.34+

7.91 + 1.98

15.5-18.5

M

8

12

17.16+0.59

2.03+0.41

11.83+2.41

13.0-20.0

July- Aug.

2

12

16.75+0.59

2.04+0.42

12.18+2.49

14.0-22.0

D

8

14

19.21+0.42

1.57+0.30

8.17+1.54

17.0-22.0

Aug.

2

15

17.40+0.48

1.88+0.34

10.80+1.97

14.0-21.0

G

8

8

16.50+0.54

1.53+0.38

7.63±1.91

15.0-19.0

Aug.

2

10

15.96+0.40

1.26+0.28

7.89+1.76

13.5-18.0

L

8

27

15.70+0.37

1.93+0.26

12.29+1.67

12.0-21.0

Aug.-Sept.

2

20

15.25+0.37

1.67+0.26

10.95+1.73

12.0-19.0

S

8

14

16.93+0.40

1.56+0.28

9.21 + 1.68

14.5-20.5

Aug.-Sept.

2

16

15.46+0.41

1.66+0.29

10.74+1.90

13.0-18.5

c

8

11

17.09+0.47

1.57+0.33

9.19+1.96

15.0-20.0

Aug.-Oct.

2

11

16.44+0.71

2.14+0.50

13.02+3.07

13.0-21.0

J

8

9

16.77+0.36

1.07+0.25

6.38+1.50

15.0-18.0

Oct.

2

10

14.80+0.33

1.05+0.23

7.09+1.58

13.5-16.5

Q

8

12

17.33+0.33

1.16+0.24

6.69+1.36

15.0-21.0

Oct.

2

14

15.28+0.57

2.13+0.40

13.94+2.63

12.0-18.0

H

8

14

17.89+0.45

1.67+0.31

9.33+1.76

15.0-20.0

Sept.-Nov.

2

9

16.00+0.46

1.37+0.32

8.56+2.02

13.5-18.0

A

8

63

16.87+0.21

1.73+0.15

10.25+0.91

13.0-22.0

Oct.-Nov.

2

57

15.58+0.23

1.80+0.16

11.51 + 1.08

12.5-22.5

Guilday: Variation in Blarina from Pennsylvania

59

from the surrounding populations in many of the characters studied.
Groups such as this have at times been described as new taxonomic
forms by some workers. It is very improbable, if these differences are assumed
to be heritable, that they represent more than a transitory condition of
little evolutionary significance. They may, however, be a response to some
unrecognized local environmental factor which will continue to separate
this group even further from the original stock. Again, they may be non-
genetic in nature, a reflection of collecting bias, or inadequate sampling.
A formal recognition of these minor variations, even if they can readily
be separated from the surrounding populations, would seem to call for
extreme caution. Because of our inability to assign cause for effect, or

Fig. 10. Variation in total length of skull of Blarina brevicauda

TABLE 7. VARIATION IN TOTAL LENGTH OF SKULL OF BLARINA
BREVICAUDA FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 10)

Sta-

tion

N

M

( 7

V

R

Sta-

tion

N

M

a

V

R

A

73

23.44±0.06

0.57

2.43

21.8-25.0

L

49

23.01 ±0.06

0.46

1.99

22.0-24.1

B

18

23.54±0.07

0.33

1.40

23.0-24.2

M

20

23.16±0.13

0.60

2.59

22.5-23.7

C

19

23.31 ±0.11

0.50

2.14

22.5-24.4

N

14

23.20±0.24

0.92

3.96

22.5-24.0

D

28

23.40±G.08

0.47

2.00

22.4-24.4

O

13

23.32±0.16

0.59

2.53

22.7-24.3

E

57

23.37±0.07

0.53

2.27

22.4-24.6

P

18

23.02±0.19

0.81

3.52

22.2-24.3

F

20

22.77±0.10

0.44

1.93

22.2-23.7

Q

26

23.10±0.06

0.30

1.29

22.6-23.8

G

24

23.14±0.15

0.77

3.32

22.2-24.0

R

21

22.75 ±0.05

0.23

1.01

21.4-24.0

H

20

23.11±0.12

0.57

2.42

22.1-24.2

S

30

23.13±0.07

0.38

1.64

22.2-24.0

I

37

23.39±0.13

0.82

3.50

22.0-24.4

T

24

22.94±0.14

0.68

2.96

21.7-23.7

I

19

23.13±0.14

0.59

2.56

22.1-24.1

U

54

22.45±0.14

1.01

4.50

20.1-24.5

K

28

23.30±0.12

0.64

2.75

22.4-24.4

V

32

22.81±0.13

0.75

3.28

21.8-24.0

60

Annals of Carnegie Museum

VOL. 35

weigh the relative effects of environment on one hand and heredity
on the other, raising such a group to subspecific status should not be
undertaken rashly, but only after careful weighing of all of the evidence
available. The informal recognition of a local race to aid in the study
and clarification of the factors involved in its formation is one thing,
but formally describing it as a well defined entity in a phylogenetic classifi-

ABCDEFGHIJKLMNOPQRSTUV

mrvt 2
1 3 3 —

ft -E

- p

12 - —

L

j

iH

*4

Fig. 11. Variation in cranial breadth of Blarina brevicauda

Fig. 12. Variation in rostral breadth of Blarina brevicauda

ABCDEFGH IJ KLMNOPQRSTUV

§ft

3

ft

1

_L .

I

ft

i _

^ A

Fig. 13. Variation in maxillary breadth of Blarina brevicauda

Guilday: Variation in Blarina from Pennsylvania

61

cation is quite another. One of the basic goals of the study of evolution
at any taxonomic level should be the recognition of the forces acting

TABLE 8. VARIATION IN CRANIAL BREADTH OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig. 11)

Sta- Sta-

tion

N

M

a

V

R

tion

N

M

a

V

R

A

72

12.51 ±0.03

0.33

2.63

11.7-13.2

L

46

12.32±0.08

0.54

4.38

11.8-13.3

B

17

12.47±0.04

0.20

1.60

12.1-12.8

M

21

12.32±0.09

0.41

3.33

11.8-12.7

C

19

12.36±0.07

0.33

2.66

11.8-12.9

N

15

12.23rfc0.il

0.44

3.60

11.8-12.8

D

29

12.46±0.08

0.43

3.45

11.8-13.1

O

13

12.34±0.16

0.58

4.70

11.6-12.7

E

54

12.50±0.04

0.30

2.40

11.7-13.1

P

17

12.47rfc0.03

0.14

1.12

11.8-13.1

F

20

12.44±0.08

0.37

2.97

11.8-12.9

Q

24

12.37 ±0.06

0.32

2.58

11.8-12.9

G

24

12.33±0.04

0.24

1.94

11.9-12.9

R

20

12.47±0.07

0.33

2.65

11.8-13.0

H

20

12.42±0.08

0.38

3.05

11.8-13.3

S

28

12.39±0.10

0.56

4.52

11.5-13.2

1

37

12.58±0.06

0.38

3.02

11.9-13.2

T

21

12.26+0.11

0.51

4.16

11.6-13.0

1

20

12.39±0.06

0.26

2.09

12.0-13.0

U

50

12.27±0.07

0.47

3.83

11.4-13.2

K

26

12.42±:0.08

0.40

3.22

11.8-13.2

V

32

12.14rfc0.08

0.48

3.95

11.8-12.7

TABLE 9. VARIATION IN ROSTRAL BREADTH OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA. (See Fig 12)

Sta- Sta-

tion

N

M

<r

V

R

tion

N

M

<T

V

R

A

77

3.14±0.01

0.13

4.14

2.9-3.4

L

48

3. 00 ±0.02

0.14

4.66

2.7-3.3

B

18

3.17rfc0.03

0.13

4.10

3.0-3.5

M

23

3.10±0.03

0.14

4.52

2.8-S.6

C

19

3.O7rfcO.03

0.16

5.21

2.8-3.4

N

15

3.06±0.06

0.24

7.84

2.9-3. 3

D

29

3.11 ±0.02

0.14

4.50

2.9-3.4

O

13

3.09±0.05

0.17

5.50

2.9-3.3

E

55

3.16±0.02

0.17

5.38

2.8-3.4

P

18

3.11 ±0.04

0.17

5.47

2.8-3.3

F

22

3.13±0.03

0.14

4.47

2.9-3.3

Q

26

3.06±0.02

0.12

3.92

2.9-S.4

G

25

3.19±0.03

0.15

4.70

2.9-3.5

R

22

3.03±0.01

0.05

1.07

2.8-3.3

H

23

3.19±0.02

0.11

3.47

2.9-S.4

S

31

3.03±0.03

0.17

5.61

2.9-3.3

I

38

3.18±0.04

0.26

8.17

2.9-S.8

T

29

3.03±0.03

0.14

4.62

2.8-S.4

J

19

3.05±0.03

0.14

4.59

2.8-S.2

U

56

3.01 ±0.02

0.14

4.65

2.7-3.3

K

28

3.11 ±0.02

0.10

3.21

2.8-3.5

V

36

3. 08 ±0.04

0.24

7.79

2.8-S.4

TABLE

10. VARIATION IN

MAXILLARY BREADTH OF BLARINA BREVI-

CAUDA

FROM VARIOUS LOCALITIES

IN

PENNSYLVANIA.

(See Fig. 13)

Sta-

Sta-

tion

N

M

<r

V

R

tion N

M

<r

V

R

A

77

7.84±0.02

0.22

2.80

7.0-8.3

L

48

7.71 ±0.03

0.24

3.11

7.2-8.2

B

18

7.88±0.03

0.14

1.77

7.6-8.2

M

24

7.75 ±0.08

0.40

5.16

7.4-8. 1

C

19

7.77±0.03

0.17

2.11

7.5-8.2

N

15

7.82±0.06

0.25

3.19

7. 4-8. 2

D

29

7.82±0.03

0.19

2.42

7. 5 -8. 3

O

13

7.71 ±0.09

0.33

4.28

7.4-8.0

E

56

7.87±0.02

0.14

1.78

7.0-8.4

P

18

7.87±0.10

0.41

5.21

7.6-82

F

22

7.71 ±0.06

0.30

3.89

7.3-8.0

Q

26

7.86±0.04

0.23

2.93

7.4-8.1

G

25

7.84±0.03

0.14

1.78

7.6-8. 1

R

22

7.66±0.05

0.22

2.87

7.3-8.0

H

23

7.98±0.05

0.28

3.50

7.6-8.8

S

30

7.79±0.04

0.22

2.82

7.5-8.2

I

38

7.99±0.08

0.53

6.63

7.6-S.6

T

28

7.72±0.06

0.31

4.01

7.4-8. 1

J

19

7.9Q±0.05

0.22

2.78

7.5-8.3

U

56

7.75 ±0.05

0.34

4.39

7.0-8.2

K

28

7.85±0.04

0.20

2.55

7.6-8.4

V

37

7.77±0.04

0.22

2.83

7.4-8.2

62

Annals of Carnegie Museum

vol. 35

A8CDEFGH! JXLMNOPQRSTUV

12.0 ^ —

9 0

Fig. 14. Variation in length of upper tooth row of Blarina brevicauda

TABLE 11. VARIATION IN LENGTH OF UPPER TOOTH ROW
OF BL ARINA BREVICAUDA FROM VARIOUS LOCALITIES
IN PENNSYLVANIA. (See Fig. 14)

Sta- Sta-

don

N

M

a

V

R

tion

N

M

a

V

R

A

75

10.83±0.03

0.31

2.86

10.0-11.3

L

49

10.62±0.06

0.46

4.33

10.1-11.2

B

18

10.86±0.04

0.17

1.56

10.6-11.1

M

23

10.62±0.05

0.24

2.26

105-10.9

C

18

10.70±0.04

0.19

1.77

10.3-11.2

N

15

10.56±0.02

0.07

.66

10.0-11.0

D

29

10.76±0.05

0.28

2.60

10.3-11.4

O

13

10.67±0.04

0.14

1.31

10.4-11.3

E

57

10.71 ±0.04

0.31

2.89

10.0-11.6

P

18

10.50±0.11

0.47

4.48

10.1-11.2

F

22

10.38±0.09

0.42

4.05

10.0-10.9

Q

26

10.68±0.04

0.20

1.87

10.4-11.0

G

25

10.81 ±0.07

0.35

3.23

9.9-11.4

R

22

10.37±0.09

0.44

4.24

9.4-11.1

H

23

10.72±0.06

0.30

2.79

10.0-11.2

S

31

10.62±0.09

050

4.71

10.2-11.0

I

38

10.71 ±0.05

0.33

3.08

10.1-11.3

T

29

10.59±0.04

0.22

2.08

10.1-11.1

J

20

10.73±0.09

0.41

3.82

10.0-11.1

U

53

10.54±0.06

0.47

4.46

9.7-11.4

K

27

10.74±0.08

0.42

3.91

10.0-11.4

V

35

10.57±0.09

0.54

5.11

9.9-11.4

TABLE 12. VARIATION IN WIDTH OF FORAMEN MAGNUM OF
BLARINA BREVICAUDA FROM VARIOUS LOCALITIES IN PENNSYLVANIA

Sta- Sta-

tion

N

M

<r

V

R

tion

N

M

c

V

R

A

74

3.20±0.02

0.16

5.00

2.8-3.6

L

47

3.23±0.02

0.14

4.33

2.9-S.7

B

18

3.34±0.03

0.15

4.49

3.0-3.6

M

19

3.32±0.04

0.17

5.12

3.0-3.5

C

18

3.32±0.03

0.16

4.81

3.0-3.6

N

14

3.25±0.05

0.17

5.23

2.9-3.5

D

29

3.33±0.03

0.17

5.10

3. 1-3.7

O

13

3.27±0.08

0.28

8.56

2.9-35

E

57

3.25 ±0.03

0.22

6.77

2.8-3.5

P

18

3.21 ±0.04

0.17

5.29

3.0-35

F

19

3.24±0.06

0.26

8.02

3.0-3.5

Q

25

3.27±0.03

0.15

4.58

3.0-35

G

24

3.24±0.03

0.17

5.24

2.9-3.6

R

21

3.26±0.05

0.24

7.36

3.0-3.8

H

20

3.27±0.04

0.20

6.11

2.9-3.6

S

27

3.28±0.05

0.24

7.32

2.9-35

I

36

3.26±0.04

0.28

8.58

2.9-3.6

T

24

3.24±0.03

0.17

5.25

2.9-3.6

J

20

3.30±0.05

0.24

7.27

3.0-3.6

U

52

3.21±0.03

0.24

7.48

3.0-35

K

26

3.30±0.04

0.22

6.66

3.0-3.7

V

36

3.24±0.04

0.26

8.02

3.0-35

Guilday: Variation in Blarina from Pennsylvania

63

on the organism; not the relatively unproductive procedure of naming
new forms ad infinitum with little thought as to what lies behind them.

In order to work within any group the subgroups must, of necessity,
have some sort of tag on them; some method of designation must be
employed for practical reasons, if for no other. The admission of a group
to a phylogenetic classification carries with it many implications concerning
the genetic composition of that group. Morphological differences are
assumed, on good grounds in most cases, to be expressions of differing
genetic make-ups. Every variation, which may conceivably have any bearing
on the problem at hand should be recognized of course. But, without a
thorough knowledge based upon adequate samples and a comprehensive
knowledge of variation within the larger category of which they are a
part, minor secondary fluctuations may be brought to the foreground;
and the primary trends lying behind them may be overlooked. The forest
soon becomes lost in the trees.

TABLE 13. VARIATION IN INTERORBITAL BREADTH OF BLARINA
BREV1CAUDA FROM VARIOUS LOCALITIES IN PENNSYLVANIA

<7 V R

Sta-

tion N

M

A

77

5.95 ±0.02

B

17

5. 98 ±0.04

C

18

5. 93 ±0.04

D

29

5.90±0.05

E

57

5.93±0.Q4

F

21

5.86±0.Q5

G

25

5.99±0.04

H

23

6.03±0.03

I

38

5.97±0.03

J

19

5. 93 ±0.07

K

28

6.04±0.03

0.19

3.19

5.7-6.6

0.17

2.84

5.8-6.2

0.19

3.20

5.6-6.2

0.25

4.24

S.7-6.4

0.31

5.23

5.5-6.S

0.24

4.09

S.6-6.3

0.22

3.67

5.7-6.4

0.19

3.15

5. 8-6.5

0.22

3.68

5. 5 -6.4

0.32

5.39

5.5-6.1

0.14

2.32

5.7-6.5

Sta-

tion

N

M

L

48

5.86±0.05

M

24

5.92±0.04

N

15

5.91 ±0.06

O

13

5.89±0.08

P

17

5.92±0.04

Q

26

5.98±0.04

R

22

5. 90 ±0.06

S

31

5.94±0.05

T

30

5.88±0.04

U

56

5.80±0.03

V

37

5.84±0.04

<T

V

R

0.32

5.46

5.3-6.5

0.22

3.72

5. 5-6.5

0.23

3.89

5.7-6.1

0.28

4.75

5.S-6.2

0.17

2.87

5.5-6.6

0.19

3.17

5.5 -6.3

0.27

4.58

5.6-6. 1

0.30

5.05

S.6-6.4

0.20

3.40

5.4-6.2

0.22

3.79

5.3-6.S

0.22

3.77

5. 5 -6.2

TABLE 14. VARIATION IN LENGTH OF UPPER MOLARIFORM SERIES OF
BLARINA BREVICAUDA FROM VARIOUS LOCALITIES IN PENNSYLVANIA

Sta-

tion

N

M

cr

V

R

A

77

6.25 ±0.02

0.19

3.04

5.8-6.6

B

18

6.27±0.03

0.16

2.55

6.0-6.6

C

18

6.16±0.02

0.10

1.62

5.9-6.S

D

29

6.24±0.02

0.14

2.24

6.0-6.5

E

57

6.19±0.04

0.32

5.17

5.8-6.6

F

22

6.06±0.05

0.26

4.29

5.8-6.4

G

25

6.22±0.03

0.21

3.37

S.9-6.7

H

23

6.25 ±0.03

0.19

2.14

5.S-6.6

I

38

6.24±Q.Q4

0.30

4.80

5.9-6.5

I

20

6.30±0.06

0.26

4.13

6. 1-6.5

K

27

6.19±0.06

0.33

5.33

5.9-6.5

Sta-

tion

N

M

<r

V

R

L

49

6.16±0.01

0.10

1.62

5. 9-6.5

M

23

6.23±0.05

0.26

4.17

S.9-6.4

N

15

6.23±0.04

0.17

2.73

6.0-6.5

O

13

6.26±0.10

0.36

5.75

6. 1-6.5

P

18

6.17±0.06

0.24

3.89

5.8-6.S

Q

26

6.23±0.06

0.29

4.65

5.9-6.G

R

22

6.09±0.07

0.33

5.41

5.5-6.5

S

31

6.20±0.02

0.10

1.61

S.9-6.4

T

30

6.23±0.04

0.24

3.85

6.0-6.5

U

53

6.19±0.04

0.30

4.85

5.7-6.0

V

37

6.24±0.04

0.26

4.17

6.0-6.6

64

Annals of Carnegie Museum

vol. 35

TABLE 15. VARIATION IN CRANIAL HEIGHT OF BLARINA BREVICAUDA
FROM VARIOUS LOCALITIES IN PENNSYLVANIA

Sta-

tion

N

M

<T

V

R

Sta-
tion N

M

<7

V

R

A

75

6.57±0.02

0.18

2.73

6.2-6.9

L

47

6.42±0.04

0.28

4.36

6.1-6.8

B

18

6. 55 ±0.04

0.18

2.74

6. 1-6.8

M

22

6.55 ±0.04

0.17

2.59

6. 1-6.9

C

19

6.56±0.04

0.19

2.89

6.3-7.0

N

14

6.56±0.04

0.16

2.44

6.2-6.8

D

28

6.59±0.03

0.18

2.73

6.2-6.9

O

13

6.57±Q.10

0.36

5.48

6.1-6.8

E

56

6.46±0.04

0.10

4.95

5.9-7.0

P

18

6.52±0.06

0.26

3.98

6.1 -7.0

F

19

6.25±0.06

0.26

4.16

5.9-6.6

Q

26

6.50±0.05

0.26

4.00

6. 1-6.7

G

24

6.55 ±0.03

0.19

2.90

6.0-6.9

R

20

6.47 ±0.06

0.28

4.32

6.3-6.7

H

21

6.52±0.03

0.17

2.60

6.2-6.8

S

27

6.53±0.03

0.14

2.14

6.S-6.8

I

35

6.62 ±0.06

0.36

5.43

6.4-6.9

T

24

6.32±0.03

0.17

2.69

6.1-6.6

J

20

6.44±0.07

0.32

4.97

5.9-6.S

U

53

6.42 ±0.04

0.31

4.83

6. 0-6.9

K

28

6.54±0.04

0.24

3.67

6.1-6.9

V

35

6.32±0.05

0.30

4.75

6.0-6.6

CONCLUSIONS

The range of Blarina brevicauda, of which the state of Pennsylvania
comprises only a relatively small portion, extends across that portion of
the United States and southern Canada east of the Great Plains proper.
The state encompasses a rectangular area, 300 miles from east to west
and 180 miles from north to south. Pennsylvania traverses seven physio-
graphic provinces (Fenneman, 1938). These are (1) Coastal Plain Province,
(2) Piedmont Province, (3) Blue Ridge Province, (4) Ridge and Valley
Province, (5) Appalachian Plateau Province, (6) New England Province,
and (7) a small section of the Central Lowland east of the Mississippi
in extreme northwestern Pennsylvania bordering Lake Erie. Superimposed
upon, but uncorrelated with, this involved physiographic situation Blarina
brevicauda presents a relatively simple picture of clinal variation. It
becomes smaller in overall dimensions and exhibits slight morphological
differences in the crania towards the east.

The Appalachian mountain system, as it extends northward from Georgia
through Pennsylvania, becomes progressively less efficient in its ability
to isolate cryophilic forms. From an altitude of 6711 feet at Mt. Mitchell,
North Carolina, the system declines northwards through Virginia and
Maryland into the South Mountain range of southeastern Pennsylvania,
scarcely 2000 feet in elevation. Similarly, the high plateau of the Catskills
in southern New York, which reaches 4000 feet in places, slopes southward
to the Delaware River and forms, in Pennsylvania, the Pocono Plateau
of only 2000 feet elevation (Lesley, 1885). Ortman (1913, p. 347-348)
points out the interesting fact that as one progresses north along the
mountains the head-waters of the eastern draining streams push farther
and farther to the west. He says, “Coming to the Potomac drainage, we
observe that this river has cut clear across the mountains, and has reached
in northeastern West Virginia and in western Maryland, the western
[sic] boundary of the Allegheny Plateau, Allegheny Front, and at one
point has even cut through this and encroached upon the Allegheny

Guilday: Variation in Blarina from Pennsylvania

65

Plateau . . . The Susquehanna in Pennsylvania has progressed farthest in
the capture of the western streams. It has not only cut clear across
the mountains, but has also invaded a large section of the plateau, which
originally drained to the westward”.

Thus in Pennsylvania the Appalachians break down; divides are lowered;
the ranges have been completely captured by eastern drainage; and
ecological differences between physiographic provinces, have become less
efficient in limiting the distribution of mammals. This becomes apparent
when one examines the ranges of such mammals as Parascalops breweri,
Sorex cinereus, Sorex fumeus and Sorex palustris, which in the south are
restricted to the mountain provinces (Burt, 1952). The Appalachian Moun-
tains provide a southern extension of the habitat to which these forms
are adapted, but in the north their ranges fan out both east and west
as interprovincial limiting factors become less intense.

Straddling, as it does, the Appalachian mountain system at a point
where major ecological areas begin to break down and function less
efficiently as zoogeographic barriers, the fauna of Pennsylvania might
be expected to demonstrate a condition of intergradation between its
western and eastern facies. This is apparently reflected in the variation
exhibited by Blarina brevicauda within Pennsylvania.

The Southern Appalachian mountain system affords suitable habitat
for the large northern forms of Blarina which extend southward via the
mountains, pushing a tongue deep into the territory of the smaller southern
form, Blarina b. carolinensis, which surrounds the mountains on three
sides. Blarina b. carolinensis is a form of the southern lowlands, the lower
Mississippi Valley, west to the Great Plains, and the southeastern sea-
board north to Virginia, but apparently has not been successful in the
mountains proper.

Within Pennsylvania, an east-west cline has been demonstrated in
Blarina . As one progresses eastward in the state the skulls approach the
appearance of those of Blarina 5. carolinensis; the samples from the south-
eastern sector of the state are the extreme of the cline as represented in
Pennsylvania.

Assuming this cline to be primarily genetic in origin, it would seem to
demonstrate intergradation between the population of Blarina brevicauda
to the west with that of the smaller southern lowland form, north of the
Appalachian barrier. Merriam (1895), and Bole and Moulthrop (1942),
point out that Blarina appear to become smaller as one progresses from
Ontario southward, east of the Great Lakes. This again would seem to
be in the direction of carolinensis, only in this case the diminution in
size is from north to south rather than from west to east.

This hypothesis, however, is based solely upon the indirect evidence
of Pennsylvania specimens, and is offered as one possible interpretation.
The area examined is small, and therefore any attempt to relate it to
other areas can be hazardous. Further taxonomic research on Blarina
brevicauda is needed to clarify the Post-Pleistocene evolution within the
species.

66

Annals of Carnegie Museum

vol. 35

LIST OF SPECIMENS EXAMINED

Specimens from Pennsylvania numbered 1193 distributed as follows:
Beaver Co., 1 mi. NE. of Darlington, 40; Bieler Run, 0.5 mi. W. of Bealer
Run School, 17; 5 mi. N. of New Galilee, 4; 2 mi. E. of Industry, 7;
I mi. S. of Rowtown, 8. Bedford Co., 1 mi. NE. of Osterburg, 29. Berks
Co., 2 mi. SSE. of Shanesville, 3; 2.25 mi. NW. of Lenhartsville, 5; 0.5 mi.

N. of Lenhartsville, 2; 2.5 mi. N. of Shartlesville, 1; 0.25 mi. W. of

Moselem, 3; 1 mi. S. of Eckville, 7; 2.5 mi. NNW. of Shartlesville, 1;
4 mi. SW. of Eckville, 2; 2.5 mi. N. of Monocacy, 6; 0.5 mi. N. of Douglass-
ville, 2; 1 mi. SW. of Pine Forge, 2; 3 mi. NE. of Hamburg, 2; 2.5 mi.
NW. of Albany, 6; 1 mi. S. of Athol, 1; 1.5 mi. NE. of Monocacy, 1; 1.5 mi.
SW. of Weavertown, 3; 0.5 mi. SW. of Athol, 1; 0.5 mi. W. of Athol, 1;

O. 5 mi. NE. of Earlville, 1; 1.5 mi. NE. of Scarlets Mill, 4; 2.5 mi, SSE.

of Shanesville, 2. Bradford Co., 1 mi. N. of Alba, 31. Bucks Co., 4 mi.
NW. of Dublin, 2; 1.5 mi. SW. of Point Pleasant, 3; 1.5 mi. SW. of
Dublin, 1; 2 mi. N. of New Britain, 2; 2 mi. NW. of Doylestown, 2;

1.5 mi. N. of Perkasie, 1; 0.5 mi. SE. of Buckingham, 1; 1 mi. S. of New

Britain, 2; 1.5 mi. SE. of Dublin, 1; 3 mi. SE. of Furlong, 2; 2 mi. N. of

Perkasie, 1; 3 mi. NW. of Ottsville, 2. Butler Co., 2 mi. E. of Mars, 42;

3 mi. E. of Mars, 11; Mars, 1; 2 mi. E. of Middle Lancaster, 10; Thorn
Creek, 4 mi. S. of Butler, 19; Thorn Creek, 2 mi. W. of Saxonburg, 10.
Cambria Co., Cresson, 21. Cameron Co., Three Mile Run, 8 mi. NNW.
of Emporium, 66. Chester Co., 0.75 mi. and 1 mi. E. of Unionville, 12.
Clinton Co., Tamarack, 9 mi. NNW. of Renovo, 28. Crawford Co., Pynia-
tuning Swamp, 77. Delaware Co., Tinicum Island, ]/2 mi. W. of Phila-
delphia, 2. Erie Co., State Game Lands 154, 22; East Springfield, 121.
Fayette Co., 0.5 mi. and 0.25 mi. S. of Ohiopyle, 22. Huntingdon Co.,

1 mi. NE. of Spruce Creek, 19. Jefferson Co., 5.5 mi. E. of Sigel, 11;
Heath Station, 8 mi. NE. of Sigel, 5. Lawrence Co., 1 mi. S. of McConnells
Mills, 17; 4 mi. SW. of New Wilmington, 4; 0.5 mi. N. of McConnells
Mills, 8. McKean Co., Sugar Run, 10 mi. SW. of Bradford, 7. Mercer Co.,

2.5 mi. W. of Mercer, 36; 12 mi. W. of Mercer, 5; 5 mi. S. of Mercer, 10;

2 mi. N. of Clarks Mills, 3; 3 mi. E. of Sandy Lake, 1. Monroe Co., 3 mi.
N., 2 mi. NW. of Pocono Lake, 19; Buck Hill Falls, 15. Montgomery Co.,

4 mi. NE. of Pottstown, 6. Pike Co., 5 mi. SE. of Greentown, 21; 5 mi.

E. of Greentown, 18. Potter Co., 4 mi. WSW. of Ulysses, 31. Somerset Co.,
4 mi. SW. of Somerset, 49. Sullivan Co., Ogdonia Creek, 4.25 mi. SSE.
of Hillsgrove, 34. Union Co., Glen Iron, 23. Venango Co., Sandy Creek,

2 mi. W. of Polk, 34; Pecan, 4 mi. S. of Franklin, 2; Sugar Creek, 3 mi.

NW. of Franklin, 2; Lower Two Mile Run, 2 mi. E. of Franklin, 5; 4 mi.
W. of Franklin, 2; 2 mi. NE. of Franklin, 7; 9 mi. E. of Oil City, 8; 6 mi.
E. of Oil City, 5; 6 mi. N. of Oil City, 18; 1 mi. SW. of Utica, 12; 1.5
mi. NE. of Kennerdell, 1; 2 mi. S. of Franklin, 3. Warren Co., Benson
Swamp, 5 mi. E. of Columbus, 29; 1.5 mi. N. of Pittsfield, 36; 2 mi. S. of
Weldbank, 1; 2.5 mi. N. of Kinzua, 11; Miles Run, 5 mi. NW. of Pitts-
field, 9; 5 mi. N. of Pittsfield, 4; 4 mi. SW. of Youngsville, 5; 1 mi. SE.

of Garland, 17; 2 mi. N. of Pittsfield, 5; 6 mi. N. of Pittsfield, 6.

Guilday: Variation in Blarina from Pennsylvania

67

SUMMARY

This article presents a method of arriving at the relative age of speci-
mens of Blarina brevicauda based upon cranial suture closure.

Males averaged significantly greater than females in total length, length
of hind foot, weight, total length of skull and occipital height in a sample
of 121 specimens from northwestern Pennsylvania.

Age differences in the samples studied were slight. All immature shrews
collected were of essentially adult size. However, significant differences
between younger and older specimens were found in weight, total length
of skull, cranial height, maxillary breadth, and width of foramen magnum.

External measurements were found to be significantly influenced by
personal bias.

A west-east cline is demonstrated for Blarina brevicauda in Pennsylvania,
shrew crania becoming smaller and less rugged towards the east. It is
suggested that this demonstrates intergradation with the smaller Atlantic
Coast form, Blarina brevicauda carolinensis.

The most recent taxonomic work on the northern races of Blarina brevi-
cauda appears to be at variance with the results reported herein, as far
as Pennsylvania is concerned, and thus suggests the need for future study.

REFERENCES

Araback-Christie-Linde, Augusta

1907. Die Bau der Soriciden und ihre Beziehung zu andern
Saugetieren. Morphologisches Jahrbuch, v. 36, pt. 4, p.
463-514.

Bole, B. Patterson, Jr., and Philip N. Moulthrop

1942. The Ohio recent mammal collection in the Cleveland Museum
of Natural History. Scientific Publications of the Cleveland
Museum of Natural History, v. 5, no. 6, p. 83-181.

Bert, W. H.

1952. A field guide to the mammals. 200 p. Houghton, Mifflin Co.,
Boston.

Boett, J. Kenneth

1942. A review of the genus Phoca. Annals of the Carnegie Museum,
v. 29, p. 61-125.

Fenneman, Nevin M.

1938. Physiography of eastern United States. 714 p. McGraw-Hill
Book Company, New York.

Gifford, Clay L., and Ralph Whitebread

1951. Mammal survey of south central Pennsylvania. Final Report
Pittman-Robertson Project 38-R. 75 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Grimm, William C., and Harvey A. Roberts

1950. Mammal survey of southwestern Pennsylvania. Final Report
Pittman-Robertson Project 24-R. 99 p. Pennsylvania Game
Commission, Harrisburg, Pa.

68

vol. 35

\ J 1 t i ' - 5 < ^ * 5' , ( « ' ' * * %

Annals of Carnegie Museum

Grimm, William C., and Ralph Whitebread

1952. Mammal survey of northeastern Pennsylvania. Final Report
Pittman-Robertson Project 42-R. 82 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Lesley, J. P.

1885. A geological hand atlas of Pennsylvania, embodying the
results of the field work of the survey, from 1874 to 1884.
13 + cxii p. + 61 maps. Second Geological Survey of
Pennsylvania, Harrisburg, Pa. Report of Progress X.

Merriam, C. Hart

1895. Revision of the shrews of the American genera Blarina and
Notiosorex. North American fauna, no. 10, p. 5-34. Govern-
ment Printing Office, Washington, D.C.

Ortmann, Arnolde

1913. The Alleghenian Divide, and its influence upon the fresh-
water fauna. Proceedings of the American Philosophical
Society, v. 53, p. 287-396.

Pearson, O. P.

1945. Longevity of the short-tailed shrew. American Midland
Naturalist, v. 34, p. 531-546.

Pruitt, William O., Jr.

1954. Aging in the masked shrew, Sorex cinereus cinereus Kerr.
Journal of Mammalogy, v. 35, no. 1, p. 35-39.

Richmond, Neil D., and Harry R. Rosland [sic]

1949. Mammal survey of northwestern Pennsylvania. Final Report
Pittman-Robertson Project 20-R. 67 p. Pennsylvania Game
Commission and U. S. Fish and Wildlife Service, [Harris-
burg, Pa.]

Roberts, Harvey A., and Robert C. Early

1952. Mammal survey of southeastern Pennsylvania. Final Report
Pittman-Robertson Project 43-R. 70 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Roslund, Harry R.

1951. Mammal survey of northcentral Pennsylvania. Final Report
Pittman-Robertson Project 37-R. 55 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Warren, Edward Royal

1942. The mammals of Colorado, ed. 2. 330 p. University of
Oklahoma Press, Norman, Okla.

ART. 6. THREE INTERESTING LEPIDOPTERA FROM
THE PHILIPPINES
By I-Iarry K. Clench

Associate Curator, Section of Insects and Spiders
Carnegie Museum

Some twenty-five years ago Carnegie Museum received a small collection
of butterflies and moths made by C. F. Clagg on the island of Mindanao,
Philippine Islands. With the exception of a small amount taken at or
near Davao, most of Mr. Clagg's material comes from rather high eleva-
tions in the mountainous southeastern part of the island and is, thus, of
considerable interest.

For several reasons it is not feasible to prepare a complete list of his
captures. I hope, however, from time to time to be able to publish on
some of his more noteworthy species. In an earlier paper (Clench, 1953)
two new agaristid moths were described from this material. The present
note considers two butterflies, one new and the other hitherto known only
from the unique type, to which I have added a new agaristid moth re-
cently received from another source but closely related to one of those
taken by Mr. Clagg and discussed in the paper referred to above.

Delias levicki Rothschild (Pieridae)

Rothschild described this rare species (Rothschild, 1927) from a single
male from Mindanao, no more exact localization of the specimen being
known. Talbot (1928) published on a similar specimen from Mt. Apo.
In view of its slight but undeniable differences from levicki as well as
the unknown origin of the latter, Talbot came to the entirely reasonable
conclusion that his specimen represented a Mt. Apo subspecies of levicki
and that the latter accordingly must have been taken on some other moun-
tain or range on the island. He therefore proposed for his specimen the
name Delias levicki apoensis and later (Talbot, 1937, p. 402, plate 50,
fig. 1) published a figure of the type.

Among Mr. Clagg's captures is a single female (Galog R., Mt. Apo,
6000 ft., S. Mindanao) agreeing well with Talbot's redescription and
figure (Talbot, 1937, plate 62, fig. 3) of the levicki holotype, save in the
following particulars. The ground color above is a pale yellow rather than
the creamy white of levicki , the hind wing being further faintly tinted with
greenish. The subterminal spots are smaller, less definite in outline and
darker, almost orange yellow, especially subapically on the fore wing. On
the under surface there is no white patch at the fore wing cell-end and the
veins of the hind wing are not white, or only faintly so on the median veins
only. The subterminal “yellowish olive” band of levicki is feebly indi-
cated, barely recognizable.

In view of this evidence of the occurrence of levicki on Mt. Apo (and
whether or not the above differences are only individual in character)
it is clear that apoensis Talbot should be recognized as a full species,
occurring with levicki and not replacing it geographically.

Issued

Uan.28,1

69

70

Annals of Carnegie Museum

vol. 35

Danaus apoxanthus, sp. nov. (Danaidae)

Frons black-brown with a median pale yellowish line and a spot of
similar color below each antennal base. Palpi with second segment ap-
pressed to front, yellow with a ventro-mesal and a ventro-lateral longi-
tudinal black-brown line and basally and apically black laterally; third
segment porrect, black with a yellow line dorsally and ventrally, each
broadest at base. Eyes bordered dorso-mesad by a yellow dash and poster-
iorly by a yellow line; posterior to antennal bases and well between the
eyes a pair of cuneiform yellow spots. Antenna black, very faintly tinged
with ruddy, minutely dark fulvous at extreme apex; club gradually in-
crassate but strong, its greatest diameter about 2.5 times that of middle
of shaft. Patagia black-brown with a pair of rather large yellow spots
dorsally. Tegulae black-brown with a yellow spot anteriorly just below
base of fore wing costa; a larger yellow spot at, or just above, level of
costa; the whole narrowly bordered with yellow. Thorax dorsally black-
brown with longitudinal yellow line from one end to the other; laterally
with long, sparse yellowish hair-scales. Thorax ventrally with the black-
brown ground color almost obscured by profuse, rather large yellow spots.
Legs brown, paler than body color; rather strongly marked with yellow
on fore tibia, mesad on mid and hind femora and exposed portions of
trochanters. Abdomen dorsally dark gray with a tint of brown; a dorso-
lateral yellowish white longitudinal streak from base to about middle
and a much fainter mid-dorsal pale streak from base to about one-fourth;
ventrally pallid, rather yellowish pale gray; pale yellowish without gray
transversely on distal margins of sterna.

Upper side

Male. Both wings bright reddish-brown darkening abruptly to dark,
almost blackish-brown in outer half of post-discal area and very narrowly
on costa. Hyaline areas on both wings bright yellowish with a tint of
green, in worn specimens (chemical action?) brownish; arranged in a
pattern as follows. Fore wing with a single longitudinal discal cell patch
from base to very near cell apex; along costal half of cell it is clouded
over with reddish-brown scales and along the costal border of the cell
it is completely replaced by the ground; a subapical hyaline patch or band
crosses diagonally from costa opposite cell-end to midway out on Ma, com-
posed of juxtaposed spots as follows. One in Ri-R2 twice as long as wide,
filling interspace to as far basad as cell-end; a tear-drop shaped spot
in Ra-Rs filling base of interspace; a minute dot (often absent) in base
of R3-R4-5; a cuneiform one filling base of R^-M*; a subquadrate one
in M2-Mg reaching basad not quite as far as cell-end and a little over
three times as long as wide, filling interspace width; an irregularly sub-
quadrate one in M2-M3 filling interspace and larger in middle (due to
bowing of veins) than at either end, both ends excavated and the basal
edge lying opposite middle of preceding; just before margin a row of five
oval spots, one each in interspaces R^R^-Rs-Mj-M^Mg; relative sizes as
follows: small or absent, largest, smaller, smaller, occasionally absent, larger;
base of Mg-Cu, without a spot other than a feeble appearance of the one

Clench: Three Lepidoptera from the Philippines

71

beneath which may show through; base of Cu^Cua with a pair of sub-
quadrate, variably shaped spots, large, but not entirely filling interspace,
separated by a transverse reddish-brown band or line. Base of Cu2-2A
with an elongate patch, usually washed with reddish-brown scales and with
a definite though faint slender longitudinal streak of ruddy brown through
it from its base to its distal end. In M3-Cu1-Cu2-2A before margin a series
of rather large round spots subequal in size (the first slightly larger). The
submarginal paired dots in each interspace, so characteristic of most Danaus,
are here almost completely lost. They appear faintly from Ms to Cu2 in
one male (paratype) but not in holotype. Hind wing with costa above Rs
grayish with a hyaline elongate streak and a subterminal dot. Discal cell
filled with a hyaline patch save along costal edge to base of Mx and along
lower border between bases of M2 and Cu2, these edges being ruddy brown,
the former shading to grayish basad. A hyaline spot in base of Mx-M2,
longest along cell and along Mx, shortest along M2 and across from there
to Mx, broadly separated from all veins; post-discal hyaline spots in bases
of Rs-Mx and M2-M8-Cu1( absent in Mx-M2 or present only as a small
elongate streak; extreme base of Cux-Cu2 with a hyaline triangular spot
in angle between Cu2 and cell; Cu2-2A with a deeply bifid elongate hyaline
spot, each branch acutely pointed, reaching two-thirds along 2A, the branch
along Cu2 rather shorter; 2A-3A with a long slender spot similar to the
2A half of the preceding and reaching barely farther; between 3A and inner
margin with a similar but more whitish and slightly shorter one. Before outer
margin a row of round, single (Rs-M3) or oval, double (M3-Cua) spots.
Two-thirds out on Cu2 an oval black scent-patch; on 2A adjacent to it a
round gray scent-patch, the vein slightly swollen through it.

Female. Similar to the male, with the following exceptions. Fore wing
with discal cell more washed with ruddy brown scales, as also basal spot
in Cu^Cus (so much so that the spot is virtually absent) and the elongate
spot in Cu2-2A. The second spot out in cell Cux-Cu2 is differently shaped,
being diagonally elongate (basad towards Cux, distad towards Cu2). Hind
wing. Hyaline area in cell narrower; elongate hyaline patches or streaks
from Cu2 to inner margin shorter; androconial patches absent.

Under side

Male. As on upper surfaces except as follows. Distal black-brown is
here rather pallid gray-brown, which furthermore runs along inner margin
of fore wing below 2A to base. Hind wing with basal and discal ruddy
brown very feebly developed, almost all replaced by pallid gray-brown.
Hyaline areas of both wings similar in size and location save that the
basal portions on fore wing (in discal cell and in bases of cells M„-2A)
less washed with ground color scales; a basal spot in Mg-CUi not present
on upper side, though sometimes showing through. Scent-patch on Cu2 jet
black, the other (2A) grayish, barely visible, though the inflation of the
vein through it is more noticeable here than above.

Female. Similarly different from female upper surface. Fringe of both
sexes black above, gray below, broadly but inconspicuously paired white
in each interspace and white along hind wing inner margin.

72

Annals of Carnegie Museum

vol. 35

Length of fore wing. Males 40.0 (para type), 41.5 (holotype); females
both 45.0 mm.

Holotype and allotype, Seliban River, Mt. Apo, 7000 ft., S. Mindanao,
Philippine Islands, 21.ix.1930 (C. F. Clagg); C. M. Ent. Acc. No. 9163.
Paratypes: One male and one female, Sibulan River, Mt. Apo, 2000 ft.,
ll.v.1930 (Clagg), same accession number. All types, C.M. Ent. type series
no. 188.

Remarks. It is surprising that so striking a species has hitherto escaped
notice, the more so as Mt. Apo has been collected on several times in the past.

D. apoxanthus has no very close relatives, though it clearly belongs in
Fruhstorfer’s “Artengruppe Chittira Moore” (Fruhstorfer, 1910), equivalent
to Talbot’s “ Melaneus group” (Talbot, 1943). Talbot, of course, was en-
tirely correct in removing the New Guinean weiskei to a group of its
own. A tentative key to this group, as thus delimited, might begin as
follows:

la. Dark ground color of upper surface different on fore wing (black-

brown) and hind wing (chestnut red) tytia and subspecies

b. Ground color similar on both wings, though varying from one form
to another 2

2a. Antesubmarginal spots of hind wing above Rs to M3 duplex as are

those from Ms to Cu2 (the pair in Rs-Mx often fused, however)

albata and subspecies

b. These spots single (though duplex from M3 to Cu2) all others

The forms keying out to couplet 2b form an assemblage with specific
determinations and delimitations still extremely doubtful. Neither
Fruhstorfer (1910) nor Talbot (1943) is of any help in establishing specific
characters within this group, though I am certain that such characters
exist, as examination of most of the named entities suggests. Both these
authors had knowledge of the male genital structures, yet neither has com-
municated this information in usable form.

Pattern characters in this group, however, are so many and so helpful
that I am at a loss to understand why these also have been so little used.
Neither of the characters mentioned in the key above has ever been pointed
out as indicative of species , yet they are visible at first glance.

D. apoxanthus will key out in the above table to this troublesome couplet
2b. From among the many species which belong here it is difficult to
single out one of them as being the closest relative. After careful com-
parison, however, I believe that it must come closer to D. phyle Felder
than to any other. The two agree in the transverse band-like appearance
of the subapical spots of the fore wing, absence of a longitudinal dark
bifid streak in the hind wing discal cell, presence of a feeble one in cell
Cu2-2A of the fore wing, costad obsolescence by gradual shading over of
the hyaline patch in the fore wing discal cell. The androconial patches
of both are similar in position and structure though those of apoxanthus
are only about half as large as those of phyle. The abdomens of both are
similarly colored save that in apoxanthus it is ventrally rather more
yellowish.

Clench: Three Lepidoptera from the Philippines

73

Distinctions between the two are many. Besides the color of the upper
surface ground (uniform black-brown in phyle; ruddy brown, shading
distally to black-brown in apoxanthus) and of the hyaline markings (pale
bluish-green in phyle; yellowish with a greenish tint in apoxanthus), there
are these differences in pattern. Subapical spots of fore wing larger and
more nearly confluent, the one in M2-M3 basally and distally excised;
absence of the two basal spots in M^Cu^ absence of the basal spots in
Ma-Ma-CUj on hind wing; enlarged submarginal spots (which in phyle
are further completely gone from M3 to Cu2) of hind wing and, to a lesser
extent; of fore wing as well.

Philippodamias, gen. nov. (Agaristidae)

Genotype: Philippodamias jocelyna, sp. nov.

Antennae subapically thickened, apically tapering; the thickest portion
about 2.5 times the diameter of the shaft; frons with a slightly raised
median prominence, but no protuberance, and completely covered with
scales and hair; palpi with first two segments fringed ventrally with long
hair, upturned, the apex of the second reaching about to middle of eye;
the third segment with appressed scales, about half the length of second,
pointed and porrect; eyes naked; tongue fully developed; legs with middle
and hind femora fringed ventrally with rather sparse long hairs; middle
and hind tibiae with appressed scales and hairs; middle tibia with apical,
hind tibia with apical and subapical pairs of spurs, the inner one of each
pair being twice or more the length of the outer; tarsal claws of all legs
with a rather small inner tooth; approximate lengths of the leg segments
are as follows. (These measurements were made with an ocular grid
micrometer on a dissecting microscope from the legs in situ and hence
must be regarded as approximate only. For each leg is given in millimeters,
respectively: femur, tibia, tarsus I, tarsus II-V exclusive of claws.) Fore
leg, 5.2, 3.1, 2.1, 3.0; middle leg, 6.4, 5.6, 3.1, 3.5; hind leg 5.3, 6.8, 3.4, 3.5.

Venation. Fore wing. Sc from base to costa at about two-thirds; Rx from
upper cell vein at about middle, to costa at about 4/5; areole present,
R2-3-4 from its apex, the veins separating from a common point more
than two-thirds the distance from end of areole to apex of wing, the
first two to costa near apex, the last to apex; R3 from end of areole a little
separated from origin of stalked R2-a-4, to termen; from cell-end as
far below areole as width of latter at that point; Mx-M2 cross-vein long,
inbowed, feebly angulate at middle; Ma, Ma and Cux subequally spaced
at origins, respectively a little above, at, and a little before, lower cell
angle; more closely approximated to each other than breadth of areole;
Ms strongly bowed near origin; Cu2 arises at about four-fifths of cell.
Hind wing. Cell short, about one-third the distance from base to end
of M2; Sc touching cell briefly near base, thence free to apex; Rs and Mx
connate from upper cell angle; M2 from near middle of cell-end, very
little nearer Mx; M3 from lower angle of cell, lower cell-end vein (between
M2-M3) costally faint but apparently deeply and angulately inbowed; Cux
from near lower cell angle, almost connate with Ms; Cu2 from cell beyond
four-fifths.

74

Annals of Carnegie Museum

vol. 35

Philippodamias belongs in a closely interrelated group of Indo-Australian
genera which includes also the following: Fleta Jordan, Crinala Jordan,
Immetalia Jordan, Damias Walker ( Burgena Walker auct.) and Scrobigera
Jordan. (Further discussion and references to these genera may be found
in Hampson, 1901, p. 515; Jordan, 1912; Strand, 1912) This group of
genera possesses in common: naked eyes; frontal protuberance reduced
to a rounded boss; areole present, formed by the anastomosis of R4 with
R2-3; Ma from the cell, usually distinctly separate from areole. These

genera may be keyed out as follows:

la. Mid and hind tibiae with a fringe of long, erect hairs 2

b. These tibiae with appressed hairs and scales only 3

2a. Hw with Rs and Mt short stalked Fleta

b. These veins connate Crinala

3a. Distance between origins of Ms and Cux of fore wing distinctly
longer than that between origins of M2 and M3; males with andro-
conial area of long hairs in a streak from base to middle of cell-end

Scrobigera

b. Distance between origins of Mg-Mg-CUi subequal; no such andro-

conial area 4

4a. Third palpal segment nearly as long as second, with erect hairs

Damias

b. Third segment half as long as second, or less 5

5a. Second palpal segment with appressed scales Immetalia

b. This segment with ventral fringe of long, erect hairs Philippodamias

Its closest relative, rather than Immetalia as would be suggested by the
key, appears to be Damias and the three genera, Immetalia , Damias and
Philippodamias, appear to be very nearly ancestral in the group. From
some form similar in structure to these were developed such later specializa-
tions as the fringed middle and hind tibiae of Crinala and Fleta (probably
independently in each, despite the wide occurrence of this trait in the
family outside the present group); the stalked veins Rs and Mj of the
hind wing in Fleta; the very small third palpal segment of Crinala ; and
so on. Many of the characters useful and used in the classification of
the members of this group of genera appear to have had independent
origin several times, complicating the picture of their relationship.

Three genera of this group are now known to occur in the Philippines—
Crinala , Philippodamias and Scrobigera. It is interesting that two of them
( Crinala and Philippodamias) appear to be restricted to these islands,
where they seem to be quite rare, both showing distinct affinities towards
the two pre-eminently Papuan genera of the group (Immetalia and Damias).
The third, Scrobigera, is a far more typical resident of the Philippines,
being a southeastern Asiatic genus extending as far north as China and
as far east as the Celebes.

Philippodamias jocelyna, sp. nov.

Antennae black, becoming reddish on distal third, more so ventrally
than dorsally; palpus, frons, vertex of head, collar, thorax both dorsally

Clench: Three Lepidoptera from the Philippines

75

and ventrally and abdomen dorsally and ventrally, all black with metallic
blue or greenish iridescence. On the palpus laterally at the base are a
few pale brownish scales.

Upper side

Fore wing black, very slightly brownish, with a faint bluish sheen in
certain lights, crossed from costa at middle almost to tornus by a broad,
nearly straight yellowish orange band with sharply defined edges. The
band curves abruptly but slightly downward just before its lower end,
the angle on the inner edge of the band being about at the locus of lost
vein 1A, that of the outer border between Cux-Cu2. The posterior end
of the band is rather evenly rounded or faintly subangulate, ending just
before tornus and thus leaving a slender black area between it and the
adjacent wing edge about tornus. The band is of equal width throughout,
about one-fourth the length of costa in width. The orange on the anterior
edge of costa projects basad about half way to the base, visible only when
the specimen is viewed from the front. Hind wing black, distally faintly
tinged brownish, with a strong purplish-blue iridescence. This iridescence
is especially strong and bluish in the base, its outer limit straight, rather
vaguely defined, crossing wing from costa at about one-third, across cell-
end to anal angle. Beyond this the iridescence is more critical and a dark
lustrous purple. In the extreme base of cell Cuj-Cu,*, at the lower angle
of the cell, is a small patch of orange scales. Fringe of both wings con-
colorous with the ground (dark brown-black, blue iridescent).

Under side

Fore wing as above, with these exceptions. Ground color slightly browner,
the orange band more irregular, being streaked slightly basad on costa
and costal veins and again on lower cell vein and on 2A. Its distal margin
is more irregular, being displaced about one mm. distad from M2 to tornus,
the posterior end approaching tornus still closer so as to leave only the
terminal fringe dark. Hind wing as on upper side except that the veins
are prominently green-iridescent and the small orange patch is absent.
Fringe as above, everywhere dark black-brown, greenish iridescent on hind
wing.

Length of fore wing , 29.5 mm.

Holotype , male, Matuguinao, Samar Island, Philippine Islands, 7.i. 1952
(P. de Mesa). C. M. Ent. type series no. 330.

Remarks. This handsome species is congeneric with one which I de-
scribed a few years ago (Clench, 1953) as Scrohigeraf?) claggi, from the
Galog River, Mt. Apo, 6000 ft. Although at the time it was apparent
that claggi was not properly referable to Scrobigera, description of a new
genus was unwise with only the unique female specimen. With the
receipt of the present male of jocelyna, clearly congeneric, it is possible
now to assign a generic name. P. jocelyna is distinguishable from claggi
in many ways— absence of yellow on head and body; presence of the
marked iridescence (completely missing in claggi), the orange spot on
the hind wing above; darker orange of fore wing band and its slightly
different size (broader) and configuration (straighter).

76

Annals of Carnegie Museum

vol. 35

In addition to the foregoing pattern differences there are these slight
structural differences. In claggi the r-m cross-vein of the fore wing leaves
the areole at about one-fourth; in jocelyna it leaves at its middle.

Since only the female of claggi is known and only the male of jocelyna ,
it is impossible to know which of the preceding differences may be sexual
in nature and which specific.

It gives me much pleasure to name this distinctive and beautiful species
for my daughter, Jocelyne.

REFERENCES

Clench, H. K.

1953. New Indo-Australian Agaristid moths. Annals of the Car-

negie Museum, v. 33, p. 141-144.

Fmhstorfer, H.

1910. Danaidae. In Seitz. Grossschmetterlinge der Erde, v. 9, p.

191-284.

Hampson, G. F.

1901. Catalogue of the Lepidoptera Phalaenae, v. 3 (Arctiadae

Arctianae and Agaristidae). 690 p. London.

Jordan, K.

1912. Agaristidae. In Seitz. Grossschmetterlinge der Erde, v. 11,

p. 1-28.

Rothschild, W.

1927. On a new species of Delias from Mindanao, Philippine Is-

lands. Annals and Magazine of Natural History, ser. 9, v. 19,
p. 563.

Strand, E.

1912. Noctuidae Agaristinae. Lepidopterum catalogus, pt. 5, p. 1-82.

Talbot, G.

1928. Seven new form of Delias (Lep. Pieridae). Bulletin of the Hill
Museum, v. 2, p. 224-228.

1937. A Monograph of the Pierine genus Delias, pt. 6, p. 261-656.

1943. Revisional notes on the genus Danaus Kluk (Lep. Rhop.

Danaidae). Transactions of the Royal Entomological Society
of London, v. 93, p. 115-148.

W b / J

^ P k Q *4 &

v

ART. 7. WATERFOWL OF THE WESTERN PENNSYLVANIA
SKYWAYS AS PORTRAYED BY THE MIGRANTS AT ONEIDA
DAM, BUTLER COUNTY, PA.

By F. W. Preston

Director Preston Laboratories, Butler, Pa.

In Memory of Howard H. Elliott, who died suddenly in
the fall of 1952 just before this report was completed

SUMMARY

The location, size, topography, date of construction, properties of the
water, and other general data of Oneida Dam, Butler County, Pa., are out-
lined, with a brief description of its permanent flora and vertebrate fauna.
Observations of its migratory waterfowl, especially ducks, over a four-year
period, 1949-1952, inclusive, are given for both spring and fall migrations,
involving approximately one hundred and twenty visits in spring and eighty
in the fall. In particular, the number of individuals seen is reported.

The lake is used in the spring by more than three times the number of
ducks that use it in fall.

It is used by far more males than females.

The great majority of the diving ducks, at least, are in flocks predomi-
nantly of one sex, usually males.

Forty per cent of all ducks are scaup, presumably Lesser Scaup.

The proportion of Black Duck to Mallards indicates that the lake is in
the Atlantic Flyway for these species, and the frequent presence of swans
and general absence of Blue Geese agrees with this. The azimuth of the
swan's spring migration is noted.

The spring and autumn migrations are “independent” events and may
conveniently be graduated by Gaussian curves. The peaks of the migra-
tions are located, and their standard deviations determined.

The number of ducks to be expected at the peak, and the number of
species, are determined, and the standard deviations of those numbers.

Tables of earliest and latest dates on both migrations, for most of the
species, are provided. A few notes on unusual species and unusual dates are
included.

An analysis of the number of Common Loons present on 38 separate
occasions is consistent with the hypothesis that the species always migrates
solitarily, even though a number may be present simultaneously on the Dam.

The number of “bird-days” (birds x days) per season is computed for both
spring and fall migrations, for scaup and for “all ducks lumped together.”
(This gives a good estimate of the number of birds that were not observed
because no one was there to observe them.)

ACKNOWLEDGMENTS

I wish to acknowledge the help, in the field, of John M. McCormick,
formerly of the Preston Laboratories, on many of the early observations, and
of the late Howard H. Elliott of Jamisonville, Butler County, who occasion-
ally took over during my absences. The computations were made mostly by

78

Annals of Carnegie Museum

vol. 35

work. The preparing of the graphs was done mostly by Miss Mary E. Gem-
perle, and the whole of the typing by my wife, Mrs. Jane E. Preston. I wish
to thank P. H. Dowdell of the Butler Water Company for information on
the size, date of construction, and properties of the water in the Dam, and
for a list of the known fishes in it. I am indebted to Dr. O. E. Jennings,
Director Emeritus of Carnegie Museum, for a special survey of the macro-
flora of the lake bottom, and to Dr. M. Graham Netting for identification of
frogs and turtles. Finally, I am indebted to E. H. McClelland, formerly
Technology Librarian of the Carnegie Library of Pittsburgh, now retired, for
a great deal of help in arranging and editing the whole of the manuscript.

ONEIDA DAM AND THE WATERFOWL

Some forty years ago, in England, I asked my Greek and Latin professor,
an excellent amateur ornithologist, how many species of birds he supposed
one could see in my native town. “That depends,” he said, “on whether you
consider that the town extends ' usque ad caelum’ (up to the sky), for if you
count the kinds that fly over us, but seem never to land here, it would be
a very high total indeed.”

The concept of the “fly ways” is well known, largely due to Lincoln’s work.
I have avoided the term here and substituted the term “skyways,” because I
see little evidence that over Butler County there is any well defined direc-
tion into which the wild-fowl are channeled. The birds seem to go in all
directions. It is true we have but little evidence all told, but such as it is,
it indicates, I think, that the air above Butler County is, in the language of
the aviators, simply a “navigable air space” that may be navigated in any
direction whatever. This applies to the migrant wild-fowl on their spring and
fall migrations, when one might expect the movement to be simply or pre-
dominantly a north or south one. But except that there is likely to be a
northerly component to the migration in spring and a southerly one in fall,
the movement seems to cover (at any one season) almost half of the horizon.

Thus, a loon migrating in spring over the Laboratory grounds was headed
due north. A Bald Eagle in May, at a great altitude, was also headed due
north; but a Bald Eagle in February was headed almost west (slightly north
of west) on a course that would take it to Toledo. An American Egret was
headed somewhat west of north on a course that would take it to Pymatuning.
Seven wild Whistling Swans left Oneida and flew over Grove City on a true
bearing of 322°; that is, 38° west of north, on a course that would take them
over Pymatuning and Ashtabula, Ohio, on their way to the Mackenzie Delta.
If they used the same course before reaching Oneida, they should have
started from the coast somewhere near the Virginia Capes. On the other
hand, banding returns collated by Carl Warren showed that many ducks were
using a NE. to SW. flyway. Bonaparte’s Gulls over Oneida in spring seem
headed essentially west, with possibly a trace of south in their course. In the
spring I have seen Ring-bill Gulls arrive at a local cemetery pond from the
west, and again, over the Laboratory grounds, I have seen a large gull (prob-
ably a Herring Gull) also headed due east; but I have also seen several gulls
(probably Ring-bills) in a flock headed due north just over our treetops. In
the early spring I have seen a Merganser headed due west, also over the

1957 Preston: Waterfowl of western Pennsylvania Skyways 79

Laboratory grounds, and various other ducks headed almost west. In the
spring R. T. Norris reported seeing a large flock of Snow Geese over the
grounds headed almost due north, while in the fall we have often seen
Canada Geese headed due south. On the other hand, I have seen a large
flock headed south-east and watched them for miles.

In many cases the weather was fair and visibility good, so there is no
reason to suppose the birds were confused (as does happen in snowstorms),
and it is to be presumed the birds were going in their normal courses. There-
fore, I here treat the “navigable air space” as a pool of birds, flying more
or less in all directions, and the problem is simply one of finding what that
pool contains and when it contains it.

Oneida Dam (Fig. 1) five miles NE. of the center of the city of Butler, is
a net for sampling the skyways. Its efficiency is no doubt different for differ-
ent species, so it may not accurately reflect the content of the pool. But it
is the best net we have, and it has some noteworthy advantages. It is the
largest body of open water on the Foreland (the area between the Allegheny
and Ohio rivers and the terminal moraine of the Wisconsin glaciation on the
northwest) large enough to be occasionally acceptable to birds that normally
prefer much larger bodies of water, but so small that no swimming bird can
escape observation. And although it is so wide that birds near the far shore
can not always be identified with certainty by binoculars, they can with a
20-power telescope, unless the light is poor.

It has the further advantage that there are virtually no resident water-
fowl. In the winter there are none whatever, for the lake freezes over com-
pletely. In the summer there are virtually none. There are a few little Green
Herons that no doubt nest near-by, and a pair or two of Kingfishers. Once
in a while, but not every year, a pair of Black Duck will nest there. The
Wood Duck and Mallard nest in Butler County, but we have no evidence
that they nest at Oneida. Over the last 25 years, there have been occasional
years when a few egrets (mostly American Egrets) have spent the summer
there from July to September, but more years in which they are not seen
and, generally speaking, the summer months might be said to be as blank
as the winter ones.

Thus, ornithologically speaking, the lake has no existence apart from the
skyways and is merely a device for sampling those skyways. This it does in
spring from the middle of February to the end of May, and in the fall from
September to November.

In the spring the dam is full of water, and the lake is accordingly attrac-
tive to waterfowl; in the fall, in many years, the water is very low, and
extensive mud-flats are exposed at the upper (north) end. The lake is then
attractive to shore birds, but much less so to waterfowl, though some visit it.

In the spring, the ducks and other waterfowl are anxious to press on to
their breeding grounds in the north and west, and they seem to push as
far north and west as they can find open water. There is therefore a con-
centration of birds against the edge of the retreating ice, as it were. When
the weather is cold, snowy, or generally bad, if there is any open water at
Oneida, birds will be found there, sitting on the water or on the ice by the
water’s edge. But if there is a sudden warming up of the weather, there is a

80

Annals of Carnegie Museum

vol. S5

sudden drop in the bird population. Presumably they have gone north to
the ice edge. The ducks are always scouting for this edge. I have seen a
big flock of ducks come over Oneida when it was frozen from one end to
the other. The birds would come over the hills and drop down as if to
alight, fly the whole length of the lake, and then, finding no open space,
would rise and continue their search, usually to the north.

The ice edge, however, is not the only controlling factor. At a time in
the spring when the waters are open for many miles to the north or west,
the birds may run into a line-squall. Since our weather moves from west to
east and often forms a front extending SW. to NE., birds flying NW. in
perfectly good weather may run into an impassable storm front. It is then
a case of any port in a storm, and Oneida is such a port. This was the ex-
planation of our greatest visit of Whistling Swans, 133 of them descending
upon Oneida to escape a line-squall on April 1, 1949. Most of them left the
next day, but seven remained, two of them (possibly slightly injured by their
tussle with the storm) till April 23. Similar occasions frequently make it
worth while to visit Oneida, and, generally speaking, when the weather is
foul, it is good weather for ducks.

However, even this is not all. Oneida is a port of call as well as a port in
a storm. This is particularly the case with the scaup (presumably mostly the
Lesser Scaup) on spring migration. This is by far the commonest duck at
Oneida, and it can be found there in good weather as well as bad through-
out its migration season. While it is dear that birds of many species
frequently stay only one day or even part of a day, there is reason to believe
that the scaup frequently stays longer, flocks of identical numbers and
identical sex ratio being sometimes seen for several consecutive days. To
some extent this is true of other species, especially in the very early spring,
with individuals that have arrived abnormally early. Thus, a solitary male
Canvas-back was present at Oneida from February 16, 1951 to February 24,
1951, and there is no reason to doubt it was a single individual that arrived
a long while (about three weeks) ahead of schedule. Similarly, on the small
pond on the Laboratory grounds that we call the Carrie Dam, a male Ring-
necked Duck was present from February 19 to February 27. Again, the bird
was far ahead of normal schedule.

Oneida Dam is not shown on the present Butler, Pa., quadrangle of the
U. S. Geological Survey (Fig. 1). This map was published in 1911, and re-
printed in 1940, from surveys in 1908-1909. The map does show two other
dams belonging to the same system, supplying water to the city of Butler
and owned by the Butler Water Company. These other two dams are Boyds-
town and Thom. Oneida lies on the same watercourse as Boydstown and
extends from the foot of the Boydstown spillway to a point about a mile
south. Though most of Oneida Lake lies northeast and southwest, or nearly
so, the dam breast lies where the stream bends to the west, and the breast
itself lies almost north and south. In the spring the Bonaparte Gulls seem
always to come in at the north end of the lake, follow it south and then
west as they pass over the breast. The level of the lake at the overflow is
1058.8 feet,* and the hills around it rise to 1300 feet. The country is shale

♦Information supplied by P. H. Dowdell, Manager of the Butler Water Company.

1957 Preston: Waterfowl of western Pennsylvania Skyways

81

Fig. 1. Map showing Oneida Dam and other features, superimposed on Butler, Pa., quadrangle of U. S. Geological Survey

82 Annals of Carnegie Museum vol. 35

and sandstone of the Carboniferous (Pennsylvanian) system, and the lake
lies well outside the glaciated area, about half way between the Wisconsin
terminal moraine and the Allegheny River. Its position is indicated on
the map.

The dam was constructed in 1918. In the intervening years a good deal of
fine mud has collected in the upper, or north, end. This is a fine, sticky
clay resembling the glacial muds of Muddy Creek, Butler County, and is
many inches thick. The bottom of the south end of the lake, on the other
hand, is sand and shale, much easier for human beings to walk on when
the water is low, but much less attractive to shore birds. When the lake is
full,* in spring, the dabbling ducks are, naturally, found almost entirely at
the north end. The diving ducks are predominantly in the middle reaches,
while mergansers, loons, and the like may be found over this same area and
all the way to the breast of the dam. Fresh-water clams (Anodonta grandis)
are plentiful in the mud, and painted turtles ( Chrysemys picta marginata)
are present in modest numbers. Years ago, there were some very large snap-
ping turtles ( Chelydra serpentina ). Spring peepers ( Hyla crucifer crucifer)
seemed to be the diet of an American Egret that appeared on April 15,
1951, and stayed a day or two, and other species of frogs are known to be
present, though no inventory of the amphibians and reptiles has been taken,
and it is not known whether they form any significant part of the food of
the birds. John M. McCormick observed bitterns feeding on frogs, and it is
possible that the Little Green Herons use them to some extent. The fish
include bass (some very large), and bluegills, sunfish, perch, and carp. Loons,
mergansers, and Greater Yellow-legs have been seen to catch some of these.

Dr. O. E. Jennings (with Dr. L. K. Henry) examined the exposed mud-
flats on November 30, 1949, when the water was quite low. He divided the
flats into three zones, beginning at “high-water mark,” and continuing on
into the water itself (zone 4). He reports the flora as follows:

Zone 1. Ordinary land plants. These evidently not much disturbed by
inundations, which were probably of brief duration.

Zone 2. Rather firm mud apparently not covered by water since spring
or early summer. A sedge ( Scirpus pauciflorus) about 1-2 inches
tall, forms extensive brown patches over the whole zone, with
occasional plants of an Eleocharis (another member of the sedge
family) commonly known as spike-rush; and also occasional
plants of the low cudweed ( Gnaphalium uliginosum), a common
plant of muddy ditches.

Zone 3. This area was probably covered with water until late summer
or early autumn. At the time of this visit it was soft mud. This
area had rather extensive patches of the aquatic shallow water
Naias ( Naias flexilis), and occasional plants of the water-weed
( Elodea canadensis ), sometimes listed as Anacharis or Philotria.

Zone 4. In the water there were in places dense growths of the water-
weed (Elodea canadensis) and a few pondweeds, mostly Pota-
•Capacity (original) 589,000,000 gallons. The pH of the water through the
pipe-line at the filtration plant varies from 5.6 to 7.3; no information on its
value in the lake itself. At the same point, the temperature varies during the
year from 36® to 78® F.

1957 Preston: Waterfowl of western Pennsylvania Skyways 83

mogeton diver si folius. At the inlet at the upper end of the lake
the water was choked with masses of these plants, mostly Elodea.

Dr. Jennings adds that on an earlier visit, in August, he found that the
pondweed (Potamogeton) had been fruiting abundantly. There is, therefore,
some food for even vegetarian fowl, but Oneida seems to be less a feeding
ground than a resting ground, and the ducks are often observed to be float-
ing asleep, or sitting asleep on the ice or the shore.

OBSERVATIONS HERE REPORTED

Although I have some acquaintance with Oneida extending back almost
thirty years, the observations here reported are confined to a four-year
period, the years 1949 to 1952, inclusive.* In these years the work was essen-
tially quantitative. An attempt was made, on most occasions, to get an exact
count of every species that was present. Usually that was not difficult, for
the waterfowl were sometimes less than a dozen individuals, and only rarely
did they exceed a hundred. None the less, it was not always possible, perhaps
because the birds were continually flying up and down the lake, or diving
a great deal, or because the light was poor, or for other reasons. It is believed
that very few mistakes in identification have been made, and that these
few are probably not important in the general picture. Visits could not be
made every day, and the period of observation on any one day was usually
not more than 30 to 60 minutes. The object was not to ascertain every bird
that visited the Dam in the course of a day, but to get an “instantaneous”
sample of the skyway pool. On some days the picture changed from hour to
hour, but when more than one visit was made on any given day, only one
visit is taken into account for our present purpose. On most occasions the
lake was examined with good binoculars and with a 20-power spotting tele-
scope. For the most part this was done simply from the west bank, where
Pennsylvania state highway 38 passes close by the shore, and usually in the
late afternoon, as the light is then best.

The observations are almost entirely mine, but I had the company on
many occasions of J. M. McCormick in 1949 and 1950, of young James
Glenn, in 1951 and 1952, and of Mrs. Preston in all four years. In March
1951, when it was certain that I must be away for several weeks, the late
Howard Elliott undertook to visit the Dam and keep a quantitative account,
and his results for the last three weeks of March in that year are included
in most of what follows. Except for this, no results are included except
those obtained when I was personally present.

The spring migration % The object here was to determine the duration
of the migration, for ducks only ; to determine the peak of that migration,
and to estimate the most likely number of individuals that would be seen
on any given date within that period. It is shown that all this can be repre-
sented fairly well by a Gaussian “normal” curve. A further object is to
determine the most likely number of species to be found on a visit on any
day within this period. For this purpose the species are expanded to include

*A few notes of later date added while this article was in proof.

84

Annals of Carnegie Museum

vol. 35

the ducks, geese, swans, loons, grebes, coots, gulls, terns, egrets, and herons.
Excluded (on the spring migration, for this section of the investigation) are
shore birds, ospreys and eagles, kingfishers, and passerines. This arrange-
ment is somewhat arbitrary, but some definition of “waterfowl” is necessary.
It might have been well to exclude the Little Green Heron, since it is not
part of the skyway pool, but a summer resident. However, this was not done,
and it makes only a slight difference.

Fig. 2 shows a scatter-plot of every available observation for the four years
during the spring migration, giving the total count of ducks on the lake at
each visit. In this figure the abscissa represents the date, regarded as so
many days after January 31 of each year. The ordinate is the number of
birds seen on each date, plotted logarithmically. This is a convenient way
to plot numbers that fluctuate widely, and where the significant thing is

200r-

o = 1949
x 3 1950
® = 1951
A= 1952

100

90

80

o

70

A

z

tr

60

a.

50

A^ °

CO

oA ®

w

40

-

°o o

Q

UJ

>

30

-

®

o

x ° A O

O'

Id

%

®

<]

o

X

<1

CO

CO

20

X ^ ® O

o

°A

CO

o

® o

®

A O

3

o

10

9

_

® o

A

A *°

Ax®

OX

© A

2-

S1NGLES-*

i i I
10 20 28
FEB.

J J L

12 22 31
MARCH

I I ! —

! I 21 30

APRIL

-La-

11 21 3!

MAY

_L_ 1

10 20
JUNE

ZEROS-* A Ax Ax A AA ® ® A

Fig. 2. Scatter-plot. Observed numbers of individuals (ducks only) seen
on spring visits in the years 1949-1952. The vertical scale is logarithmic

AVERAGE NUMBER OF INDIVIDUALS (DUCKS)
OBSERVED PER VISIT

1957 Preston: Waterfowl of western Pennsylvania Skyways 85

not the arithmetical difference between two observations but the ratio be-
tween the two.

Although there is much scatter, it is immediately obvious that a first
approximation to the facts would be a parabolic curve with its axis early in
April and its vertex at a few dozen ducks. Such a curve is logical. When
restored to arithmetical (instead of logarithmic) ordinates, it will be a
Gaussian “normal” curve. Accordingly, it seemed worth while to fit a
parabola to the logarithmic plotting, by the method of least squares. In
order to reduce the amount of work, all the observations over a 10-day
period (the same 10-day period in each of the four years) were averaged,
as in Fig. 3, and it is to these points that the parabola, the transformed
Gaussian curve, is fitted. Fig. 4 shows the same data as Fig. 3 with the loga-
rithmic scale eliminated and the ordinates plotted arithmetically. The curve
now takes the “normal” or Gaussian “bell-shaped” form.

( NUMBER OF DAYS AFTER JAN. 31 )

Fig. 3. Averages of 10-day intervals, for the same data as in Fig. 2, plotted
on the same logarithmic scale

86

Annals of Carnegie Museum

vol. 35

The tentative inference from this is that the overhead traffic in the
skyways pool is heaviest on April 6 (a date close to the week-end usually
chosen by the Audubon Society of Western Pennsylvania to visit Oneida),
but that there is a wide spread in the dates at which a fair number of ducks
may be seen. In slightly more mathematical language the standard devia-
tion, cr, is 24.1 days, so that one-quarter of the ducks will have passed
through 16 days ahead of April 6, and one-quarter still remain to pass
through 16 days after April 6.

Fig. 4. Data of Fig. 3 with logarithmic ordinates eliminated

Now although this is perhaps as logical and satisfactory a “graduation”
as can be made of the data relating to total individuals, the picture is not
at this stage complete, or might be regarded as of less than average interest,
because the scaup duck (presumably in nearly all cases the Lesser Scaup)
forms such a large proportion of the total population.

In Fig. 5, we give the scatter-plot for number of species observed instead
of number of individuals. Here we have used arithmetical ordinates, but in
Fig. 6 we have graduated the logarithms, and then converted back to arith-
metical ordinates in Fig. 7. The data can be fitted reasonably well to a
Gaussian normal curve, as before, and its crest is at a not very different
date. Since the number of species of ducks proper observable on any one
day is likely to be low, we have here taken waterfowl in a somewhat broader
sense, including geese, swans, grebes, loons, coots, gulls, terns, herons, and
egrets, but excluding shore birds, ospreys, eagles, and kingfishers. The king-
fishers are properly excluded, being summer residents. The exclusion of the

1957 Preston: Waterfowl of western Pennsylvania Skyways

87

others is somewhat arbitrary, since only the Spotted Sandpiper and Killdeer
are summer residents. The exclusion of ospreys and eagles makes little dif-
ference, since there is only one species of each, never more than one in-
dividual, and rarely that. The shore birds also are very few in species and
individuals in the spring, because the lake is full and there are no mud-flats.

There is some slight possibility that the curve is bimodal, involving a
preliminary small migration in February (as the ice begins to melt, on a
temporary basis, at the north or inlet end, which frequently freezes), and
then the main migration starting in early March. It is only the second

17-

16

15

o

> 14
or

Ixl

£ 13

O

12

$

£ n

DC
111

£ 10

8 -

o 7

UJ
CL

40 6

5 -

or

iii

co 4

2

3

z 3

2 -
I -

xO

XX x

O x©

x x ®x O

O O O x

XX A A A OO XCX

A» <® ® O g SSgO X
X X Q£> A O OO ®

XX QDOGD AO® X

A ® j

A ADX

©OX® A A A

X ®©® O © ® A

o * 1949
x = 1950
® = 1951
1952

<38xx

O © ®

X A

J aJ-

J 1 t

10 20 28 12 22 31 II 21 30 II 21 31

FEB, MARCH APRIL MAY

Fig. 5. Scatter-plot of observed numbers of species of waterfowl in spring,
1949-1952. Arithmetical ordinates

88

Annals of Carnegie Museum

vol. 35

phase that contains scaup ducks. The earlier one contains a substantial
number of species, such as the Golden-eye (the “ice -duck” of the Norwegians),
Red-breasted Merganser, American Merganser, Ring-necked Duck, Redhead,
Canvas-back, Mallard, Black Duck, and Gadwall, as well as a gull or two.
The Baldpate, though it has been recorded more than once in February
on the Laboratory grounds, has not been recorded at Oneida till March 1,
which is also the first date at Oneida for scaup (two individuals), but there
is no other Oneida date for scaup earlier than March 11.

Since observations are relatively few early in March, because I am regu-
larly out of town for some days, the bimodal aspect should not be emphasized.

(FEB. 1 = 1) DAYS AFTER JAN. 31

Fig. 6. Observations of Fig. 5 reduced to 10-day averages (circles). Loga-
rithmic ordinates

Spring migration of the scaup. Fig. 8 shows a scatter-plot of the four-
year observations on this species, believed to be mostly Lesser Scaup, though,
since both species (greater and lesser) have been seen at the Laboratory
grounds, it is most likely that both have been seen at Oneida. Fig. 9 reduces
the observations to 10-day intervals and fits a parabola to them, while
Fig. 10 removes the logarithmic feature. The vertex seems to be at about
April 13, which is later than for ducks in their entirety, and agrees with
the qualitative observation that the scaup does not appear in February or
even early March, is the last bird to linger into late May, and that big flights
go through late in the season.

1957 Preston: Waterfowl of western Pennsylvania Skyways 89

Todd (1940, p. 103) however says, “The Lesser Scaup Duck moves north
very early in the season,” (italics mine). Todd's work is so painstaking and
thorough that any phenomena that seem to disagree with his findings should
be examined with extreme care to see what the meaning of the discrepancy
may be. There is no question here that his findings are right, in the sense
that at least some scaup do move north very early. Todd’s earliest date is
February 26, but at Toledo, Ohio, which is farther north than Oneida, it
is sometimes much earlier even than this. Thus, on February 7, 1953, in
company of Mr. McCormick, I found more than a thousand scaup on Lake
Erie near Toledo. They were almost all males. We estimated that males
outnumbered females at least twenty to one. I reported the matter to
Harold Mayfield, who said that the birds must be spring migrants, since
nothing like that number of scaup had been present during the winter.

>

(FEB. 1= I) DAYS AFTER JAN. 31

Fig. 7. Observations and curve of Fig. 6 with logarithmic scale eliminated

The situation that develops then is that at Toledo the scaup is a very
early spring migrant, whereas at Oneida it is one of the latest. None the
less, at both places there is a remarkable predominance of males in the
early movements, and perhaps later.

Again, on February 3, 1953, in the company of Dr. E. S. Thomas, of
Ohio State Museum, and Dr. Milton Trautman, I found at Buckeye Lake,
Ohio, a few scaup, females as well as males.

90

Annals of Carnegie Museum

vol. 35

Concerning this species Trautman (1940, p. 194-195) says, “The first
transients (at Buckeye Lake) arrived by March 5,” but since, as above men-
tioned, he was present on February 3, 1953, he agrees that sometimes the
first transients are now more than a month earlier.

Fig. 8. Scatter-plot. Observed numbers of scaup in spring, 1949-1952.

Logarithmic ordinates

1957 Preston: Waterfowl of western Pennsylvania Skyways 91

All these observations confirm and even emphasize Todd’s comment that
the scaup is very early, in the sense that the first transients of the species
are very early, earlier even than he and Trautman have recorded perhaps;
but so far as the average scaup duck is concerned, the Oneida evidence is
conclusive that there, at any rate, the peak migration is later than that of
most of the other ducks.

( FEB. 1 = 1) DAYS AFTER JAN. 31

Fig. 9. Observations of Fig. 8 reduced to 10-day intervals and fitted by

a parabola

The interpretation, in mathematical terms, of our joint findings would
seem to be this: the scaup is by far the commonest of all our ducks, as
Table 1 shows. Therefore, although the peak of its migration is late, and
the standard deviation of its coming does not appear to be great (18.4 days),
its very numbers give us a better chance to see something of the ends of its
Gaussian distribution, and in particular to see some scaup very early.

92

Annals of Carnegie Museum

vol. 35

Our own findings, that the Lesser Scaup is on the average a late, not an
early, spring migrant agree with those of Kortright (1943, p. 259), “In the
spring their return north is . . . tardy,” and again, “These Scaups are late
breeders, next to the White Winged Scoter probably the latest.”

{ FEB. 1*1) DAYS AFTER JAN. 31

Fig. 10. Data and curve of Fig. 9 with logarithmic scale eliminated

The fall migration. Ducks. As already mentioned, Oneida attracts fewer
ducks and waterfowl in the fall, and more shore birds. The greater attraction
for shore birds is obviously connected with the drop in the water-level, so
that there is less acreage of water, and many acres of exposed mud-flat.
However, there are probably several other factors as far as the ducks are
concerned. Lake Erie, Pymatuning, Conneaut Lake, and Edinboro Lake lie
from 60 to 100 miles north. Weary birds from the north would probably
rest there, and when taking off again, fresh and revived, would not be
attracted to Oneida. Again, in good weather, and the fall is predominantly
good weather, the winds are in the west or northwest, and the birds are
not likely to run into a “front” that forces them down. Thus the “net” is
less efficient in the fall. Obviously the ducks are more numerous, somewhere,
in the fall than in the spring, for the spring migration consists of the same
birds that came south in the fall, minus those that perished over the winter.

1957 Preston: Waterfowl of western Pennsylvania Skyways 93

It is possible that the fowl use a somewhat different flyway in fall from that
which they use in spring, and the fall one may not pass over Oneida; but
this is conjecture and seems much less likely than that the dam is less attrac-
tive and efficient in the fall.

Years ago I thought that the relative scarcity of ducks on Oneida in the
fall, as compared with the spring, might be because there is a hunting season
in the fall. Later I came to think that it might be due to the normally much
shrunken condition of the autumn lake. It does not seem likely that the
latter hypothesis can be right, because F. D. Walker, of Edinboro, Pa., re-
ports that on Edinboro Lake, some 70 or 80 miles north of Oneida, the
ducks in spring are three or four times as numerous as in fall, though the
lake is full of water at both seasons. It is a lake comparable in length with
Oneida, but considerably wider. Nor does it appear likely that hunting has
much to do with it, because the hunting season does not open till the crest
of the migration has passed. The birds can hardly be supposed to be avoid-
ing the hunting season.

We seem therefore to be left with some such hypothesis as the one that
assumes that these lakes are not so favorably located for the fall migration
as for the spring one, possibly due to their proximity to Lake Erie, as pre-
viously mentioned, or else that, on the spring migration there is some bar-
rier, climatic or constitutional, to the birds’ pressing on to the north, while
there is no such barrier to their southward movement in the autumn.

Fig. 11 shows the scatter-plot. Fig. 12 graduates it to a parabola, and
Fig. 13 eliminates the logarithmic ordinates.

The average number of birds seen per visit was approximately 8, com-
pared with 28 in spring. Once more, the figure refers to ducks only, and
excludes swans, geese, and other water-birds.

Sex ratio of the spring migrants. In my notes for February 20, 1951, is
a comment to the effect that Mallards seem to come north in mated pairs,
but most of the other species are predominantly unmated, and that the great
majority of the birds are males. I have found occasional flocks of females,
but the general picture seems to be that males are far more numerous.
Whether this means that males really do outnumber females by a huge
margin, or whether it means that males use Oneida in spring much more
than females do, is not clear from these observations alone. At one time I
assumed that the males went north to establish territories, and, running
into adverse climatic conditions by starting too early, were forced to spend
a little time at Oneida. The females, presumably following later, would not
meet with the same obstacles and would pass right over. Possibly a count of
sexes at various times on the wintering grounds in Florida or elsewhere
might throw some light on the subject.

Davis (1952), quoting Hochbaum (1944), says that ducks form mated pairs
in the south before leaving their wintering grounds. This seems to be true
of Mallards and Black Duck, but the diving ducks at Oneida in spring do
not give that impression. Some diving ducks do appear in pairs, it is true,
but the big flocks seem to have an overwhelming preponderance of males.
The sex composition of the flocks was not always recorded, but very often
it was, and a few examples are given below:

94

Annals of Carnegie Museum

vol. 35

1949. Spring. Red-breasted Merganser . March 14; 5, all males. April
10; 7, 6 males and 1 female. April 27; 13, all females. April 28; 9, all females.

1949. Spring. Bufflehead. March 27; 6, all males. April 5; 12, 6 males
and 6 females.

1950. Spring. Redhead. March 1; 44, 34 males and 10 females. March
12; 4, 3 males and 1 female. March 14; 10, “mostly males.” March 15; 7, 6

ZEROS X AA A 0 9 9 A A AAA 9 9 0

Fig. 11. Scatter-plot. Observed numbers of individuals (ducks only) seen
on autumn visits in the years 1949-1952. Logarithmic ordinates

1957 Preston: Waterfowl of western Pennsylvania Skyways 95

males and 1 female. March 16; 8, 6 males and 2 females. March 19; 4, all
males.

1950. Spring. Scaup. March 12; 7, 6 males and 1 female. March 14; 5,
all males. March 15; 10, all males. March 16; 7, all males. March 19; 10, all
males. March 24; 11, 10 males and 1 female. March 29; 12, 8 males and 4
females. April 11; 33, “mixed sexes.” April 13; “18 scaup and Ring-necked
Ducks, mostly males”.

1951. Spring. Golden-eye. Males and females often appeared sepa-
rately; when together, the males outnumbered the females. The species was
present intermittently from February 16 to April 13; males from February 16

Fig. 12. Observations of Fig. 8 reduced to 10-day intervals and graduated

by a parabola

96

Annals of Carnegie Museum

vol. 35

to March 4, females from February 17 to April 13. Maximum number of
females on any one day, 2; males, 6. Total observed males 37; females, 12.

1952. Spring. Scaup. March 13; 46 Ring-necked and scaup duck, 40
males and 6 females. March 21; scaup, 17, 16 males and 1 female. March
25; 7, 3 males and 4 females. March 29; 14, 7 males and 7 females. March
30; 14, “mostly males.” March 31; 20, 18 males and 2 females. April 6; 34,
“mixed sexes.” April 13; 62, “mostly males, but some females”.

1952. Spring. Bufflehead. March 30; 23, 18 males and 5 females.

1950. All species of genus Aythya. March 1 to April 13. Approximately
175 males and 50 females. One-third of the females did not appear till April,
but only about one-fifth of the males appeared in that month.

There is a definite statistical pattern here, but its biological meaning is
obscure.

( AUG. 6=1) DAYS

Fig. 13. Data and curve of Fig. 8 with logarithmic scale eliminated

Trautman (1940, p. 194-195) has noticed the early preponderance of male
scaup at Buckeye Lake, Ohio. He says, “The March flocks contained more
well-marked males than females or males in changing plumage; early April
flocks were rather evenly divided between these classes; and in late April
and thereafter the flocks contained many more females and immature males
than well-marked males”. These statements are undoubtedly true, but they
leave the impression that, averaged over the whole spring migration, one
sees females about as often as males. This is not the case. Both Dr. Trautman
and Dr. E. S. Thomas (1953), who also is very familiar with Buckeye Lake
and was frequently with Dr. Trautman there, agree that, averaged over the
season, observations of males outnumber those of females at least four to
one; and further, they are inclined to put upon it the same interpretation as
mine— that the males, coming early, dawdle by the way and spend four
times as long in going through as the females do, when the latter come later
and keep going.

Todd (1904, p. 520, and 1940, p. 103) has no very definite observations on
this sex ratio question, nor does Sutton (1928, p. 74). The phenomenon is
none the less a very striking one.

Shore birds. The fall of 1949 was a good time for shore birds at Oneida,
and led me to suppose that many species might be seen by careful watching

1957 Preston: Waterfowl of western Pennsylvania Skyways 97

in subsequent years. In 1950, however, the summer and fall were relatively
wet, and the dam was not drawn down enough to attract shore birds. In
1951 it was low, and in 1952 very low, but the count of shore birds, though
higher than in 1950 (when there were none) did not reach the 1949 level
by any means.

Shore birds are not so easily counted as ducks, and diving ducks, when
diving, not so easily as surface-feeding ducks or sleeping ducks. So long as
the numbers are not large, however, or the birds too active, two or more
observers can always agree about ducks. A good test was provided by the
swans of April 1, 1952. Mr. McCormick counted them and I counted them.
His count was 131, and mine, 133, on a single trial, and considering that
the birds were moving about somewhat actively, this was satisfactory, and
we thought we could compromise amicably on 132. But the same number
of shore birds, unless sleeping, would be very hard to estimate with a
probable error of less than 10 per cent. Killdeers are the birds that appear
in greatest numbers on the mud-flats, and they are continually flying about,
so that it is impossible to keep accurate track of them. In consequence, in
the present paper, no quantitative work is reported on shore birds.

Species-composition of the spring migrants , excluding gulls , terns , and
shore birds. The faunal assemblage as seen at Oneida is tabulated below.
It may not agree exactly with some of the other tabulations, since I fear I
have not been absolutely consistent throughout this paper in admitting, or
excluding, observations by Howard H. Elliott, and may have made other
minor adjustments in one section or other of the paper. There is no reason
to suspect, however, that such discrepancies affect the validity of any con-
clusions that are here drawn. In the present section, only birds seen at
Oneida are recorded, and Boydstown, Thorn, and the Laboratory grounds
are rigorously excluded. Some of the observations are by Mr. Elliott, but the
great majority are my own. The report covers the bird-days for each species
as accumulated over the four-year period, on observations of the spring
migration only.

TABLE 1. OBSERVATIONS OF SPRING MIGRANTS

Common Loon

... 109

Green-winged Teal

3

Horned Grebe

... 68

Blue-winged Teal

100

Pied-billed Grebe

... Ill

Shoveller

2

Double-crested Cormorant ...

1

Wood Duck

12

Great Blue Heron

... 15

Redhead

195

American Egret

... 4

Canvas-back

46

American Bittern

... 2

Ring-necked Duck

170

Least Bittern

1

Scaup

1253

Coot

... 45

American Golden-eye

95

Whistling Swan

... 167

Buffle-head

103

Canada Goose

... 174

Old-squaw

11

Mallard

... 47

White-winged Scoter

3

Black Duck

... 240

Ruddy Duck

97

Gadwall

... 24

Hooded Merganser

78

Baldpate

... 403

American Merganser

18

Pintail

... 131

Red-breasted Merganser

178

98

Annals of Carnegie Museum

vol. 35

A few interesting points emerge from this table. Although the table
accounts for only 3209 ducks (excluding geese and swans), the proportion
of scaup, 1253, is still just about the same percentage as in previous tables,
viz., 39.1 per cent.

Black ducks outnumber Mallards five to one. Since the Black Duck is
primarily a bird of the eastern seaboard, and the Mallard of the interior,
Oneida seems here to be definitely in the Atlantic flyway where these two
species are concerned. From observations made in the company of Dr.
E. S. Thomas near Columbus, Ohio, I get the impression that there, only
200 miles west of us. Black Ducks and Mallards are about equal in num-
bers, and Dr. Thomas believes that the immigration of the Black Ducks
is a recent phenomenon.

The Red-breasted Merganser is ten times as common as the American
Merganser (Goosander), and twice as common as the Hooded Merganser.

The Ring-necked Duck and the Redhead are about equally common
(170 and 195 respectively), and the Ruddy Duck is about as common as
the Buffle-head (97 and 103), or as the Golden-eye (95).

Although the Black Duck is so much commoner than the Mallard, it
is not the commonest of the pond ducks. That distinction belongs to
the Baldpate (403 vs. 240), but this is due to the Baldpate often appearing
in very large flocks, occasionally of 60 or 80 individuals. The Black Duck
is seen more often, but the largest flocks observed were of 16 and 20
individuals. Baldpates were seen (in spring) on only 34 occasions, while
Black Ducks were observed 57 times.

With those species that move in large flocks, a single observation, made
or missed, can make a vast difference in the count. Thus of the 167
swan-days reported, 157 are due to the single large flock of 132 birds
forced down by a line-squall, as previously mentioned. Half of the goose-
days are due to a single flock of 85 birds observed by Howard Elliott
one evening.

Superficially, there is a great difference from year to year, even with
those species where the total count is not due primarily to a few large
flocks. Thus the annual spring count for scaup goes: 140, 466, 353, 294;
but for Ring-necked Ducks the progression is quite different, viz. 81, 52,
16, 21.

The count for Black Duck was 77, 99, 57, 7; and for Mallards, 12, 19, 16, 0.

These differences do not seem to be due to the relative frequency of my
visits to Oneida, and it is not clear that they are due to timing, though
both may play some part.

Black Duck seen per visit (counting only those visits on which at
least one was seen) come out as follows: 1949, 4.5; 1950, 4.7; 1951, 3.8;
1952, 1.7. But it is possible that the differences are not significant, and
it seems improbable, in view of the tendency of this species to move at
times in pairs and at others in flocks, that any simple statistical treatment
would help.

The Common Loon. It is likely that, even when a species is not ob-
served on more than a few dozen occasions, or in numbers exceeding

1957 Preston: Waterfowl of western Pennsylvania Skyways 99

a few score, a detailed count on each occasion may permit the testing
of various hypotheses or current notions on how the bird behaves.

This has been done in some detail in a separate paper (Preston, 1956)*,
and it is therefore unnecessary to do more than summarize it here.

The Common Loon is an extremely late migrant in spring, lagging far
behind the ducks as a whole and even behind the Lesser Scaup. The peak
of its migration is at April 22. Further, this species migrates much more
nearly simultaneously than ducks as a whole, or even than the scaup.
The standard deviation of the timing is 11.6 days for the Loon, 18.4 for
the scaup.

This suggests that the Loon does not start north till its breeding grounds
are ready for it; and when it starts, it goes right through. Thus the Loon
is probably commoner in the skyways than its numbers, compared with
the scaup, would suggest. With a cruising speed, unhurried, in excess
of sixty miles an hour (Bent, 1946; Preston, 1951), it does not need to
dally at Oneida. We see few of them unless they are forced down by fog
or rain.

Normally the Loons we see at Oneida are not organic flocks, but random
assemblages. At the peak of the season the most likely number to be seen
is three, and they are usually well separated and paying no attention to
each other. There is no evidence in the field that any of them are mated
pairs, and a mathematical analysis of the subject suggests that they are
moving as individuals, not as pairs.

Table 2 gives a conspectus of the migrant water-birds seen, in spring
and fall (S and F) for the four years 1949 to 1952 inclusive. It is quantita-
tive only as to species. Some species may be represented by only a single
individual, as with the Northern Phalarope, and others by more than
a hundred, as with the Whistling Swan on April 1, 1949.

A few comments on Table 2 may be useful. First of all, some species
are conspicuous by their absence, and others have not shown up to advan-
tage in this four-year period. Thus, I once saw a large flock of Snow
Geese on Oneida under the thick mists of early morning, and Mr. Norris
has reported a large flock of Snow Geese flying north over the Laboratory
grounds. On November 16, 1948, J. M. McCormick found two immature
Blue Geese at Boydstown Dam, immediately above Oneida, and it is
reasonable to suppose this species occasionally visits Oneida. Only one,
slightly uncertain, record is given for Oneida itself, the bird disappearing
before I could be absolutely sure of it. Again, only one record is given
for the Caspian Tern, in the spring of 1950, but another individual (or
the same one) spent some time there in the fall of 1948 (September). The
Dunlin, or Red-backed Sandpiper, was present in the fall of 1948, too.

The Shoveller is recorded on Oneida only in 1950 (spring); but it oc-
curred on other, smaller, local ponds once or twice in this four-year period,
and I saw it in greater numbers on Oneida a few years earlier. The Canada
Goose probably occurs every spring and most falls, but it is somewhat
easily scared out and probably remains only from dusk to dawn. On one

•Charles S. Pearce, Manager of the American Ceramic Society, of Columbus,
Ohio, says that paper should be entitled “The marital status of migrating loons”.

100

Annals of Carnegie Museum

TABLE 2. MIGRANT WATER BIRDS

1949 1950

Species S F S F

Common Loon x x x x

Red-throated Loon x

Horned Grebe . . x x x

Pied-billed Grebe x x x x

Double-crested Cormorant x

Great Blue Heron x x x

American Egret

Little Blue Heron x

American Bittern x

Whistling Swan x x

Canada Goose x

Blue Goose x

Mallard x x x x

Black Duck x x x x

Gad wall x

Baldpate x x x

Pintail x x x

Green-winged Teal x

Blue-winged Teal x x x

Shoveller x

Wood Duck x

Redhead x x

Ring-necked Duck x x x

Canvas-back x

Scaup x x x

American Golden-eye x x

Buffle-head x x x x

Old-squaw x x

White-winged Scoter

Ruddy Duck x x x x

Hooded Merganser x x x

American Merganser (Goosander) x

Red-breasted Merganser x x x

American Coot x x x x

Semipalmated Plover x

Killdeer x x x

Wilson’s Snipe

Solitary Sandpiper

Greater Yellow-legs x x x

Lesser Yellow-legs x x

Pectoral Sandpiper x

White-rumped Sandpiper x

Baird's Sandpiper x

Least Sandpiper x

Red-backed Sandpiper (Dunlin) x

Stilt Sandpiper x

Semipalmated Sandpiper x

Northern Phalarope x

Herring Gull x x

Ring-billed Gull x

Bonaparte’s Gull x x

Common Tern x

Caspian Tern x

Black Tern x

Forster’s Tern x

1951

x x

X X

X

X X

X

X X

X X

X

X X

X X

X

X X

X X

X X

X
X

X X

X X

X

X X

X X

X X

X

X X

X X

X

X

X

X

X

vol. 35
1952

X

X X
X X

X

X X

X

X X

X

X

X X

X

X X

X X

X

X

X X

X X

X

X X

X
X
X
X
X
X
X

X

X

X

X

* * H K X X 05

1957 Preston: Waterfowl of western Pennsylvania Skyways 101

day in the quadrennium, Howard Elliott saw a large flock come in at
dusk half an hour after I had left Oneida. In 1952 a single individual came
in to the Laboratory grounds at 4:30 P.M. and stayed into the night but
was gone early next morning. The Black-crowned Night Heron presumably
occurs regularly at Oneida, but is abroad only at night, when I am not
there. I have seen the species at Thorn Dam at dusk, and at the Laboratory
grounds for weeks on end.

Forster’s Tern is a new species for western Pennsylvania, and Mr. Todd,
properly, rejects the observation. I did not at the time realize that it had
never been seen here before, or I could perfectly well have collected it, but
since Mr. McCormick and myself are both completely satisfied, I can not
very well reject my own records. The bird may be seen in some numbers
at Toledo, Ohio. The Black Tern is probably more frequent than the
records indicate. It has several times been seen on the Laboratory grounds
and at even smaller ponds.

Bonaparte’s Gull seems to be regular in spring, but absent in the fall.
The other gulls and terns are seen at both seasons. The Canvas-back
and Old-squaw also seem to be spring migrants, while the other ducks,
geese, and swans are likely to show up in the fall also, though not in
such numbers or with equal certainty as in the spring.

I have expected the Gray (Black-bellied) Plover and the Golden Plover
to show up in the fall on the mud-flats, some of which are wet and some
fairly dry, but I have never seen either species, either in the quadrennium
or earlier. They occur in some numbers near Sandusky, and near Columbus,
and in great numbers on the New Jersey marshes, so presumably one of
these days a few strays will reach Oneida. At present I am inclined to suspect
that the skyway over Butler County does not normally contain this species at
all, but there may be other explanations. Near Toledo and Sandusky the birds
are remarkably local, though very plentiful where present at all.*

Of the other birds that might be expected to be attracted to Oneida,
some mention must be made. The Bald Eagle is occasionally there but
only as a bird of passage and very briefly. I have seen it attacked by crows.
The Osprey is much more regular. In fact, in April it is present all the
way from Oneida northeast to New York state. Apparently it catches fish
in the riffles of the streams, from the Allegheny River at the junction of
the Clarion over to the Susquehanna watershed. It regularly appears at the
Laboratory grounds, and catches fish in the Carrie Dam. Occasionally it
is present as late as June.

The Belted Kingfisher is present about nine months of the year, and
nests (or has nested) below the Oneida spillway. The Little Green Heron
is present on all the streams, probably, of the district, and regularly nests
at the Laboratory grounds. It is, therefore, omitted from the tabulation.
The Spotted Sandpiper nests at Oneida, as well as at the Frith (the inclosed
grounds of the Preston Laboratories, three miles due west of the center
of Butler, Pa.), but at the former place it appears to me less common
than it was years ago. This may be because of the great increase in the

•A single individual of the Black-bellied species finally put in an appearance
on September 29, 1953.

102

Annals of Carnegie Museum

vol. 35

number of fishermen that crowd the banks. The Solitary Sandpiper is
more common at the Laboratory than at Oneida, perhaps because it is a
woodland bird. The American Merganser is occasional at Oneida, but
probably regular on our swift-flowing streams, such as the Slippery Rock
Creek in the Main Gorge, and the lower Connoquenessing above its junction
with the Slippery Rock.

Only one American Egret has been seen at Oneida on migration proper
(spring of 1951). Five individuals were present at Oneida from August 8
to September 19, 1948; and, many years earlier, about 1928, I saw about
a dozen of them there. An adult bald eagle that flew over the Laboratory
grounds at a great height on May 13, 1951, headed due north, was perhaps
a Florida eagle headed for the Great Lakes or the Canadian north woods.

I myself have never seen the Florida Gallinule at Oneida, but a bird
of the year was caught there by one of my assistants (Sept. 14, 1951). The
rarity of the bird in Butler County is a great surprise to me, since in
England, under the name of moorhen or water-hen, it is the most widely
met of all waterfowl.

Of the passerine birds there is little to say, since few of these are water-
birds in any real sense. Oneida is a good place to see the early swallows, and
on the cold spring mornings the Carrie Dam in the Frith is a good place,
too. The water supplies the few insects that are abroad when the ground is
chilled to the freezing point but the water is open. The American Pipit is
a tundra bird. In the fall I have seen it in hundreds among the dead tree
stumps of the mud-flats of Mosquito Creek Dam, just over the border in
Ohio. The only place I myself have seen it in Butler County is at Oneida
Dam, on two occasions, once in some numbers. But Howard Elliott and
Earl Schriver have reported flocks of the species in other parts of the County,
so neither water nor mud-flats are essential. Crows are rather plentiful
around Oneida, and in the fall vast flocks of Starlings, with some Red-wings
and perhaps other species, roost in the woods adjoining the lake, but this
has no obvious connection with the presence of the lake itself: it is not the
only Starling roost of the County, nor a principal Crow roost.

The Rusty Blackbird occasionally appears at the Laboratory, but I have
not yet recorded it from Oneida, where it might be expected.

Rarities , unusual dates, etc. Birds not often seen in western Pennsyl-
vania, at any rate not often seen south of the lakes area in the extreme
northwest corner, or reported by Todd (1940) to be infrequently observed
there, are as follows:

The Red- throated Loon (November 15-17, 1949) has been reported very
seldom. One of the earlier records, however, was from Oneida.

The Double-crested Cormorant (April 13, 1950) is quite uncommon, though
one made itself conspicuous along the Allegheny River a few years ago.

The Whistling Swan is common enough, and 1 mention it here because
its flyway appears to be narrow. The bird is common near Youngstown
and on Lake Erie near Toledo, but unusual near Columbus, Ohio. It
would seem that Columbus lies southwest of the flyway.

The Gadwall is fairly regular in its appearance. Todd seems to have
thought it rather uncommon short of the Lakes.

1957 Preston: Waterfowl of western Pennsylvania Skyways 103

Todd says the Long-tailed Duck, or Old-squaw, used to be common on
Lake Erie but has been all but exterminated in the interests of the fisheries.
None the less it has visited Oneida every spring in the four years under
review, though only in small numbers.

The White-winged Scoter is reputed to be regular at Youngstown, but
has been reported only twice at Oneida, both times in the spring of 1951.

The Northern Phalarope has appeared twice, and the 1952 individual
stayed for about a week. The bird is regularly seen near Toledo, but not
in any numbers; it is rare at Buckeye Lake, Ohio. It may be less rare
overhead in Butler County than at present appears.

The Caspian Tern and Forster’s Tern have already been commented on.
The Caspian Tern’s date was very early (April 11, 1950). On the previous
day, driving home from Toledo, I had seen a couple of these birds at a
small pond near Fitchville, Ohio. A very strong west wind was blowing, and
I thought it possible some migrants might be blown into Pennsylvania. I
suggested to Mr. McCormick that we go to Oneida and find a Caspian Tern.
We went, and there was the bird.

The Little Blue Heron was, of course, an immature in white plumage.
The same species, in the same plumage, has been seen on the Laboratory
grounds, but only once.

The White-rumped Sandpiper is represented by a single individual at
Oneida, but again, a single individual was also seen in an earlier year, at
the Laboratory.

The Greater Yellow-legs stays somewhat later than Todd reports. In
1949, one individual was present on November 20, an all-time record accord-
ing to Todd, but 20 were present on November 10, and eight on No-
vember 15.

Three Dunlins (Red-backed Sandpipers) were present on November 20,
1949, again an all-time record according to Todd, but one was present on
November 24, Thanksgiving Day. The Dunlin is a very late migrant at
Toledo, and I think Todd underestimated its tendency to be late on the
fall migration. It is the last of the Sandpipers to go through, and it out-
stays even the Greater Yellow-legs.

Three Killdeers remained till December 4 in 1949, which is quite late
with us.

In 1950, a male Hooded Merganser appeared on March 15, which is a
comparatively early record.

On April 20, 1950, Mr. McCormick and myself found at Oneida two
Greater Yellow-legs with one Lesser Yellow-legs with them, so there was
no question of its size. It is a very early record for the latter species.

On February 16, 1951, I observed four male Golden-eyes and one male
Canvas-back. The Golden-eyes were early, and the Canvas-back three weeks
earlier than anything reported by Todd. The first Ring-necked Duck, a
female, appeared on February 23, 1951, which is very early, but a male had
been at the Carrie Dam in the Frith since February 19.

Eleven Redheads appeared on February 28, 1951, a day or so earlier than
Todd’s records, but agreeing closely with our own date of March 1, 1950.

I

104 Annals of Carnegie Museum vol. 35

The American Egret of April 15, 1951, was a somewhat unusual observa-
tion, spring visitors being rare in Western Pennsylvania.

Pectoral Sandpipers remained at Oneida in 1952 until October 26.

From the detailed records, it is possible to tabulate the first and last
(observed) appearance of each species on both spring and fall migrations
within the period of four years, but, though such information is not too
plentiful, especially as regards last appearances, it is doubtful whether it
has much utility. At the best, it represents only the tail-ends of more or
less Gaussian distributions, which theoretically extend indefinitely to left
and right. There is a good deal of evidence in Todd (1940) that many species
of ducks will stay on Lake Erie all winter if the winter is exceptionally
mild, and a good deal of evidence that many unexpected species will stay
and breed at Pymatuning if they find conditions to their liking. Both these
facts emphasize the Gaussian character of the time “limits” of the spring
and fall migrations. The birds are not set on reaching a particular geo-
graphical region, but only on reaching a satisfactory environment. If we
could maintain a suitable environment all year round, the birds might cease
to migrate at all. This is not altogether surmise. In Britain, with its equable
climate, the birds tend to be much less migratory than the conspecific forms
of continental Europe. The recent tendency of the European White Stork
to nest in South Africa, hitherto merely its wintering grounds, may be an
example of the same sort; and the colonization by the Gadwall of the Outer
Banks of North Carolina (Pea Island), a thousand miles from its normal
prairie summering grounds, indicates the same thing. Migration may be a
deeply ingrained instinct, but it seems to be one that is readily dropped as
soon as it ceases to be a painful necessity.

UTILITY OF THESE OBSERVATIONS

It is highly probable that if Oneida had been watched continuously over
this four-year period, or if it had even been visited every day, a number of
additional species would have been seen, and many more significant observa-
tions made. As it is, it is probable that no small body of water in our region
has been watched equally closely or quantitatively. None the less the period
is short and a decade might have been more useful. However, it is probable
that the picture is changing. There are vastly more small farm ponds on
the Foreland than there were 10 or 20 years ago, and these attract success-
fully many species, including birds as large as geese, and ducks as relatively
uncommon at Oneida as the Shoveller. The number of such ponds will
undoubtedly increase, and some may prove superior to Oneida in their
seclusion, for Oneida is much disturbed by fishermen and other visitors.
When Oneida was new, thirty years or so ago, and less thickly surrounded
by anglers throughout most of the year, it is probable that it was more
visited by waterfowl and waders (shore birds), especially in the fall. Unfor-
tunately, we have no quantitative records, known to me, of those days.

It seems to me that, if Oneida is watched long enough, everything that
has ever been seen at Pymatuning or Erie, in the line of waterfowl, will be
seen also at Oneida, and perhaps some species that have not been seen
farther north. If this is so, it seems clear that the proposed lake in the valley
of Muddy Creek, a few miles to the west, will be a far greater attraction;

1957 Preston: Waterfowl of western Pennsylvania Skyways 105

for Oneida is a reservoir of drinking water, and is not operated with any
eye on the needs of water-birds or the interests of bird lovers or ornitholo-
gists. The lake is small, largely free of vegetation, quite free of islands, and,
on the face of things, quite unattractive to birds. The proposed lake in
Muddy Creek would be many times as large, much more complex in out-
line, with much more extensive shore -lines, and would intercept a much
greater width of flyway.

Whereas no part of Oneida is free from frequent, almost incessant, dis-
turbance from human beings, being completely and closely surrounded by
roads and, for most of the year, lined with fishermen, there is a possibility
that Muddy Creek can, at least in part, be set aside for wildlife.

At Oneida, while with Mr. McCormick, I have seen a fine antlered deer
swimming the lake, and with Mrs. Preston have seen a small herd of deer
come down to drink. On Boydstown, beaver could sometimes be seen, and
there were until recently several beaver dams upstream from Boydstown.
Both species of mammal still occur in Muddy Creek, and there also I have
seen the Bald Eagle, the Bittern, Virginia Rail, American Egret, and many
other species, though at present there is no expanse of water there. The
Marsh Hawk and Red-shouldered Hawk (a swamp species) both nest there,
and I have never seen either at Oneida. The Snow Goose has been reported
on Muddy Creek, and the Wood Duck and Mallard nest there. Howard
Elliott has reported the King Rail nesting at a point not much beyond it,
and it would seem that Oneida is only a very tattered net for sampling
our skyways.

METHOD OF MAKING COMPUTATIONS

Fig. 2 shows the scatter-plot representing all observations, in four spring-
times, of the ducks (excluding geese and swans) observed on the water (not
overhead) at Oneida. Such a scatter-plot really contains more information
than any other plot that can be devised, but it is too complex to be grasped,
and therefore lacks utility. Fig. 1 1 shows a similar plot for the four autumns,
and the same comments apply. It is difficult to make out what is going on.

Accordingly, it is convenient, and useful, for some purposes, to reduce
the number of plotted points to a modest number of average points. This
has been done by taking 10-day periods, the first being February 1 to 10,
inclusive; the second, February 11 to 20, and so on. February 29, 1952, is
treated as non-existent, so that March 1 is in every year treated as the
29th day, and March S as the first day of the fourth 10-day period. Since
observations were not made on every day, it may happen that the average
time of observation during the 10 days does not come at the mid-abscissa of
that period. Its true position is calculated, and the average of the number of
birds seen on each visit constitutes the ordinate for that period. By this
means we reduce the plotted points to fifteen or so, and the general shape
of a well-fitting curve becomes apparent. (Fig. 3, Fig. 12, etc.) In Fig. 3 the
circles are observed points, and the figure suggests that the cut-off in late
May is sharper in practice than the computed curve would imply. However,
very few visits were made in late May and in June, so the evidence is here
incomplete and the computed curve is possibly fairly accurate.

106

Annals of Carnegie Museum

vol. 35

Curves are sometimes fitted on a purely empirical basis. This is not a
particularly useful proceeding. The curve that is to be fitted should be
logical, not merely a good fit; and of two curves that fit reasonably well,
the more logical should be chosen even if it is a slightly poorer fit than
the other. The question of logic involves a reference to the physics or
biology of the problem and can not be decided solely by statistical ideas.

Thus, the fact that we see a period of maximum abundance of ducks on
Oneida near April 6 and again near October 16 does not constitute all
that we know about the skyway pool. We know that the birds in spring are
headed one way (conventionally “north”) and in the fall are headed the
opposite way (conventionally “south”). If we take this vector knowledge into
account, the question we are really asking is: “How full is the skyway of
birds flying north?”, or “How much northing is the bird population over-
head accomplishing per day?” When expressed in this form, the fall migrants
are negative birds, because they are losing “northing.” If spring and fall
migrants are plotted on the same diagram, the fall migrants should be
plotted below the zero line (Fig. 14). We are then led to treat the annual
migration as a single event, and should logically “graduate” the curve as a
sine curve with a periodicity of 12 months. Such a sine curve has some ele-
ments of logic, but the diagram rather discourages it; there is too long a
period in the summer, and too long a period in the winter, with zero birds.
The sine curve would pass through zero twice a year, but the zeros ought
to be short and sharp. The observations themselves accordingly discourage
this concept, and this discouragement is fortified by our knowledge of
their biology. The birds are not flying north or south to accommodate
themselves to the rhythmic movement of the isotherms. They are moving
north to breed and moult, a business that takes time, and so there is
naturally a complete absence of birds for some time in the summer. In the
winter they have to eat, and they will naturally be absent from the frost
and freezing, till open water becomes dependable.

On this basis, the birds’ biological requirements could be met by the whole
of them moving north simultaneously on April 6, and all moving south
on October 16. This would be similar to the daily “rat race” that develops
in Pittsburgh and other large cities, when all the suburbanites crowd into
the down-town districts about 8 A.M. and all rush out at 5 P.M.

The birds also spread or “stagger” their comings and goings over a sub-
stantial period. This arrangement presumably arises from the fact that the
birds are not exactly synchronized in their internal biological rhythm, and
they are not all flying the same routes or starting from the same points.
Accordingly, a reasonable assumption would be that enough variable fac-
tors are present to induce in the observations the appearance of a “normal,"
or Gaussian, “curve of errors.”

The two migrations, spring and fall, are then to be thought of as separate
events, even though one does involve the other. But since the birds do not
necessarily follow the same routes coming and going (a matter we know
for certain with some other species), and since their behavior, and the com-
position of the flocks, is most likely quite different at the two seasons, we

1957 Preston: Waterfowl of western Pennsylvania Skyways 107

are justified in regarding the two events as independent problems, and
graduating each one with a single Gaussian curve.

Consider the spring migration. Let its peak occur at £ days after
January 31. Let the most probable number of birds to be seen on a
single visit at the peak be tj birds. Let us calculate the most probable
number ( N ) of birds to be seen on any other occasion, dated x days from
January 31. (x — £)a

We have N ~~ rj . e (1)

where <r is the “standard deviation" in days. (The standard deviation will
be discussed below.)

Our problem is to fit the observations with a curve of this type so
as to get the best estimates of ??, £ and a. This can be done in various
ways. We have chosen to do it by the simplest and most obvious way,
which is good enough in view of the great scatter in the scatter-plots, though
it may not be the most accurate method.

108

Annals of Carnegie Museum vol. 35

Taking logarithms to base e (natural logarithms), equation (1) may be
written

(* - *)2

i0ge N = loge V

2<r*

or writing y for loge N and k (a constant) for loge 7\ or loge Nmax,
1

y =

2a

i (*2 - 2&e + e)

(2)

we have

(3)

This is the equation of an ordinary (quadratic) parabola in y and x.
It is a parabola with a vertical axis, and its vertex at x = £, y = k.

Accordingly, we can graduate our observations, if we wish, by plotting
the logarithms of the number of birds observed, and fitting a quadratic
parabola among the points. This is a reasonable procedure provided
that we ignore 10-day periods when we saw no ducks whatever. The
logarithm of nothing is minus infinity, and we should have difficulty
in accommodating points at minus infinity.* Fortunately, if 10-day periods
are taken, the average is not zero till we get outside the real migration
period. Our problem then is to fit a simple quadratic parabola

y = A + Bx 4- Cxa (4)

to our observations by the method of least squares; that is, to find the
best values of A, B and C in equation (4) which are related to the
constants of equation (3) as follows:

C = -i/2<73
B = Z/cr*

A = k - |2/2<r£

or

o-s = -1/2 c

Z = - B/2C
k — A — B2/4C
t} z= antiloge k

(5)

Let x, y now denote the co-ordinates of each of the n observed points
(average points in the 10-day periods), so that

S x = xx + x2 -f x3 -f -f xn, and so on.

Then, for best fit, we have the well known result:

An -f J32x + C2x2 =

AVx + JB2xa + C2x3 = 2xy (6)

A 2x2 + 52X8 + C2x4 = 2x2y

whose solution, written in the form of determinants, is simply

A = -B = C = -1 (7)

2x 2x2 —2 y n 2x2 —2 y n 2x —2 y n 2x 2x2

2x2 2x3 —2 xy 2x 2xs —2 xy 2x 2x* —2 xy 2x 2xa 2x8

2x8 2x4 —2 x*y 2x2 2x4 —2 x2y 2x2 2x3 — 2x2y 2x2 2x8 2x4

This permits us to determine A, B and C quickly, and hence to compute
<r, £, and k from equation (5). These quantities, as we have seen, define
respectively the time of the “migration peak,” the most likely number
of birds to be seen at that time (on an “instantaneous” visit), and the
extent to which the migration is spread out over the days.

#Hald (1952) has apparently given an alternative method, but I have not yet
seen this work.

1957 Preston: Waterfowl of western Pennsylvania Skyways 109

It seemed likely that it would be worth while to make these computations
only for a few cases— for ducks (all species lumped together) in spring,
for ducks in fall, and for scaup (both species lumped together but believed
to be chiefly Lesser Scaup) in spring. For other species there are not
enough individuals or enough observations to be very satisfactory.
Observations every day for 10 years would probably provide enough
material for a fair estimate of a dozen or more species of ducks, for
the Canada Goose and Whistling Swan, and for several species of shore
birds.

Ducks in spring

Peak of
migration (£)

April 6

Individuals
at peak (77)

37.11

Standard
deviation (a)

24.1 days

(individuals)
Ducks in fall

October 16

10.89

24.5 days

Scaup in spring

April 13

20.49

18.4 days

Species in spring

April 6

6.74

32.1 days

From this it follows that scaup in spring are just one week late com-
pared with ducks in general (including themselves), or almost two weeks
late compared with all species of ducks other than scaup.

Again, the ducks are equally “spread out” in time on the fall migration
and in spring, as shown by the standard deviations of 24.1 and 24.5 days
respectively.

However, the period in spring that is rich in species is longer than
the period that is rich in individuals, as shown by standard deviations of
32 and 24 days respectively. This may be due in part to our having
included species other than ducks proper in our analysis of the species-
richness.

One other point may be worth noting. The period from April 6 to
October 16 is very close to six months, so that the ducks do in fact divide
their time fairly equally between their summer and winter homes. This
would be a fundamental requirement if we wished to graduate the data
as an annual rhythmic sine curve; the requirement is in fact met, but
the other requirements, previously referred to, are not, and hence we
have rejected this method of graduation.

The “peak of migration,” or Rvalue, is not strictly a local phenomenon.
When a maximum number of birds are observed at Oneida, it presumably
means that a maximum number of birds are shifting from north to south,
or vice versa, and this would appear to be a phenomenon continental
in scope, or at any rate valid for a very large region.

There is another standard deviation (<r2) that might be computed. What
we have computed above is the “spread of the migration,” that is, the
dispersion in time of the birds’ passage. We can, however, also make an
estimate of the spread in numbers likely to be observed at any time
during the migration. For a 20-day period near the peak of migration
this comes out approximately as follows:

110

Annals of Carnegie Museum

vol. 35

Individuals
at peak

Standard
deviation (<r2)

0.6745 c 2
15.6
13
2

Ducks in spring
Scaup in spring
Species in spring

37

21

7

23

20

3

In other words, 50 per cent, of all ducks pass through Oneida in spring
within a period of about 32 days, from March 21 to April 22. One quarter
of them pass through before March 21, and one quarter after April 22.
(Fifty per cent, of the area of a Gaussian curve lies between plus and
minus 0.6745 <r2 from the peak.) And if we make a large number of
random visits to Oneida in the peak periods of spring, half our visits
should result in our observing a population of ducks between 22 and 52
individuals, a population of scaup between 8 and 34 individuals, and a
number of species between 5 and 9.

The high standard deviations ( cr2 ) mean that we may, not seldom,
draw something approximating a blank and on other occasions run into
a great concourse of individuals and species, since an occasion with a
2<r2 deviation is by definition not a really rare event. This high value
is what makes the scatter-plots so hard to interpret until they have been
averaged down to 10-day intervals. The reason for the high values is
probably simple. For the wildfowl to travel it must not only be the
right season but also the right weather, and the weather is highly variable.
The influence of weather upon migration has been studied for many
decades, and is an active subject at the present time [Gunn and Crocker
(1951), Mayfield (1947), Bagg (1950), Bennett (1952), and many others].

Integral of the curve. Suppose we visit Oneida once every day in
the migration period, and count the ducks. We may count the same
individual duck two or more times, if it is present on more than one
day. On the other hand, some ducks may come and go between two of
our consecutive visits, so that we never see them. These two factors tend
to cancel one another, but one may well be much greater than the other,
so that the cancellation is far from perfect. What we finish with in
any case is a daily count of birds, and if we add up all these counts
throughout the migration, we get an estimate of the birds X days or
“bird-days” per season. Call this Q.

Now

- (x - or

2a2

. dx = 1

J-foo

— 00 ri .e

- (x - or

or

. dx = V2 tt . <r . 7f = 2.5 <r . 7} = Q ....(8)

Hence we have

Q =

Ducks in spring Scaup in spring
2240 940

Ducks in fall
665

1957 Preston: Waterfowl of western Pennsylvania Skyways 111

From this it follows that 42 per cent, of all ducks seen at Oneida in
the spring will be scaup. Also, the number of ducks using Oneida in
the fall is about 30 per cent, of the number using it in spring. Or, seeing
that the birds going south in the fall are necessarily more numerous than
those going north in the spring, the efficiency of Oneida Dam in the
fall is considerably less than 30 per cent, of its efficiency in the spring.

These figures may be compared with the actual observations over the
four-year period. In the spring, 3335 duck-days were counted, of which
1323, or 40 per cent, were scaup-days. In the fall we had 622 duck-days,
but this must be corrected for the fact that the number of visits was not
identical in spring and fall (120 and 80 respectively). Adjusting this to
equal numbers of visits would give a figure of 933 or 27 per cent., which
is close to the calculated 30 per cent. The 40 per cent, figure above
refers to identical visits, and is close to the theoretical value of 42 per
cent. Thus both figures agree well with theoretical estimates.

CONCLUSION

The data we have so far accumulated are sufficient, with a few species,
to indicate with fair accuracy the peaks of the migrations, but they are
not sufficient to define with any real certainty the normal beginning and
end of a migration. However, it may be interesting, though unimportant,
to list the earliest and latest dates so far observed, for those species that
have been observed often enough. There are two reasons for this. One
is that a few of the dates already lie outside the range given by Todd.
The other is that it provides targets to be shot at in their turn.

In what follows I have included a few observations made at the Carrie
Dam on the grounds of the Preston Laboratories, and a very few from
other local ponds and dams of Butler County. Although I should have
preferred to use a single “net” such as Oneida Dam alone, it seemed to
me that a slightly better picture of the skyways was presented by utilizing
the other data also, in this particular appendix (Table 3).

Since we have decided that both spring and fall migrations appear
to be represented by Gaussian curves, it follows that it is only a question
of making enough observations and examples will be found that fall
outside the limits established by a few observations. In other words,
all “records” set in the present report must be expected to fall, and to
fall, for the most part, rather promptly.*

•Since this report was completed, part of the spring migration of 1953 has been
observed, and six species were observed earlier than in any previous year, and
one species later than in any previous year, thus bearing out the prophecy.

TABLE 3. EARLIEST AND LATEST DATES, ON SPRING AND FALL MIGRATION, OF BIRDS THAT HAVE BEEN
SEEN ON AT LEAST FOUR MIGRATIONS OUT OF EIGHT. (ONEIDA AND OTHER LOCAL DAMS)

112

Annals of Carnegie Museum

vol. 35

OO

CO

CM

T3

05

co

CO

o>

C3

CO

CO

o

CO

PM

1

>

>

CM

QJ

rfM

CM

CM

CJ

o

£

Dm

<U

o

Z

Dm

QJ

mJ

CJ

f-s

QJ

>

W

u

Dm

>

o

Dm

CJ

CM

UO

o

co

CO

o
. — '

O

O

QJ

CO

Z

CO

05

05

05

o

o

CM

CO

c~~

o

>-

.CM

CM

CM

CM

05

CM

CO

*4

o3

05

Ms

c3

03

Pm

<

Sh

03

M<‘

03

Jh

o3

5-J

c3

o
5^ CM

* s

s s a

CM

Dm

QJ

CO

O

CM

L

03

CM

^ co
CM

00

CO

s z ri 121 i?
z o

50

P

<

eo
. >
> O
o *7-

Oh

QJ

CO

CO

CO

CM

IT)

CO

o

03

1—1

CM

CM

CM

>>

Oj

Ms’ 5*
Dm^

•D

QJ

^ Dm

Ms’

Dm

J-1

03

.D

QJ

QJ

s

C

PM

<

<

pM

pM

xh

^ > 00 >
• 0.0
tJ S5

QJ O

CO w

CM

o

05

m CM ^

CO

CO

o ^ ®

° z

!>■

rf

CM

™ >

oz

t - OO

M 00 04 OD « ,
* CM ^ CM

^ It

< H

50

3

<

05

05

>—• iT5

CM CM

CM

—i >-n

•— M

03 03

U

<n

S-s

03

^ ^
QJ QJ ^

Jh*

03

*i *5! Mi

o3 oS

m

©

CM

o

CM

CO

CM

GO

i>

O

o

o

^ >

Pm

CM

mJ

CJ

>

o

2:

mJ u

DmO

QJ

u

M.O

M-M

Dm

Dm

w

u

o

^ o

^ z

Dm

05

o

co

QJ

QJ

QJ

QJ

TjH

CO

CO

CO

CO

05

CO

m

CM

CM

05

OO

o

— <

CM

CM

*”• f-J

CM

CM

CM

co

Ms

Dm

<

Ms

Dm

<

S3

2

Apr.

aF

JD

.<u

P>H

x>

QJ

-O

Pm

50

3

<

_ OO

so

qj

• »“M

u

QJ

Dm

CO

3

O

O

H-!

3

O

e

a

o

U

qj

•D

qj

Mm

o

w

(0

-Q

QJ

(m

o

T5

qj

»— <
• »-*

-a

qj

S

Green-winged Teal Nov. 15 Mar. 21 Oct. 28 Sept. 21?

Nov. 20 May 6 Nov. 4

1957

Preston: Waterfowl of western Pennsylvania Skyways

113

©

©

©

Pn *-<

CM

CM

. cm

CL

CU

tS

tj

CM

©

cu

co

o

o

o

OO

©

CM

cl

<

u

o

u

c

©

OO

©

©

©

04 ©

r-1

co

CM

f-H

©

CM

©

>-©

CM

CO

CM

r^

$4 ©

5-1

CM

»4

. CL •

rS

pS

O, .

CL

CL

.d

•g < £

S xi

v4

a

&H*

n .

s4

PS

s

< -f

<

s4

cS

<

<3L>

Pl

QJ L-U

f* £

qj

Pl

2

Wh

<

2

s

cu

ptl

2

$

> O ^
O

Z « °r

n. QJ
QJ ' to
0/3

PL —*
bD

* O

; Z

CO
CO
. >
> o

° z

3

<

<U

z

CM

O

o

z

O

z

oo

O

©

0

©

z

6 w
CO ©

CM

©

00

©

©

CO

CM

CM

CM

1—1 ’~l

CM

>^CM

f-H

CM

’“H

v4

S_<

CM

v4 t4

£

Ma

Feb.

X3

QJ

fr4

Ma

Feb.

L

CL

CL

<

CL

<

>4

Pj

CL . CL
< a.< o.

rt

2

2

Uh

2

<

CJ

cu

2

< <

00 > M-

^ O

X Z X

« CL

° Jj

© <M

r-H i

PS

u 3 >->

a/

Ui

f-H

OO

F-H

oo

oo

f-H

CM

f-H

CM

CM

ilO

©

>

>

>

>

CM

>

>

o

>

o

>

o

>

O

o

0

Z

z

o

z

Z

o

Z

z

0

&

Z

Oct

Z

t"

oo

c-~

00

©

©

CM

CM

©

CO

1

©

©

^ CM

’~i

re

10 ©

©

CM

<

f-H

CM

S-.CM

f-H

*4

.CM

CM

^ CM

CU

«(

C3

a

CL •

C^!

nS

PS

>4

ctf

S M

»4

ps

<

>4

ps

s

Jh

s4

S 1

Cl,

J§

;4

PS

< S3

(4

pS

y »4

F=i fX,

^ CL

2

2

2

s

<

<

s

S

<

<

©

<m >:

s*

^ >

OO

CM

©

0O ^

s*

$ d8

> Q >

O o

Z Z

2c £ 00

&

CO

27

©

CM

00

CM

©

CM

25

27

oo t4

ct

©

<4

fS

>4

PS

xi

s4

cS

Apr.

4>

Pl

s

<u

Pl

CL

<

a,

C

c

r*

bo

&

s

w

-o

w

bo

•3

SJ

H

0

3

PS

M

CJ

3

(m

QJ

2

3

QJ

«

'3

fi

-3

cs

u

OJ

5

QJ

XJ

cS

3

Q

3

cu

cu

CL

oo

'%

<L>

3

cu

XS

<u

e

bo

c

CL

3

5

FH

S

<u

s

*3

?r

3

3^

3

3

QJ

3

O

0

t-i

XI

i

s

cZ

3

c/3

<

P5

0

X

&

bo

qj

%

o

><

Sd

JJ

£

-2

*53

>

Bonaparte's Gull Mar. 24 Mar. 29 Apr. 1 Apr. 13

Apr. 23 Apr. 20

114

Annals of Carnegie Museum

vol. 35

These results may perhaps be usefully condensed to give the earliest and

latest dates so far observed for migrant species in spring and fall, as follows:

Common Loon March 19— May 18

October 19— November 25

Horned Grebe March 28— April 23

November 13— November 23

Pied-billed Grebe February 16— May 12

September 1— November 24

Great Blue Heron March 22— April 27

August 12— October 22

Whistling Swan March 29— April 25

October 29— November 17

Canada Goose March 19— March 27

October 13

Mallard February 12— May 1

September 10— November 11

Black Duck February 12— May 14

August 30— November 20

Gadwall February 20— May 11

September 3

Baldpate February 21— May 1

September 14— November 18

Pintail March 15— March 20

August 29— November 18

Green-winged Teal March 21— May 6

September 21— November 20

Blue-winged Teal March 25— May 12

August 12— November 19

Redhead February 9— March 1

November 18

Ring-necked Duck February 9— April 27

October 18— November 11

Scaup March 1— May 18

October 20 — November 23

American Golden-eye February 9— March 12

October 20— November 23

Buffle-head March 21— May 1

October 28— November 27

Old-squaw April 6— cApril 25

(no Fall observations)

Ruddy Duck March 21— May 13

October 17— November 23

Hooded Merganser March 13— April 20

October 4— December 4

Red-breasted Merganser February 9— May 1

October 28— November 20

1957 Preston: Waterfowl of western Pennsylvania Skyways

115

Coot

PSeptember 10— November 24

Greater Yellow-legs

April 9— May 14

PSeptember 1— November 20

Lesser Yellow-legs

April 20

Bonaparte’s Gull

September 10— PSeptember 28

March 24— April 23

(no Fall observations)

116 Annals of Carnegie Museum vol. 35

REFERENCES

Bagg, A. M.

1950. Barometric pressure patterns and spring bird migration. Wil-
son Bulletin, v. 62, p. 5-19.

Bennett, Holly Reed

1952. Fall migration of birds at Chicago. Wilson Bulletin, v. 64,
p. 197-220.

Bent, A. C.

1946. Life histories of North American diving birds. 237 p. Dodd
Mead & Co., New York.

Davis, D. E.

1952. Social behavior and reproduction. Auk, v. 69, p. 171-182.

Gunn, W. W. H., and A. M. Crocker

1951. Analysis of unusual bird migration in North America during
the storm of April 4-7, 1947. Auk, v. 68, p. 139-163.

Hald, A.

1952. Statistical tables and formulas. 92 p. Wiley, New York.

Hochbaum, H. A.

1944. The canvasback on a prairie marsh. American Wildlife Insti-
tute, Washington, p. 1-201.

Kortright, Francis H.

1943. The ducks, geese and swans of North America. 476 p. Ameri-
can Wildlife Institute, Washington.

Mayfield, H. F.

1947. Spring migration, Ohio-Michigan region. Audubon Field Notes,
v. 1, p. 153-154.

Preston, F. W.

1951. Flight speed of common loon. Wilson Bulletin, v. 63, p. 198.

Preston, F. W.

1956. The migrant loons of western Pennsylvania. Auk, v. 73, p.
235-251.

Sutton, George Miksch

1928. The birds of Pymatuning Swamp and Conneaut Lake, Craw-
ford County, Pennsylvania. Annals of the Carnegie Museum,
v. 18, p. 19-239.

Thomas, Edward S.

1953. Personal communications.

Todd, W. E. Clyde

1904. The birds of Erie and Presque Isle, Erie County, Pennsyl-
, of the Carnegie Museum, v. 2, p. 481 -596-.

Todd, W. E. Clyde ’ ’

1940. Birds of western Pennsylvania. 710 p. University of Pittsburgh
Press, Pittsburgh.

Trautman, Milton R.

1940. The birds of Buckeye Lake, Ohio. 466 p. University of Michi-
gan Press, Ann Arbor.

£ O ?4 * ^

P h

MAY 2 0 1959

ART. 8. SYSTEMATIC NOTES ON NORTH AMERICAN BIRDS

2. The Waterfowl (Anatidae)

By Kenneth C. Parkes
Associate Curator of Birds, Carnegie Museum

This is the second of a series of papers on the systematics and nomencla-
ture of certain North American birds. A general introduction and acknowl-
edgments for the series as a whole will be found in the first paper (Parkes,
1955c).

The waterfowl have long been among the most popular birds, and there
has been a recent renewal of interest in their classification. The now classic
paper by Delacour and Mayr (1945), in which a radically new alinement
of the swans, geese and ducks was proposed, may be regarded as a major
turning-point in waterfowl taxonomy. Most of the changes from the tradi-
tional classification employed by Peters (1931), even though sometimes
quite drastic, have been accepted by most students of this family. In the
decade since the revised classification of the waterfowl appeared, many
papers have been published which supplement or correct some of the state-
ments made by Delacour and Mayr, or take issue with them on specific points.
Differences of opinion are to be expected among taxonomists, and Delacour
and Mayr would be the last to claim that theirs was the final word. In
general, however, it may be said that our knowledge of the systematics of
the waterfowl, when compared with other bird families of similar size, is
exceptionally thorough.

Special mention must be made of the work of Verheyen, who has taken
sharp issue with Delacour and Mayr, and proposed his own classification
of the waterfowl (summarized in Verheyen, 1955b) based primarily on com-
parative osteology. A thorough critique of Verheyen’s work may be accom-
plished only by a comparative anatomist. It is quite apparent, however,
that Verheyen has fallen victim to the tendency described by Simpson (1945,
p. 23) as “the tendency to raise the ranks of groups without need, that is,
without gaining any practical advantage. One of the more evident symptoms
of this tendency is the appearance of many monotypic groups in classifica-
tion. If a classifier makes mostly monotypic families, genera, etc., it is a
fair statement that he is giving family rank to what should be called genera,
generic rank to species, etc.” That this is descriptive of Verheyen’s classifica-
tion is made plain when it is pointed out that he has divided the relatively
homogeneous family Anatidae into sixteen families, no less than six of
which are erected to include a single species!

Brief mention should also be made of the work of Yamashina (1952),
who proposed a classification of the Anatidae based entirely on cytological
and hybrid sterility data. Although he examined only fifty species, his classi-
fication agrees in many respects with that of Delacour and Mayr. Yamashina,
however, has “lumped” genera to an even greater extent than did Delacour
and Mayr, and his classification represents the opposite extreme in taxonomic
practice from that of Verheyen.

Submitted for publication, November 15, 1957
Issued November 7, 1958

117

INSTITUTION A™* * 15,3

118

Annals of Carnegie Museum

vol. 35

Since the appearance of the first paper in the present series, the long-
awaited fifth edition of the American Ornithologists’ Union “Check-list of
North American birds” has been published (1957). The classification of
waterfowl employed by the A.O.U. is a conservative one, based chiefly on
that of Peters (1931). Reference will be made beyond to some of the ques-
tions on which I feel that the A.O.U. Check-list is conservative to the point
of being reactionary.

1. The Swans

The classification of the swans hinges on the relative importance assigned
to certain osteological features. Although Delacour and Mayr (1945) omitted
all reference to the striking internal differences among these superficially
similar birds, these structural features are described in Delacour’s recent
book (1954, p. 57, 71). Wetmore (1951) believed that two genera should be
recognized; Cygnus for those species in which the trachea passes directly into
the thorax without entering the sternum, and Olor for those in which the
trachea loops into the sternum, the furculum being modified at the sym-
physis to accommodate this loop. In addition to these major anatomical
features, the two genera may be separated by certain relatively minor ex-
ternal characters, such as the cuneate tail of Cygnus versus the rounded tail
of Olor. The two groups also differ in behavior patterns, and it seems worth
while to follow Wetmore in recognizing two genera.

There are also two schools of thought regarding specific limits within the
genus Olor. Delacour and Mayr (1945, p. 8) followed the suggestion of
Hartert (1920, p. 1275) that the New World O. columbianus and O. buccina-
tor be considered conspecific with the Old World O. bewickii and O. cygnus ,
respectively. In the case of the Whistling Swan (O. columbianus ), such action
is justifiable. A glance at the range map published by Delacour (1945,
p. 84-85) shows that the New World columbianus and the two Old World
forms bewickii and jankowskii are obvious geographic representatives, dif-
fering slightly in size and conspicuously in the relative amount of yellow
at the base of the bill. All three are highly migratory Arctic nesters. I there-
fore agree that these two Old World forms should be considered subspecies
of Olor columbianus.

The second case, that of O. buccinator and O. cygnus, is quite different.
Here Wetmore (1951) and Delacour (1954, p. 71) differ on a question of
fact rather than merely on interpretation of accepted facts. Wetmore recog-
nized the subgenus Clangocycnus Oberholser for the Trumpeter Swan,
O. buccinator, basing this on the fact that in this species the trachea makes
“a dorsal loop as it enters sternum, protected by a bony case that projects
into the anterior end of the body cavity.” Delacour (1954, p. 71) claims
that this is also true of the Whooper Swan, O. cygnus, which he therefore
combined with buccinator as a single species. Wetmore, on the other hand,
placed cygnus in the typical subgenus Olor. I have examined sterna of the
swans in question, and find that Wetmore is correct; the sternum of O.
cygnus does not have the enlarged “bony case” typical of O. buccinator, but
is somewhat intermediate toward that of O. columbianus. This is illustrated
by Schiller (1925, pi. 65).

1958

Park.es: Systematic Notes on North American Birds

119

As mentioned above,- Olor columbianus and O. bewickii are clearly geo-
graphic representatives of one another. This can not be said for O. buc-
cinator and O. cygnus * The Whooper Swan is primarily an Arctic nester,
although it does extend into central Asia. It is strongly migratory. The
Trumpeter Swan, on the other hand, is decidedly a bird of temperate re-
gions, and is relatively sedentary.

The Whooper Swan resembles Bewick’s Swan in having a large area of
yellow at the base of the bill, a feature of which there is no hint in the
Trumpeter Swan. The total of the evidence seems to suggest that Olor
cygnus is a larger derivative of the same stock which gave rise to the O.
columbianus group, and is no more than generically related to O. buccinator.
1 believe, therefore, that the two North American swans should be known
as Olor columbianus columbianus (Ord) and Olor buccinator (Richardson),
respectively. This differs from the current treatment in the A.O.U. Check-
list only in considering the Whistling Swan to be the New World representa-
tive of a polytypic species.

2. The genus Branta

Verheyen (1955a, p. 9) has introduced into the literature the name
Eubranta , apparently as a new genus intended to include the Barnacle
Goose, Branta leucopsis, and the Red-breasted Goose, B. ruficollis. He gives
no diagnosis of this new genus, apparently basing it primarily on the fact
that these two species have, on the average, two less vertebrae (one cervical,
one sacral) than do B. bernicla and B. canadensis. Verheyen also neglected
to designate a type species for “Eubranta” . Even if the segregation of these
two species as a separate genus from Branta were warranted, which I do not
believe, the introduction of a new generic name was completely unnecessary.
Both of thse species, according to the synonymies presented by Hellmayr and
Conover (1948, p. 294-295), have been named as monotypes of new genera;
Leucopareia Reichenbach, 1852, for leucopsis, and both Rufibrenta Bona-
parte, 1856, and Ptocas Heine, 1890, for ruficollis. In order to allocate the
name Eubranta Verheyen, 1955, I here designate its type as Anas leucopsis
Bechstein, and thus place it as an absolute synonym of Leucopareia Reichen-
bach, 1852, a genus currently believed inseparable from Branta Scopoli, 1769.

A brief distributional note on the Canada Goose may well be placed here.
Hanson and Smith (1950, p. 76-77) believed southern New Jersey to be the
northernmost part of the Atlantic coast reached by migrating and wintering
Branta canadensis interior Todd. The A.O.U. Check-list (1957, p. 61) goes
even further, stating “Not recorded from the Atlantic coast north of Mary-
land”. However, I have examined three specimens of this race from Montauk
Point, Long Island, N. Y., in the collection of the American Museum of
Natural History (A.M.N.H. 350131, December 3, 1909; A.M.N.H. 350133 and
350134, March 14, 1902). Hellmayr and Conover (1948, p. 305) listed a
specimen in the Chicago Natural History Museum from Rockaway Beach,
Queens County, New York.

3. The genus Chen

The status of the Blue and Snow geese, long one of the most difficult
questions in systematic ornithology, has been under investigation for some

120

Annals of Carnegie Museum

vol. 35

years by Graham Cooch, and I do not venture to discuss the matter here.
I can not, however, refrain from stating that I can see no justification for
the continued recognition of the Blue Goose, Chen caerulescens, as a sepa-
rate species in the A.O.U. Check-list (1957). Whether the relationship of the
Blue and Lesser Snow geese is best expressed by calling them subspecies or
color phases will be determined by specialists in this group, but to call these
two forms separate species is to ignore half a century of progress in the
study of evolution.

Many modern authors include Chen in an expanded genus Anser. This
is entirely a matter of subjective preference as to size and scope of genera,
since the two are undeniably closely related.

4. Anser albifrons

Delacour (1954, p. 110, and personal communication) has stated that all
records known to him of presently existing specimens of White-fronted
Geese along the east coast of North America pertain to the Greenland
race, Anser albifrons flavirostris Dalgety and Scott. One of the most striking
characteristics of this race is the possession of an orange-yellow rather than a
pink bill. Stoner (1944) published the details of a specimen killed near
Rouses Point, Lake Champlain, N. Y., on October 22, 1943. Although the
specimen was not preserved, both measurements and color notes were taken.
The description of the bill as “pink” precludes the identification of this
specimen as flavirostris . The measurements indicate that it was neither the
small European A . a. albifrons nor the very large far western A. a. gambelli.
This leaves, by elimination, A. a. frontalis , the common White-fronted
Goose of western North America, which migrates chiefly west of the Mis-
sississippi. An occasional eastern stray of this form would hardly be sur-
prising.

I have examined the series of European specimens of Anser albifrons men-
tioned by Todd (1950, p. 64). Comparing four from Holland with eight
from Austria, Montenegro and Albania, Todd wrote that the former “differ
in the darker, browner coloration of their upper parts and wings and in
the more brownish suffusion of the neck and under parts generally. The
significance of this variation I do not presume to explain beyond suggest-
ing that, since it cannot be seasonal, it could be racial.” Although not directly
pertinent to a North American bird, this matter may appropriately be
settled in the present discussion of the species involved. The color differ-
ences between the two series noted by Mr. Todd may be easily explained,
and have nothing to do with geographic variation. The Holland birds were
collected in 1892 and 1900, and are foxed and stained. The others were
collected in 1929 and 1932, are clean specimens, and have never been on
exhibition as the Holland birds were.

5. The Mallard and its relatives

In an earlier paper (Parkes, 1954, p. 152) I commented on a statement
made by Delacour and Mayr (1945, p. 21) who wrote that “it seems obvious
that the Mexican and Black Ducks ( diazi and rubripes) are only sub-
specifically distinct from the Dusky Duck ( fulvigula ).” Delacour and Mayr
united these three forms under fulvigula, the oldest name. My comments
were as follows: “It is my belief that the case is by no means so ‘obvious'.

1958

Parkes: Systematic Notes on North American Birds

121

The Mexican Duck, Anas diazi, is so close to the Mallard, A. platyrhynchos,
that a case might be made for considering it a rather restricted ‘hen-feathered'
subspecies of Mallard, except that diazi and platyrhynchos are sympat-
ric (Lindsey, 1946, p. 484). A comparatively recent origin of diazi from
platyrhynchos is suggested by the high frequency of hybridization (Lindsey,
1946, p. 484).” I then went on to point out that the Dusky Duck, A fulvi-
gula, is in many respects about midway between the Mallard and the Black
Duck.

Delacour himself apparently came to doubt the “obvious” conspecificity of
rubripes, diazi and fulvigula. In his recent book (Delacour, 1956) he unites
the latter two forms with platyrhynchos , allowing rubripes to stand as a
full species. He gives no references to support this treatment, nor does he
explain his own change of mind. The range map (on page 41 of his book)
does not show the overlap of the breeding ranges of diazi and platyrhynchos ,
He states that “They [diazi] do not seem to mix with wintering Mallards
which are often found at the same localities during the winter”, a state-
ment completely at variance with the New Mexico observations of Lindsey
(1946). He mentions the fact that drakes of diazi “sometimes have more or
less curled up central tail feathers and traces of bright colours . . . thus
showing a close relationship to the common Mallard.” He gives no indica-
tion as to whether such birds are ever found outside the area of overlap
of diazi and platyrhynchos , nor does he even mention the hybridization de-
scribed by Lindsey and mentioned in the A.O.U. Check-list (1957, p. 72).
The treatment of the Dusky Duck ( fulvigula ), now called Florida Duck by
Delacour, is equally scanty and devoid of explanation. At present I can see
no compelling reason to alter my statement of 1954 that “All in all, I prefer
to consider the Mallard, Black Duck, Dusky Duck and Mexican Duck as
specific entities.”

6. The Green-winged Teal

The conservatism of the A.O.U. Check-list is nowhere illustrated better
than by its persisting in giving full specific rank to the American Green-
winged Teal (Anas carolinensis). This provincial viewpoint has been aban-
doned by virtually all students of waterfowl the world around. As is well
known, the females of the American and the European (A. crecca ) Green-
winged Teal are virtually indistinguishable. The two are geographic repre-
sentatives, but individual birds of each of the races occasionally stray within
the range of the other. Hybridization among ducks is so common, of course,
that it can not be used as a sole criterion of conspecificity, but it is inter-
esting to note that Cruickshank (1986) and Poole (1940) have described
apparent hybrids or intergrades between carolinensis and crecca. (See Parkes,
1955b, p. 38, for further discussion of this case.) There are no behavioral
characters, often useful in duck classification, to separate the two forms. The
American Green-winged Teal should be known as Anas crecca carolinensis
Gmelin.

7. The Shovellers

Delacour and Mayr (1945, p. 17) and Delacour (1956, p. 19) have re-
emphasized the extremely close relationship among the four shovellers
(“Spatula”) and the three blue-winged “teal” (“Querquedula”). They make

122

Annals of Carnegie Museum

vol. 35

the interesting and, to me, highly plausible suggestion that “the shoveller
group is polyphyletic, owing its origin to the repeated development of
large-sized and large-billed species from the original blue-winged duck stock”
(Delacour and Mayr, 1945, p. 17). Oliver (1954, p. 193) objected to this
idea, with its necessary corollary of combining “Spatula” with Anas (in-
cluding “ Querquedula” ). Oliver’s full statement is as follows: “Spatula differs
from Anas in important bill characters and consists of closely related species
occupying different continents. This shows [italics mine] that the species
have not risen independently in each continent from different species of
Anas . Spatula is so different from Anas that its union with that genus
would cover up an important morphological characteristic and make Anas
indefinable. It certainly should be kept as a genus distinct from Anas.” This
treatment is adopted by the conservative A.O.U. Check-list (1957), which
interposes the widgeons Mareca ”) between the Shoveller and its closest
relatives, the Blue-winged Teal and Cinnamon Teal, included in Anas by
the A.O.U. Check-list. Meinertzhagen (1951, p. 444) pursues the concept of
monophyletic origin of the shovellers to the ultimate extreme, and makes
all of them subspecies of Spatula clypeata!

As amply shown by Delacour and Mayr, the shovellers and the blue-
winged “teal”, taken together, constitute a well-knit group within the ex-
panded genus Anas. Extreme variation in bill size notwithstanding, these
ducks are obviously more closely related to one another than any is to the
rest of the genus Anas. This is supported by plumage pattern, feeding habits,
courtship display, and, to some extent, voice. Two wild-taken hybrids be-
tween Anas (“ Spatula ”) clypeata and A. (“ Querquedula ”) discors have been
described by Childs (1952). I saw what appeared to be such a bird myself
at the Montezuma Federal Waterfowl Refuge, in central New York, on
October 11, 1952. Delacour (1956, p. 182-183) makes the interesting point
that such hybrids are extremely similar in appearance to the Australian
Shoveller (A. rhynchotis).

In order, then, to justify retention of the genus Spatula for the four
shovellers, Oliver (and presumably the A.O.U. Check-list) must rely entirely
on the shape of the bill, a notoriously unreliable character in avian taxonomy
at generic and higher levels. The geographic distribution of the shovellers
(one holarctic, three in the Southern Hemisphere) and the remarkable simi-
larity in color and plumage between the Cinnamon Teal ( cyanoptera ) and
the South American Shoveller (platalea), and between the Blue-winged Teal
(discors) and the Australian Shoveller (rhynchotis), are strong evidence in
favor of independent origin of shovellers from blue-winged duck stock in
at least three different cases. Such a polyphyletic origin would, of course,
preclude segregation of the large-billed forms as a genus Spatula , much less
combining them all as subspecies of Spatula clypeata. The Cape Shoveller
(Anas smithi = Spatula capensis of authors) is a somewhat different case.
This African species is virtually a “hen-feathered” version of the holarctic
A. clypeata, and may well be derived from the latter species, which migrates
to Africa. These two could be considered to constitute a superspecies; the
“shoveller” bill would then have been independently derived only three
times.

1958

Parkes: Systematic Notes on North American Birds

12S

8. The Greater Scaup

Witherby et al. (1939, p. 308), Scott (1949, caption to plate XV), and
some other authors have cast doubt on the validity of the New World sub-
species of the Greater Scaup, Aythya marila nearctica Stejneger. With this
in mind I examined the extensive series of this species in the American
Museum of Natural History, and found that nearctica is readily recognizable
by the coarse black barring of the upper parts, exactly as characterized by
Hellmayr and Conover (1948, p. 371, footnote). The geographically inter-
mediate A. m. mariloides (Vigors) of eastern Asia is somewhat intermediate
in color between marila and nearctica, but is smaller than either. (For
measurements, see Hartert, 1920, p. 1344.)

9. The Spectacled Eider

The fifth edition of the A.O.U. Check-list (1957, p. 91) places the Spec-
tacled Eider in the genus “Lampronetta” . In the thirty-first supplement to
the fourth edition of the Check-list (Wetmore et al., 1956, p. 448), the pro-
posed change from Arctonetta Gray, 1856, as used in the fourth edition, to
Lampronetta Brandt, 1847, was announced. The reference cited for this
change was a paper of mine (Parkes, 1955a). It is true that in this paper
I pointed out that Lampronetta antedated Arctonetta. However, my wording
was intended to make it plain that I was certainly not advocating the con-
tinued recognition of a monotypic genus for the Spectacled Eider under
any name; I pointed out that this species possesses no trenchant characters
to separate it from Somateria, and mentioned the priority of Lampronetta
over Arctonetta to indicate that a name change for this species was inevitable
in any case. Lest the citation of my paper in connection with the change in
A.O.U. usage be misleading, I wish to reiterate my firm belief that the
Spectacled Eider can not be separated from the genus Somateria .

REFERENCES

American Ornithologists’ Union

1957. Check-list of North American birds, ed. 5. 691 p.

Childs, Henry E., Jr.

1952. Hybrid between a Shoveller and a Blue-winged Teal. Condor,
v. 54, p. 67-68.

Cruickshank, Allan D.

1936. Some observations of the European Teal. Auk, v. 53, p. 321*322.
Delacour, Jean

1954. The waterfowl of the world, v. 1, 284 p. London.

1956. The same. v. 2, 232 p. London.

Delacour, Jean, and Ernst Mayr

1945. The family Anatidae. Wilson Bulletin, v. 57, p. 1-55.

Hanson, Harold C., and Robert H. Smith

1950. Canada geese of the Mississippi Fly way. Bulletin of the Illinois
Natural History Survey, v. 25, p. 67-210.

Hartert, Ernst

1920. Die Vogel der palaarktischen Fauna, v. 2, pt. 4, p. 1217-1344.

124

Annals of Carnegie Museum

vol. 35

Hellmayr, Charles E., and Boardman Conover

1948. Catalogue of birds of the Americas and the adjacent islands.
Field Museum of Natural History Zoological Series, v. 13, pt. 1,
no. 2. 434 p.

Lindsey, Alton A.

1946. The nesting of the New Mexican Duck. Auk, v. 63, p. 483-492.
Meinertzhagen, R.

1951. Some relationships between African, Oriental and Palaearctic
genera and species, with a review of the genus Monticola.
Ibis, v. 93, p. 443-459.

Oliver, W. R. B.

1954. Avian taxonomy. Emu, v. 54, p. 190-198.

Parkes, Kenneth C.

1954. Notes on some birds of the Adirondack and Catskill moun-
tains, New York. Annals of the Carnegie Museum, v. 33, p.
149-178.

1955a. The generic name of the Spectacled Eider. Auk. v. 72, p. 85-86.
1955b. Sympatry, allopatry, and the subspecies in birds. Systematic
Zoology, v. 4, p. 35-40.

1955c. Systematic notes on North American birds. 1. The herons and
ibises (Ciconiiformes). Annals of the Carnegie Museum, v. 33,
p. 287-293.

Peters, James L.

1931. Check-list of birds of the world, v. 1, 345 p. Cambridge, Mass.

Poole, Earl L.

1940. Recent records from Lake Ontelaunee, Pennsylvania. Auk.

v. 57, p. 577-578.

Schiller, E. L.

1925. Danmarks Fugle, v. 1, 552 p. Copenhagen.

Scott, Peter

1949. Key to the wildfowl of the world. Severn Wildfowl Trust
Second Annual Report, p. 91-137.

Simpson, George G.

1945. The principles of classification and a classification of mammals.
Bulletin of the American Museum of Natural History, v. 85,
350 p.

Stoner, Dayton

1944. White-fronted Goose at Rouses Point, New York. Auk, v. 61,
p. 651-652.

Todd, W. E. Clyde

1950. Nomenclature of the White-fronted Goose. Condor, v. 52, p.

63-68.

Verheyen, Rene

1955a. La syst&natique des Anseriformes bas^e sur l'ost^ologie com-
pare (suite). Bulletin Institut Royal des Sciences Naturelles
de Belgique, v. 31, no. 36, p. 1-16.

1955b. The same (fin). Bulletin Institut Royal des Sciences Naturelles
de Belgique, v. 31, no. 38, p. 1-16.

1958

Parkes: Systematic Notes on North American Birds

125

Wetmore, Alexander

1951. Observations on the genera of the swans. Journal of the Wash-
ington Academy of Sciences, v. 41, p. 338-340.

Wetmore, Alexander (Chairman) and others

1956. Thirty-first supplement to the American Ornithologists’ Union
Check-list of North American birds. Auk, v. 73, p. 447-449.

Witherby, H. F., F. C. R. Jourdain, N. F. Ticehurst and E. W. Tucker
1939. The handbook of British birds, v. 3, 387 p. London.

Yamashina, Yoshimaro

1952. Classification of the Anatidae based on the cyto-genetics. Papers
from the Coordinating Committee for Research in Genetics,
v. 3, p. 1-34.

y -

PUf^

ART. 9. A RECENT FISSURE
BEDFORD COUNTY, PENNS
By John E. Guilday* and Martin

This paper is a report on a collection of vertebrate skeletal material re-
covered from the New Paris sink-holes. These fissures are located 1.5 miles
northeast of New Paris, West St. Clair Township, Bedford County, Pa., on
the property of Oscar Miller. Fig. 1.

The bone deposits were discovered by J. Howard Taylor of New Paris,
Pa., who broke into the top of fissure (sink-hole) no. 2, then evidenced by
a slight depression on the surface of the ground. Mr. Taylor made a small
collection of elk bones which included the most interesting single specimen
from the fissures, two elk vertebrae with an imbedded flint arrowhead. The
locality was visited in February, 1932, by Charles E. Mohr, and reported on
by him in Stone’s ‘ Pennsylvania Caves” (ed. 2, 1932) and ‘‘Descriptions of
Pennsylvania’s Undeveloped Caves” (1953). In April, 1948 a group com-
posed of members of the Pittsburgh Explorers Club and the Pittsburgh
Grotto, National Speleological Society, visited the area. Mr. Taylor led the
group and volunteered valuable information about the site. As a result of
this trip the major part of an elk skeleton was excavated and donated to
Carnegie Museum. With the aid of a grant from the A. W. Mellon Educa-
tional and Charitable Trust, the Museum was able to institute field work
at the site. Excavations were carried out during 1949 and 1950 under the
direction of Dr. J. Leroy Kay. Actual field work was under the direction of
Albert C. Lloyd. Oscar Miller not only granted the Museum permission to
work at the site but also served as an active and enthusiastic member of the
excavating party. The project is indebted to the Pittsburgh Grotto, National
Speleological Society, whose members have put so much time and effort on
it. Special thanks are due Albert H. Bauer, Ralph C. Bossart and John A.
Leppla of Pittsburgh for their assistance in the excavation.

The writers would like to express their gratitude to Dr. J. Leroy Kay and
Dr. J. Kenneth Doutt for the privilege of studying the collection and for
much helpful advice and assistance; to Neil D. Richmond, Curator of Am-
phibians and Reptiles, Carnegie Museum, for examining the snake remains;
to George B. Thorp and Miss Lillian Heeren for the surveying and drafting
of the map, respectively; and to Miss Caroline A. Heppenstall, Assistant
Curator of Mammals, and Mrs. Alice M. Guilday for their assistance in the
project.

A popular article concerning the fissures appeared in Carnegie Magazine
(Guilday, 1948) and in 1955 a report upon the carnivores of the New Paris
sink-holes by the junior author was submitted to the University of Pitts-
burgh in partial fulfillment of the requirements of a Master of Science
degree, and is on file at the University of Pittsburgh library (Bender, [1955]).

The New Paris sink-holes treated in this report are a group of vertical
limestone caverns. They occur in the Helderburg (lower Devonian) Lime-
stone. It is not obvious whether the sink-holes are solution enlargements of

* Assistant Curator of Comparative Anatomy, Carnegie Museum.

•[Instructor in Geology, University of Pittsburgh.

Submitted for publication, November 5, 1957.

Issued November 20, 1958.

127

3ivn i

institution

1 1958

128

Annals of Carnegie Museum

vol. 35

structural rock joints or whether they were positioned by other factors, but
the former seems more probable.

The Helderburg Limestone is exposed on the Schellsburg anticline (locally
known as Chestnut Ridge) as a series of inliers which occupy the valley areas
on the flanks of the anticline with low ridges between them. These inliers
are surrounded by the outcropping Oriskany formations that form the high
ground of the anticline.

Physiographically the New Paris sink-holes lie within the Ridge and Valley
Province, just east of Allegheny Mountain. This mountain lies at the eastern
boundary of the Allegheny Plateau. They are on the west flank of Chestnut
Ridge, at an altitude of 1500-1600 feet. The immediate area of the fissures
is well drained and covered with hardwoods. The surrounding areas are
under cultivation. Biologically the area lies within a zone of transition be-
tween the Carolinian Biotic Province and the more northerly Canadian
Biotic Province. (Dice, 1943). Gifford and Whitebread (1951) state that
the Ridge and Valley section generally is difficult to assign to either of these
two broad classes. They state (p. 18), “No groups of mammals were found
that could be classed as either northern or southern. For the most part the
mammals that occur together in any one locality here are determined by
the availability of suitable habitats and the adaptability of the species.”

The collection includes samples from three sink-holes, all geologically
related but not in direct contact (See Fig. 1). Sink-hole no. 4 or “Lloyd’s rock
hole” is currently being excavated by the Pittsburgh Grotto, National
Speleological Society, but, to date, the work has produced no fossil fauna.*
Sink-hole no. 1 or “ Kantner’ s Sink”.

Known as the “Round sink-hole”. Excavated by the Museum party, this
sink-hole proved unproductive and received only cursory attention. A small
collection of small mammal bones was made.

Sink-hole no. 2 or the “ Elk hole”.

This was the original bone deposit discovered by Mr. Taylor and it was
completely excavated by the Museum party. This sink-hole was not open to
the surface at the time of discovery but was sealed by a cover of tree roots,
dirt and organic debris. This fissure had an initial depth of 47 feet. Excava-
tions were carried to a further depth of 12 feet at which point the bone de-
posit was flooded by ground-water and operations ceased. The bulk of this
report concerns the collection from this fissure.

Sink-hole no. 3.

A small collection of both large and small mammal bones from this
fissure was donated by members of the Pittsburgh Grotto, National Speleo-
logical Society, and the Pittsburgh Explorers Club. This was augmented by
collections made by Arnold D. Lewis, Allen D. McCrady and the senior
author of this paper, prior to the Museum excavations. This material was
picked up on or close to the surface of the rock debris at the bottom of the
fissure in two areas— one at the bottom of the first landing (approximately
50 feet) and the other at the bottom of the sink-hole (87 feet). The con-
temporaneity of these two deposits was indicated by the finding of frag-

#Since the above was written, an extensive Pleistocene fauna has been recovered from
sink-hole no. 4, and is currently under study.

1958

Guilday and Bender: Fissure Deposit in Pennsylvania

129

Fig. 1. Maps showing location of New Paris sink-holes

Main map gives detailed compass directions for individual sink-holes. Insert map
gives general location of sink-holes described in this article. To reach the point
marked Miller house, proceed north on Route 96, 0.3 mile from intersection in
center of New Paris, take right fork (NE.) for 0.3 mile, turn right (SE.) 0.1 mile,
turn left (NE.), continue straight 0.45 mile to Miller house.

130

Annals of Carnegie Museum

vol. 35

ments of the same black bear skull at both levels. This fissure was not ex-
cavated and more skeletal material remains but any extensive digging would
produce a rock slide.

The bone-bearing deposit at the bottom of the fissures was a heterogeneous
mixture of clay, apparently resulting from chemical breakdown of the lime-
stone walls, angular fragments of limestone varying in size from pebbles to
blocks many tons in weight, organic debris and bones. No evidence of selec-
tive sorting by water was noted and no evidence of internal stratification
was observed in the deposits. Apparently the animals blundered into the holes
over a period of many years; were either killed by the fall or starved to death
and the deposit grew, augmented through the years by falls of rock, soil and
additional victims. Oscar Miller, the owner of the property, states that the
fissures of this area have been alternately opened and closed to the surface
by varying accidents of forest litter accumulation and periodical cave-ins.

The bones were, for the most part, unbroken, well preserved, and in the
case of some of the skeletons from sink-hole no. 2 still lay in attitudes of
articulation. Some of the bones from this fissure were spattered with a thin
calcareous patina, but none of the bones from the fissures showed any
noticeable degree of mineral replacement.

The numbers of bones listed below refer only to items which were readily
diagnostic; such as limb bones, skulls, and teeth, and on these were based
the estimates for the minimum number of individuals represented. No
attempt was made to list the thousands of minor skeletal parts present,
except in special cases. To attempt to do so would have proved both futile
and impractical. The assumption is made that a given animal was trapped
in its entirety and, for census purposes, a count of skull elements would
give a reliable estimate. All bones were carefully examined however.

The bones recovered from the fissure deposits represent species all living
today, although several do not inhabit the area at present. The elk and the
wolf were exterminated in Pennsylvania in the late 1800’s (Rhoads, 1903).
A flint arrowhead, embedded in two coalesced cervical vertebrae of an elk,
places the age of the deposit in Post-Wisconsin times, most probably well
into the Recent and possibly as late as the eighteenth century. Absence of re-
mains of the red fox and the opossum (from sink-hole no. 2), two animals
now present in the area but believed to have been absent during late pre-
historic times, suggests the possibility that the deposit predates the period
of profound ecological changes ushered in by European colonization. Some
of the smaller mammals have been trapped since the most recent opening
of the fissure. This is very apparent in the fresh condition of some of the
rabbit remains. A woodchuck fell in during the excavating operations. The
bulk of the collection appears to be of some historical antiquity, although
geologically recent.

There are 67 forms of mammals representing 47 genera recorded from
Bedford County within recent times. (Gifford and Whitebreacl, 1951). By
comparison, 26 species of 25 genera were found in the fissure.

While the animals recovered from the fissures in no way represent an
adequate cross-section of the total mammalian fauna within the area during
historic times, they clo offer important clues to the environment of that
area during the time the fissure traps were in operation. Remains of such

1958

Guilday and Bender: Fissure Deposit in Pennsylvania

131

forms as the flying squirrel, red squirrel, gray squirrel and porcupine, and
the high numbers of chipmunk and pine mouse remains (See discussion
under Pitymys) all point directly to a former forest cover. The total fauna is
one that would be expected in the area at the present day, with the obvious
exception of those forms which were exterminated by man. In this instance
the forested nature of the country during the recent past is well known
and the fact comes as no great surprise, but it does serve to strengthen the
reliability of fossil faunas as indicators of past climatic and ecologic condi-
tions, provided the habitat requirements of the component species are well
enough known.

ANNOTATED LIST OF VERTEBRATE REMAINS
(Unless otherwise noted, all specimens are from sink-hole no. 2)

amphibians

Remains of three large toads, Bufo sp., represented by tibiotarsi and skull
fragments. These agreed closely with specimens of Bufo americanus used
for comparison and are probably referable to that species.

Two tibiotarsi were referred to the genus Rana. The wood frog Rana
sylvatica was found living in the sink-hole and the surrounding woodland.

reptiles
Order Chelonia
Family Emydidae

Box turtle. Terrapene Carolina (Linnaeus)

Remains of four individuals. Shell fragments of one individual were re-
covered from sink-hole no. 3.

Order Serpentes
Family Colubridae
Black rat snake. Elaphe obsoleta Say

One right maxilla, one left mandible and vertebra could not be distin-
guished from those of a 52-inch female black rat snake collected from the
fissure area during the course of the field work.

Family Crotalidae
Timber rattlesnake. Crotalus horridus Linnaeus

At least three individuals. Represented by skull elements which include
three right and one left maxilla, mandibles and numerous vertebrae.

The unidentified material, largely vertebrae, was kindly examined by
Neil D. Richmond who reports as follows:

Most of the snake vertebrae are Crotalidae. These appear to be of two types — a
light, more delicate type and a large massive type that are referable to Crotalus.
The maxilla and other bones of the skull are definitely referable to Crotalus.
Whether the more delicate vertebrae are young Crotalus or are those of Ancistrodon
(copperhead) can not be determined with the material at hand.

The most numerous Colubrid vertebrae are those of Coluber. The vertebrae from
the sinks agree in minute detail with Pennsylvania Coluber constrictor skeletons.

One lot of several vertebrae agrees with Elaphe.

Two lots of vertebrae, apparently of quite different ages as one is discolored by
iron salts, agree with Heterodon (hog-nosed snake).

Several small Colubrid vertebrae have been left unidentified as they were not
found together and are probably anterior thoracic vertebrae of one of the other
Colubrid snakes.

132

Annals of Carnegie Museum

vol. 35

BIRDS

Order Strigiformes
Family Strigidae

Great horned owl. Bubo virginianus
One partial skeleton.

Unidentified. Partial skeleton of one small perching bird.

MAMMALS

Order Marsupialia
Family Didelphiidae
Opossum. Didelphis marsupialis virgin iana Kerr

No opossum remains were recovered from sink-hole no. 2. One skull and
one cervical vertebra of an adult opossum were picked up from the super-
ficial deposits in sink-hole no. 3. It should be noted that the latter fissure had
been open to the surface at the time of discovery. The opossum is a recent
addition to the fauna of Pennsylvania. With the opening up of the country
and the removal of much of the primeval forest the opossum spread to the
north. It is a common animal in Bedford County today but was rare or
absent during prehistoric times. Parker (1953) reports the finding of an
opossum “in an advanced state of purification” in one of the fissures sub-
sequent to the Museum excavations.

Order Insectivora
Family Soricidae

Long-tailed shrew. Sorex sp.?

At least three individuals.

Short-tailed shrew. Blarina brevicauda (Say)

At least 31 individuals represented by 74 skull and mandible fragments.

Family Talpidae

Hairy-tailed mole. Parascalops breweri (Bachman)

One right clavicle and one right humerus.

Order Chiroptera
Family Vespertilionidae
Little brown bat. Myotis sp.?

At least three individuals.

One bat skull, without mandibles, from sink-hole no. 3, was identified by
Dr. J. Kenneth Doutt as Myotis , probably lucifugus.

Bat. species?

At least two individuals.

Order Lagomorpha
Family Leporidae

Cottontail rabbit. Sylvilagus floridanus (J. A. Allen)

Two complete skulls. In addition, one complete skull from sink-hole no. 3;
this appeared to be extremely recent in origin compared with other bones
from the same fissure.

New England cottontail rabbit. Sylvilagus transitionalis (Bangs)

One complete skull and one pair of frontal bones seem to be referable to
this species. Both are of adult animals and in both the supraorbital processes
agree in size and shape with those of the transitionalis skulls used for com-
parison. In addition, two complete skulls from sink-hole no. 3 are referred
to this species.

1958

Guilday and Bender: Fissure Deposit in Pennsylvania

133

Cottontail rabbit. Sylvilagus (either floridanus or transitionalis )

At least 13 individuals represented by 48 skull and jaw fragments, plus 87
postcranial bones and fragments. Seven mandibles, five skull fragments,
and 34 postcranial bones from sink-hole no. 3.

There is some evidence that S. transitionalis was the only form of the cot-
tontail rabbit present in Bedford County during prehistoric times (Rhoads,
1903). Subsequent destruction of suitable habitat within the area has reduced
its former range. The New England cottontail is now rare and local in Bed-
ford County (Gifford and Whitebread, 1951, p. 31).

Order Rodentia
Family Sciuridae

Woodchuck. Marmota monax (Linnaeus)

At least eight individuals represented by four skulls, 13 disassociated lower
jaws and 23 postcranial bones. During the period of excavation a wood-
chuck was found dead at the bottom of the fissure. It is preserved as a
study specimen in the Section of Mammals, Carnegie Museum. Two mandi-
bles and seven postcranial bones from sink-hole no. 3.

Chipmunk. Tamias striatus (Linnaeus)

At least 109 individuals, represented by 139 skulls and skull fragments,
215 mandibles and 553 postcranial bones. This was the commonest mammal
encountered in the deposit. One right mandible was recovered from sink-
hole no. 3.

Gray squirrel. Sciurus carolinensis Gmelin

Two skulls are certainly this species. Four skull fragments, six mandibles
and 15 postcranial bones are referred on the basis of size. A total of four in-
dividuals is represented.

Red squirrel. Tamiasciurus hudsonicus (Erxleben)

One complete skull from sink-hole no. 3.

Flying squirrel. Glaucomys volans (Linnaeus)

At least 10 individuals, represented by 27 skull and mandible fragments
and 50 postcranial bones. It is conceivable that the northern flying squirrel,
Glaucomys sabrinus (Shaw), could be represented as well in this collection.
It has been taken as close as 56+ miles at McGees Mills, Clearfield County,
Pa. (Roslund, 1951). All of the flying squirrel bones from the fissure agree
in size with specimens of Glaucomys volans, but the skeletal differences be-
tween the two species are too subtle to permit an accurate specific identifica-
tion of the more or less fragmentary material. Glaucomys sabrinus has not
been taken in the Ridge and Valley Section of Pennsylvania to date (Gifford
and Whitebread, 1951).

Family Cricetidae

White-footed mouse. Peromyscus sp.?

At least 28 individuals represented by 74 fragments of skull and mandibles.
The species is undoubtedly either Peromyscus leucopus (Rafinesque) or
Peromyscus maniculatus (Wagner). An additional mandible was recovered
from sink-hole no. 3.

Lemming vole. Synaptomys cooperi Baird

One individual, represented by a skull and associated lower jaws.

Meadow vole. Microtus pennsylvanicus (Ord)

Three individuals represented by three skulls.

134

Annals of Carnegie Museum

vol. 35

Pine vole. Pitymys pinetorum (LeConte)

Fifty-one individuals, represented by 48 skulls, four additional skull frag-
ments and three mandibles.

Remains of small mammals are generally assumed to be more reliable
indicators of past environmental conditions than are those of the more
widely ranging larger mammals (Hibbard, 1955, p. 200). The high percent-
age of pine vole remains and the low percentage of meadow voles is inter-
esting. The pine vole is much more of a burrowing animal than is the
meadow vole and appears to be limited to areas of loose friable soils irre-
spective of the type of ground cover. In the mountainous areas of Pennsyl-
vania it is local in distribution and more commonly met with in heavily
wooded areas than is the meadow vole which prefers a more open meadow
habitat and is rarely taken by the collector in wooded areas. These require-
ments are approximations of optimum habitats and in periods of high vole
abundance may have lessening significance. The ratio of 17:1 (51 Pitymys ,
3 Microtus) in favor of the pine vole reflects the forested nature of the area
during the time the sink-holes were open to the surface. Both of these rodents
are present in Bedford County today, Microtus being the more common, but
the amount of woodland is drastically different now from what it was in
prehistoric times. In 1945 only 35.8% of Bedford County was classed as
forested (Gifford and Whitebread, 1951). '

Meadow mouse or pine mouse. Microtinae sp.?

Ninety-six mandibles and three skull fragments were either too fragmen-
tary or had lost too many teeth to permit a satisfactory identification closer
than subfamily. Most of the mandibles undoubtedly belong to the skulls
listed above as Pitymys pinetorum. One Microtine mandible was recovered
from sink-hole no. 3.

Muskrat. Ondatra zibethicus (Linnaeus)

Remains of at least one individual, represented by one skull and four
postcranial bones from sink-hole no. 3.

Muskrats wander widely at times. The hilly nature of the sink-hole area
is and was entirely unsuitable as muskrat range. This was undoubtedly a
chance inclusion of a stray animal and has no bearing upon the past en-
vironment of the fissure area.

Family Erelhizontidae
Porcupine. Erethizon dorsatum (Linnaeus)

At least two individuals represented by one skull, four mandibles and 12
postcranial bones. At least five individuals represented by two skulls, seven
mandibles (two left, five right), 10 isolated teeth and 13 postcranial bones
from sink-hole no. 3.

Order Carnivora
Family Canidae

Timber wolf. Canis lupus lycaon Schreber)

Two individuals, both adult but varying somewhat in size, were recovered.
The bones were badly shattered and the remains of the two wolves were
intermhigled. The fragmentary skeletons were found immediately beneath
an articulated, undamaged elk skeleton.

1958

Guilday and Bender: Fissure Deposit in Pennsylvania

135

Gray fox. Urocyon cinereoargenteus (Schreber)

One adult and four pups, possibly litter mates. Represented by two skulls,
two skull fragments, eight mandibles and eight postcranial bones. At least
one individual (adult) represented by six fragments, from sink-hole no. 3.

Family Ursidae

Black bear. Euarctos americanus (Pallas)

One partial skull and associated teeth. This was a young animal, still re-
taining the milk dentition and with open sutures. In sink-hole no. 3 a right
mandible, a right maxilla, and one right humerus of a large and, judging
from the amount of tooth wear, very old black bear.

Family Procyonidae
Raccoon. Procyon lotor (Linnaeus)

One upper molar and one partial right mandible of an immature animal.

Family Mustelidae
New York weasel. Mustela frenata Lichtenstein

One skull and one left mandible. Judging from its large size this specimen
was a male.

Striped skunk. Mephitis mephitis (Schreber)

At least five individuals (four adult and one juvenile) represented by
four skulls, five mandibles and 1 1 postcranial bones. In addition, three
postcranial bones from sink-hole no. 3.

) Order Artiodactyla
Family Cervidae

Elk. Cervus canadensis canadensis (Erxleben)

Two articulated skeletons were found. The posterior half of a skull bear-
ing the antler bases, several segments of antler, a humerus and two cervical
vertebrae were found by J. Howard Taylor of New Paris, Pa., during his
original investigations of the fissure. The remaining bones of the same skele-
ton were recovered by the members of the Pittsburgh Grotto, National
Speleological Society. Both collections are now at Carnegie Museum. This
was an adult male. Although the antlers were fragmentary, enough remains
to show they were of some size and were firmly attached to the skull at the
time of death. The antlers were broken from the skull at a point out on the
antlers themselves, not at the pedicel where natural shedding occurs. With
the exception of the antlers, which were badly rotted, the skeleton was in
an excellent state of preservation. The following bones were missing: pre-
maxillae; left maxilla and tooth row; nasals; palate; zygoma; cervical verte-
brae, numbers one, two and four; thoracic vertebrae, number one; caudal
vertebrae, all missing; 12 phalanges; all “dew claws”; ribs number 11, 12
and 13; two fibulae; right metatarsal; right patella; left radius; left meta-
carpal; all carpals and hyoids. Cervical vertebrae numbers six and seven were
completely coalesced as a result of regenerative processes involved in the
healing of what must have been a severe arrow wound. The flint arrowhead,
with its base broken off, was embedded at the base of the left transverse
process of the sixth cervical. The animal aparently survived and the wound
had completely healed. The arrowhead was so extensively enveloped in new
bone tissue that it can not be removed without destroying some of the ver-
tebra itself. An x-ray of the specimen, taken by Dr. J. Kenneth Doutt and

136

Annals of Carnegie Museum

vol. 35

Dr. E. G. Meisel, yielded no additional information. Mr. Taylor is to be
congratulated for his efforts to preserve this unique and important specimen
as well as for his generosity in presenting it to Carnegie Museum.

The specimen is illustrated in Fig. 2.

The second skeleton, that of a doe, was found by A. C. Lloyd of the
Museum field party. It was a more nearly complete skeleton than that of
the bull. Its skull, protected from crushing by the surrounding rocks, was
recovered intact.

Elk remains were also recovered from sink-hole no. 3. These were dis-
articulated, represented parts of more than one individual and in some
cases were badly shattered. They include the following: 4 upper molars, 3
upper premolars, sacrum, 13 vertebrae, 7 ribs, 2 sternal ribs, 3 sternebrae,
1 pelvis, 2 femora, 2 tibiae, 2 metatarsals, 2 humeri, 2 radii, 2 ulnae, various
carpals, tarsals and phalanges. Also a large number of bones of a very young
elk fawn, including a complete but disarticulated skull.

White-tailed deer. Odocoileus virginianus (Boddaert)

Remains of three individuals, presumably two antlered bucks and one
fawn. Represented by five skull fragments, three broken antlers and 14 post-
cranial bones. In addition, a right tibia, a patella, fragmentary scapula,
teeth and sacrum of apparently one adult, as well as one right mandible of
a fawn were recovered from sink-hole no. 3.

There is ample evidence in Pennsylvania and neighboring states for cave
deposits dating from the Pleistocene (Hay, 1923). While the fauna recovered
to date from the New Paris fissures is recent and of no geological antiquity,
it is possible that further prospecting along Chestnut Ridge (Bedford
County) would be profitable. The area is rife with fissures and it is possible
that some of them might contain the fossil remains of Pleistocene mammals.

REFERENCES

Bender, Martin S.

[1955.] The carnivores of the New Paris sink-holes. Ill p.

Master’s thesis at University of Pittsburgh.

Dice, Lee R.

1943. The biotic provinces of North America. University of Michi-
gan Press, Ann Arbor, 78 p.

Gifford, Clay L., and Ralph Whitebread

1951. Mammal survey of south central Pennsylvania. Final Report
Pittman-Robertson Project 38-R. 75 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Guilday, John E.

1948. The fallen monarch. Carnegie Magazine, v. 22, no. 2, p. 46-49.
Hay, Oliver P.

1923. The pleistocene of North America and its vertebrated animals
from the states east of the Mississippi River and from the Ca-
nadian Provinces east of longitude 95°. Carnegie Institution
of Washington. 499 p. (Publication no. 322.)

1958

Guilday and Bender: Fissure Deposit in Pennsylvania

137

Fig. 2. Sixth and seventh cervical vertebrae of an adult male elk ( Cervus
canadensis ), showing embedded arrowhead

A. Spinous process, 7th cervical. B, B\ Demifacets for the 1st ribs. C. Centrum of 7th
cervical. D, D'. Postzygapophyses, 7th cervical. E, E'. Transverse process, 6th cervi-
cal. F. Embedded arrowhead. F'. Tip of arrowhead.

138

Annals of Carnegie Museum

vol. 35

Hibbard, Claude W.

1955. The Jinglebob interglacial (Sangamon?) fauna from Kansas and
its climatic significance. Contributions from the Museum of
Paleontology, University of Michigan, Ann Arbor, v. 12, no. 10,
p. 179-228.

Parker, John D.

1953. Bone caves in Pennsylvania. The American Caver, December,
p. 10-14. (National Speleological Society, Bulletin 15.)

Rhoads, Samuel N.

1903. The mammals of Pennsylvania and New Jersey. Privately print-
ed, Philadelphia. 266 p.

Roslund, Harry R.

1951. Mammal survey of north central Pennsylvania. Final Report
Pittman-Robertson Project 37-R. 55 p. Pennsylvania Game
Commission, Harrisburg, Pa.

Stone, Ralph W.

1932. Pennsylvania caves. Pennsylvania Geological Survey, ser. 4, Bul-
letin G3, ed. 2. p. 1-135.

1953. Descriptions of Pennsylvania’s undeveloped caves. The Ameri-
can Caver, December, p. 51-137. (National Speleological So-
ciety, Bulletin 15.)

.P V

JVTHSO^^

ART. 10. ARCHAIC HUNTERS
UPPER OHIO VALLEY*
By Don W. Dragoo 'v
Assistant Curator of Section of Man, Carm

PREFACE

The Archaic cultures of the eastern United States are of great importance
in the understanding of the peopling and cultural development of the New
World. Until recently, most of our knowledge of the Archaic was based on
the work of William S. Webb, T. M. N. Lewis, and Madeline Kneberg in
Kentucky and the Tennessee Valley, and the studies of William Ritchie in
New York state. In the vast region between these two distant areas little
was known of the Archaic cultures. The cultures described for the two
known areas appeared to be so different that only vague references were
made to their similarities.

This report presents the results of recent studies of Archaic cultures con-
ducted by Carnegie Museum, Pittsburgh, in the Upper Ohio Valley. This
region, encompassing portions of the states of Pennsylvania, Ohio, New
York, West Virginia and Maryland, is a key link between Kentucky and
Tennessee Valley Archaic and Northeast Archaic. Answers to the following
questions were a prime purpose of this project.

1. What anthropological concepts are important to the understanding
and definition of the Archaic?

2. What Archaic cultures were present in the Upper Ohio Valley and
what were their cultural, temporal, and spatial relationships to each other?

3. What were the relationships of the Upper Ohio Valley Archaic cul-
tures with those of surrounding areas?

4. What were the relationships of the Archaic cultures with the earlier
Paleo-Indian?

5. What were the origins and routes of migration for the Archaic cultures
into the Upper Ohio Valley and the Northeast?

Since most of the Archaic sites in the Upper Ohio Valley are small and
contain no depth of debris, the excavation of the Dixon and Rohr shelters
produced a sequence of Archaic and later materials which has become an
important aid in the chronological ordering of many small sites known only
by surface collections.

The information gained from the Dixon and Rohr excavations has made
it possible to revaluate several recently reported sites. In a number of cases,
new interpretations of the cultural and temporal relationships of these sites
are stated in this article.

Since I began this project in 1954, I have had the valuable aid and assist-
ance of many people. I wish to express my appreciation to Albert Bauer,
Betty Bauer, John Leppla, and Dorothy Dragoo for their labors in the Dixon
and Rohr excavations. William S. Webb, of the University of Kentucky,
kindly made available to me artifacts and depth locations of these artifacts
for Kentucky Archaic sites. Through the courtesy of Glenn A. Black, of

* Portions of this article appeared in “Archaic Cultures of the Upper Ohio Valley,”
a doctoral dissertation presented to Indiana University, June 1957.

139

Submitted for publication, January 17, 1958
Issued January 5, 1959

SMITH A BD | "ft

140

Annals of Carnegie Museum

vol. 35

the Indiana Historical Society, the materials of the McCain site were loaned
to me for restudy. I wish to express my appreciation also to Paul Gebhard,
Dorothy Cross, Irving Rouse, William A. Ritchie, William S. Webb and
John Witthoft for information on specific problems encountered during the
writing of this report.

I especially want to extend my thanks to James L. Swauger and Mrs.
Dorothy Dragoo for their aid in the preparation of the manuscript. Their
criticisms and suggestions were invaluable.

INTRODUCTION
What is the Archaic?

The term “The Archaic” was first used by William A. Ritchie (1932) in
his classification of certain early manifestations in New York state. Since
its inception the term has seen considerable usage. It has been both accepted
and rejected. It is a term that has been subjected to several conflicting inter-
pretations.

Ritchie’s (1944, p. 321) first extensive formulation of the Archaic was
as follows:

“The long postulated archaic level in New York recently confirmed by intensive
work in the Southeast consists of an aggregate of discrete foci, sharing a hunting-
fishing-gathering economy. Its chief characteristics are the absence of horticultural
traces, ceramics, and the smoking pipe.”

Ritchie (1944, p. 319) presented certain specific elements which he believed
typical of the Archaic in New York.

“Analysis of the archaic level as a whole shows (a) a large variety and numerical
abundance of chipped stone types; (b) the lack of all the so-called “problematical”
group of polished stone artifacts, except the bannerstone of several simple forms;
(c) a considerable typological range in and a large number of bone tools; (d) the
prevalence of copper tools and the total lack of copper ornaments; (e) the general
absence of shell artifacts; (f) the complete dearth of pipes; (g) the want of pottery
except perhaps in the closing phases; (h) the non-existence of agricultural traces;
and (i) the large variety of burial practices, generally not involving mortuary
offerings.”

Ritchie (1955, p. 3) defined his most recent concept of the Archaic in the
Northeast as follows:

“I conceive of the Archaic as essentially a cultural stage of several thousand years
duration prior to the introduction of ceramics, the smoking pipe and horticultural
practices, correlated with a mobile or seminomadic way of life, with small band
organization, simple social structure, flimsy settlements, and a food-seeking sub-
sistence pattern. I do not suppose that this was abruptly terminated or replaced,
or that it everywhere yielded to identical changes, certainly not simultaneously. It
is amply evident that the Archaic pattern of existence persisted in certain relatively
isolated districts of the Northeast long subsequent to the establishment of the
Woodland pattern elsewhere in the area, also that many later manifestations are
firmly rooted in this older soil.”

Thus for Ritchie (1955, p. 4) the Archaic was essentially a developmental
stage between the simpler nomadic Paleo-Indians and the more complicated
semi-sedentary way of life developed by the agricultural Woodland groups.
Ritchie believes that the Archaic in the Northeast was forest adapted at all
times, but it also probably underwent some modifications in response to

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

141

climatic and physiographic changes. For Ritchie the Archaic is not uniform
but comprises a complexity and variety in the details of its equipment be-
cause of the interaction between dissimilar traditions (Lamoka and Lauren-
tian) in the Northeast and with other connections outside the area.

While viewing the Archaic as a stage in man’s development in the North-
east, Ritchie (1955, p. 9) also called upon local environmental conditions
and independent invention as factors in the differentiation of the Archaic
into foci or phases. He does not conceive of several regional variations as
developing in a vacuum, but stresses the interplay of the process of selective
diffusion among semi-isolated and localized centers.

Ritchie’s term “Archaic” was readily taken over by several workers in the
lower Ohio Valley and the Southeast. After the excavation of many non-
pottery sites in the Tennessee Valley, Webb and Dejarnette (1942, p. 319)
stated:

“Since the first report of a non-pottery site in New York, other such sites have
been investigated . . . Ritchie has designated this pattern as Archaic, partly on a
basis of stratigraphy and because of the absence in the complex of agriculture and
pottery . . . Because it now appears that this non-agricultural, non-pottery, hunter-
fisher-collector pattern of culture may have been widespread in Eastern United
States in early aboriginal times, the term “archaic” is here adopted to designate this
pattern manifestation in Kentucky and Alabama.”

Implicit in Webb’s writings is his belief that the Archaic has three basic
criteria— age, hunter-fisher-collector economy, and the absence of ceramics.
In a report by Webb and Haag (1940, p. 109) it was also implied that the
Archaic constituted a level or stage in culture.

Other workers in the Southeast such as Lewis and Kneberg (1947, 1955),
and Fairbanks (1942) accepted the term Archaic for cultural manifestations
representative of a hunter-fisher-collector economy devoid of ceramics. Im-
plied in the writings of these workers as well as in the works of Webb is
the concept of a culture level or stage with rather definite time or age con-
notations. After having excavated many deeply stratified sites in the South-
east, it would have been difficult for these workers not to have had some
definite concept of age for the Archaic in their formulations.

The first major objection to the use of the term Archaic was made by
Griffin (1946, p. 42). He believed that the artifact assemblage of Archaic
was not typologically Archaic and that a great majority of the pre-ceramic
units of the eastern United States were relatively late in the pre-ceramic
cultural development of the country as a whole. For Griffin the term
Archaic was as much a misnomer for the eastern United States as it was
in middle America where it had been inappropriately applied to a late
cultural assemblage. The term Archaic should have been reserved for the
oldest manifestations in North America when such materials became known.
Griffin (1946) in his paper on change and continuity in eastern United States
preferred to discuss what other workers were calling Archaic as part of the
Paleo-Indian problem.

Griffin (1952, p. 352-364) later dropped his opposition to the term Archaic
and presented a series of prehistoric periods for the eastern United States
which he called the (1) Paleo-Indian, (2) Archaic, (3) Early Woodland, (4)

142

Annals of Carnegie Museum

vol. 35

Middle Woodland, and (5) Mississippi and Late Woodland. Under this
scheme Griffin also recognized both early and late Archaic periods, each with
a number of distinctive subdivisions. For Griffin each of the above men-
tioned periods represents a stage or level in the progressive evolution of
culture.

William H. Sears (1948, p. 122-124) expressed a number of adverse criti-
cisms of the use of the term Archaic. His argument was that Ritchie’s de-
tailed trait list for the Archaic did not serve to distinguish the complex
from Early Woodland levels, and thus it could not be used to establish a
distinct Archaic pattern. He believed that the concept of the Archaic as a
complex which is non-ceramic, non-horticultural, old, and has a hunter-
fisher-collector economy to be “entirely too broad to be of any use.”

Sears (1948, p. 124) was viewing the problem of Archaic cultures from the
standpoint of its inclusion within the framework of the Midwestern Taxo-
nomic Classification established by McKern and other workers in the eastern
United States. Important to the function of this classification is the prepara-
tion and comparison of trait lists for many sites. Sears prepared trait lists
for many sites classified as Archaic which revealed that their relationship
to each other was sufficient to include them all in a pattern. He then com-
pared his Archaic trait lists with the statement of the First Conference on
the Woodland Pattern (Angnymous, 1943, p. 393-400). His conclusion was
that there were some variations, but except for ceramics in Woodland there
were no differences of importance. In regard to ceramics, Sears pointed out
that in many instances Woodland ceramics appear in the closing phases of
a manifestation with no accompanying change in the non-ceramic inventory.
As examples he gave Ritchie’s Brewerton sites and Webb’s Pickwick Basin
sites. Sears’s (1948, p. 124) conclusion was as follows:

“It appears then that the manifestations which have been called the Archaic
represent the early phases of the continuum known as Woodland. It is not, however,
necessary to classify them in the Woodland Pattern. This depends on the criteria
for the establishment of, or inclusion in, a pattern, which is after all a classificatory
device, and on the fact that addition of non-ceramic manifestations to a complex
whose most useful criterion is ceramics, would confuse things considerably. A
compromise might be to group these non-pottery sites under some such term as
Pre- or Non-Pottery Woodland. This would keep the groups separated termino-
logically and at the same time indicate the relationship. In general, I would be
inclined to accept the non-pottery section of the trait list offered by the First
Conference on the Woodland Pattern as the most useful set of criteria for the
delineation of this non-ceramic pattern or section of the Woodland Pattern.”

It is evident that in 1948 Sears viewed the Archaic as only the early or
beginning phase of the Woodland Pattern. Woodland was seen as a cul-
tural continuum with both non-ceramic and ceramic phases. It is, indeed,
true that there are sites on which pottery makes its appearance without greatly
changing the non-ceramic inventory, but I view such sites as being on the
periphery of areas in which the ceramic Woodland cultures were becoming
established. Only certain Woodland traits, such as ceramics, had diffused
to these areas where they were incorporated into long-established Archaic
complexes; There is ample evidence in nearly all of these areas to indicate
that the Archaic way of life soon gave way to new ways of living.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

143

Equally important as the introduction of ceramics, agriculture, smoking
pipes, and other new cultural items was the arrival of peoples of a physical
type different from that of the Archaic (Neumann, 1952, p. 13-34). Un-
doubtedly, many of the peoples and traits of the Archaic became part of
this new way of life. In view of the evidence for the introduction of several
new and important culture elements as well as new peoples, I certainly
hesitate to call the Archaic the non-pottery section of the Woodland Pattern.
A recent publication by Sears (1954, p. 23-36) indicated that he also has
accepted the Archaic as an established cultural entity.

Just as there has been confusion of Archaic and Woodland, there has been
confusion of Archaic with Paleo-Indian. The term Paleo-Indian was origi-
nally used by Frank H. H. Roberts (1940, p. 51-116) for American cultural
assemblages which appeared to be chronologically early on the basis of
geologic, faunal, or typological evidence. As was previously mentioned,
Griffin (1946, p. 42) extended Roberts’s use of Paleo-Indian to include com-
plexes generally known as Archaic. John Witthoft (1952, p. 465) took note
of this confusion and stated the following:

“The term Archaic may lead to some confusion because this word has been used
to include what I call Paleo-Indian, even as Paleo-Indian has been used to include
what 1 call Archaic. I think of the local archeology as classified first in a chrono-
logical framework and only after that in terms of intra-areal relationship. Such
an outline includes three major technological stages, which equate, to the best of
our knowledge, with major time divisions. The Paleo-Indian Epoch, so far not
divisible into periods, lacked stone pecking and grinding techniques. The Archaic
Epoch, provisionally divided into two periods, has as its ‘index fossils’ ground
stone tools but lacks pottery. The Woodland Epoch, now seen as four or five
periods, was characterized by pottery-making. This triple division seems to be
valid almost throughout the continent; it is not to be considered as a major tenet
of American archeology, but rather as a tool for segregating cultures into natural
classes for further study.”

Witthoft finds good technological reasons for separating Archaic from
either Paleo-Indian or Woodland cultures. He sees three major technological
advances which have chronological significance in the progressive evolution
of culture in the New World. Each of these divisions he would term an
“Epoch,” but he cautions that these divisions are only tools for segregating
cultures into “natural classes” for study. Witthoft (1952, p. 465) rejected
the Midwestern Taxonomic System as a method for ordering archeological
data. Thus, he did not fall into the statistical pitfall that led Sears (1948)
to proclaim Archaic as the early non-ceramic part of a Woodland continuum.

A survey of recent literature indicated that the term “Archaic” has sur-
vived the many criticisms leveled against its use. To use Witthoft’s term,
it has become an accepted “Epoch” in the culture history of the eastern
United States. During the past few years many contributions have been
made which have aided the understanding of the content and form of many
manifestations of the Archaic. Equally important has been the attempt to
state and define the theoretical foundations on which the Archaic, or any
other culture “Epoch” or “stage,” is based. In this matter, a recent paper
by David A. Baerreis (1955, p. 1) warrants attention.

Baerreis gave his definition of an Archaic “stage” as being represented by

144

Annals of Carnegie Museum

vol. 35

cultures with a hunting and gathering economy and lacking horticulture.
The appearance of ceramics in an area was taken as a terminal date or
horizon marker as cultures representing an Archaic stage may persist even
into historic times. Baerreis believed that the appearance of horticulture
constitutes an alternative and more logical terminal horizon marker than
the appearance of pottery but he pointed out that the initial acceptance of
agriculture is difficult to recognize in the present state of archeological
knowledge.

Baerreis also called attention to the fact that the term Archaic can be
used in at least two primary senses. First, it may have reference to a specific
stage of cultural development characterizing a regional sequence. Second,
Archaic may be used in reference to a cultural tradition which represents a
particular configuration of culture traits that persist through a number of
culture stages. Baerreis warned against speaking of an archaic tradition for
he believed that the archaic stage may comprise many recognizable cultural
traditions. The spread of distinct traditions or the diffusion of specific ele-
ments from these traditions may have created the impression of develop-
mental sub-stages within the archaic.

It is evident that Baerreis took as granted the existence of culture evolu-
tion. His acceptance and use of the term stage indicated his belief in the
progressive development of culture through time. He was, however, very
much aware of factors such as diffusion and environmental influences in
the shaping of any culture of a specific area. Speaking of the differences
seen in the Archaic Baerreis (1955, p. 10) stated:

“No single answer can be given if a question as to why these differences appear
is raised. A simple solution might be to say that we are dealing with a later time
period when a higher level of cultural growth and complexity had been attained.
Yet radiocarbon dates are pushing the limits of Archaic cultures farther back in
time so that at present time we are beginning to see a clear temporal overlap in
the two broad culture types (Paleo-Indian and Archaic). Differences in environ-
mental factors may also be posed and certainly they do have their influence on
the total configuration. Yet it would be pushing an already weak explanation to
the point of absurdity to postulate, for example, that the greater diversity of game
in the eastern United States could be responsible for a more heterogeneous assem-
blage of projectile point types and to extend this reasoning to other aspects of the
culture. The problem would appear to largely revolve around differential situations
that facilitated acculturation in one region in contrast to the other.”

As to how differential situations have aided or retarded acculturation
Baerreis (1955, p. 10) suggested that the differences that set apart Paleo-
Indian and Archaic culture types are perhaps due to basic differences in
social structure. The economy of Paleo-Indian groups was based upon hunt-
ing. It is assumed that there were a number of widely scattered patrilocal
bands. Because only a small number of male adults would have been manu-
facturing projectile points the techniques would tend to have been perpetu-
ated within the patrilocal band. Only women would have been directly in-
volved in bringing in new ideas from the outside. Among Archaic groups,
seed gathering was of importance as well as hunting. The importance of
women, who are usually the gatherers, in the economy may have caused a
shift from patrilocal to matrilocal, or at least bilocal, bands. The marriage

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

145

pattern would then have brought in new males to the local group and
with them new ideas concerning styles of projectile points and other items.

Griffin (1952, p. 354) also attempted to reconstruct the social organization
of the Archaic; but, unlike Baerreis who saw women’s part as collectors as
highly important in the Archaic, he postulated a patrilineal descent for the
Archaic on the basis of the male dominance in hunting and fishing activities.
I look favorably upon Baerreis’s suggestion of the presence of matrilocal or
at least bilocal bands in the Archaic. Most Archaic sites produce abundant
evidence (charred seeds and nuts, grinding stones, etc.) pointing to the
importance of seed gathering and other collecting activities.

Before passing on to a working definition of the Archaic as a guide to
this present study, it is necessary that we look more closely at the concepts
of “stage,” “tradition,” and “co-tradition.” As has been noted in previous
paragraphs, these concepts are being strongly emphasized by present-day
American archeologists and are intimately involved in any definition of the
Archaic.

The idea of a series of stages as reflecting the chronological development
of culture is not new to archeology. Following the work of Morgan (1877)
and other evolutionary ethnologists of the nineteenth century, the idea of
stages became exemplified by the Stone-Bronze-Iron classification used by
students of European prehistory. The facts of cultural distribution as now
known have led largely to the abandonment of evolutionary schemes as a
means of presenting chronology, but such schemes are still found useful
in expressing the degree of cultural development. In the New World the
evolutionary approach to chronology was recently used by Strong (in Bennett,
1948, p. 98) in the Viru Valley of Peru where he used a series of “cultural
epochs” or stages of cultural development which he believes also applicable
to many other areas of the world. Rouse (1953, p. 72) pointed out that
Strong’s formulation has two advantages over previous studies of the
same kind:

“ (1) in so far as possible, it is based upon observed distributions rather than on
conjecture, and (2) for the most part it is defined in terms of degree of develop-
ment rather than of specific cultural content. It therefore provides a means of
expressing the broad sweep of chronology over a whole series of areas without
bringing in the differences in detail from area to area.”

This view of the cultural stage as outlined above by Rouse is like the one
expressed by Baerreis in his formulation of an Archaic stage. Griffin (1952)
and his associates took a somewhat similar position in their book “Archeology
of Eastern United States.”

Another concept currently receiving wide use by American archeologists
is that of the “tradition.” Kroeber (1944, p. 108-111) applied the term
“horizon style” to the phenomenon of one or more traits extending from
area to area through a series of contemporaneous periods. Willey (1945)
applied the term “tradition” when one or more culture traits survived from
period to period within a single geographic area. The “tradition” for Willey
was the counterpart of Kroeber’s “horizon style.” Willey (1953, p. 374) sum-
marized his concept of “tradition” as follows:

“The theory underlying the tradition concept is well expressed in the term

146

Annals of Carnegie Museum

vol. 35

itself. For “tradition” implies deep-set and channeled activity or patterned ways
in which the vitality of a culture expresses itself in strong preference to other
possible ways. The conditions surrounding this rigidity of expression which results
in the long-time traditional expression are an interesting problem for future in-
vestigation. In dealing with such things as pottery decoration, the archeologist is
undoubtedly investigating what is relatively trivial in past human events. The
failure of polychrome painting to take hold on the northern coast of Peru in
the face of the white-on-red tradition, despite several attempts to introduce it,
may eventually be revealed as nothing more mysterious than an absence of suitable
mineral pigments in that part of the country. On the other hand, the unraveling
of seemingly insignificant threads in an attempt to factor out causality may lead
us to a greater understanding of ‘tradition set’ and ‘tradition persistence’ in
institutions which loom larger in human affairs.”

Related to the concept of tradition is the idea of the area co-tradition
(Bennett, 1948). The co-tradition is based upon the persistence of a number
of closely interrelated traditions found within a specified area. It presents
a means by which the archeologist can plot the history of cultural continuity
within the confines of an area. As Willey (1953, p. 374) pointed out, the co-
tradition is the culture area with time depth.

Although the concepts of “stage,” “horizon style,” “tradition,” and “co-
tradition” were formulated by workers in Andean archeology, these terms
are being applied by workers in North America. Martin and Rinaldo (1951)
used the co-tradition concept with success in the southwestern United States,
and Willey (1953) conceived of three co-traditions consisting of the Archaic,
Woodland, and Mississippian in eastern North America.

In this study of the Archaic in the Upper Ohio Valley and its relationship
to other cultural manifestations of eastern North America, I have been
guided in the selection of materials and sites by the following points of
definition for the Archaic:

1. The Archaic is represented by cultures with a hunting, fishing, and
gathering economy, lacking agriculture.

2. The appearance of ceramics in the area is taken as a terminal date
for the Archaic. It is recognized, however, that many Archaic traits may per-
sist through a number of cultures which possess ceramics and agriculture.

3. The lower time limit for the Archaic is difficult to define clearly
because the final phases of Paleo-Indian cultures were often involved in
acculturational situations with early Archaic cultures. In agreement with
Baerreis (1955, p. 2), it seems “preferable to leave the lower limits for the
Archaic flexible and to simply attempt to trace the cultures of the delimited
stage as far back in time as possible.”

4. The Archaic is conceived as a “stage” or “level” in the development of
culture in eastern North America. For practical taxonomic purposes the
Archaic can be studied as if it were independent of lower or higher levels,
but it is imperative that interrelationships between the levels should not
be forgotten. The Archaic is not completely intelligible apart from reference
to the levels below and above it. A higher level, however, can not be reduced
to a lower level merely because it shares traits with the latter.

5. The Archaic is not a “tradition” but a stage in cultural development
in which a number of “traditions” may be recognized. Any tradition or par-

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

147

ticular configuration of culture traits may persist through a number of
culture stages.

6. Differences in environmental factors and acculturational situations are
important considerations in the understanding of any particular Archaic
manifestation or group of manifestations.

A major purpose of this report is an attempt to discover and define the
processes and factors involved in what we know as the “Archaic.” In the
following sections a number of recently discovered Archaic manifestations
in the Upper Ohio Valley are described and compared with others known
in eastern North America. The theoretical concepts just discussed present a
framework upon which an understanding of the Archaic can be constructed.

THE DIXON ROCK SHELTER

While I was on a Carnegie Museum survey trip to the Cheat River region
of northeastern West Virginia in early November 1953, an interesting rock
shelter on the Preston County farm of Jesse Dixon was brought to my atten-
tion by Albert H. Bauer of McKeesport, Pa. This site was named for the
owner, Dixon, and given the Upper Ohio Valley Archeological Survey num-
ber of 46Pr6. The shelter protected a considerable area from rain and chill-
ing winds. On the floor was the scattered debris of both Indian and white
occupation. According to local residents the shelter was once used as a
base by horse thieves, and only a few years ago a destitute family spent a
winter under the cliffs protection.

The geographic location of the shelter and the materials found on the
surface of its floor were of special interest. The site being near the limits
of the Ohio River drainage and in close proximity to the upper reaches of
the Potomac Valley presented this problem. Would the cultural materials
be typical of those found in the Upper Ohio Valley or would they resemble
more closely those of the Potomac area? Here was an opportunity to observe
cultural movements at the Ohio Valley’s rear door. An examination of the
few artifacts gathered from the shelter gave evidence of cultural change
among the types of pottery and projectile points. Typologically there were
objects ranging from the Late Prehistoric back to non-ceramic times. The
dark color and apparent depth of the soil in the shelter coupled with the
varied artifact types gave us hope that we had found a good stratified de-
posit. If this were the case, we would have a good check on the sequence
and development of human habitation on the eastern fringe of the Ohio
Valley.

With the kind co-operation of the owner, arrangements were made for
Carnegie Museum to excavate at the shelter during the last week of July
and the first two weeks of August, 1954. Mr. Bauer’s tireless participation
in the excavations and the processing of recovered materials did much to
make the project a success. This report gives the results of our partial ex-
cavation of the site.

NATURAL SETTING

The Dixon Shelter is located in mountainous country at an elevation of
1800 feet above sea-level. Many hilltops in the surrounding area reach an

148

Annals of Carnegie Museum

vol. 35

elevation of from 2500 to 3000 feet. The overhanging ledge or roof of the
shelter is composed of the Pottsville conglomerate, the basal formation of
the Pennsylvania period in the area (Hennen and Reger, 1915, p. 177-184).
It is a compact, grayish-white sandstone with many quartz pebbles. Under-
lying the conglomerate is a fine brown sandstone which has eroded extensive-
ly to form the two chambers of the shelter. Approximately one hundred feet
in front of the shelter and responsible for its formation is a small, fast-
flowing stream known as Lick Run. This stream has moved east and cut its
new channel about four feet into bedrock since it eroded the brown sand-
stone. About one-fourth mile south of the shelter, Lick Run enters Roaring
Creek which empties into the Cheat River 33A miles to the southwest. The
Cheat River drains a large portion of eastern West Virginia into the Ohio
River system. From its source at an elevation of 4000 feet to its mouth
where it is only 780 feet, the Cheat is a turbulent mountain stream with
many rapids.

The area is covered with fine forests of hardwoods. Only a small portion
of the land is suitable for farming and has been cleared. On the hilltops
and slopes are great stands of white, black, scarlet, and chestnut oaks mixed
with red maple, elm, and sweet birch. Along the small streams in the valleys
are stands of hemlock crowded by thick growths of mountain laurel and
rhododendron. Sassafras, silky dogwood, black cherry, and sour gum trees
are scattered throughout the area. There are many shrubs and bushes with
edible nuts, fruits, or berries. Among these are the hazelnut, elderberry,
huckleberry, dewberry, blackberry, partridge-berry, deerberry, mountain tea,
and the late low blueberry. Plants of the wild hydrangea and sunflower and
of Indian tobacco (Lobelia inflata ) were seen growing near the shelter.

In prehistoric times the Cheat River region abounded in animal life.
Deer, elk, bear, bob-cat, mountain lion, rabbit, squirrel, raccoon, and many
other species were common. In the streams were black bass, sunfish, brook
trout, catfish, suckers, and snapping turtles. Fresh-water clams were scarce
as the streams are too swift and stony. Wild turkey and grouse were common
and some waterfowl may have inhabited the marshy ponds cradled among
the hills.

EXCAVATION

The overhanging roof of the Dixon Shelter extends approximately 150
feet along the outcrops in a line from southwest to northeast. From the
front of the shelter to the rear wall, the recess varies in depth from a few
feet to as much as 25 feet. Near the front of the shelter the ceiling ranges in
height from 6 to 10 feet above the floor, but it slopes gently to a height of
only two to four feet at the rear wall. Near the center of the shelter is a
large fissure in the conglomerate stratum and a resulting breakdown of the
ceiling. When this breakdown occurred is not known but it is probable that
it has been relatively recent since cultural materials were found under the
edges of the fallen boulders. The collapsed ceiling and the boulders strewn
on the floor divide the shelter into two natural divisions at the present
time. Our excavations were confined to areas not covered by the large
boulders since we had no convenient method of removing them. (See Fig. 1.)

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

149

The area of the shelter was surveyed, and a base-line 150 feet in length was
established parallel with the front of the shelter. This line was then divided
into three divisions each 50 feet in length with division A at the northeast
end of the shelter and division C at the southwest end. Division B was at
the center and almost entirely covered by the large boulders. A grid system
consisting of five- by five-foot squares was established in each division. The
horizontal location of all materials removed in the excavations was recorded
by division and block designation as well as vertical level. (See Fig. 2.)

A datum point was established and vertical measurements were made in
reference to this point. Each excavated block was dug by arbitrary six-inch
levels. The levels of all the blocks dug in division A were designated in
relationship to the datum point. When the soil profiles and the recovered
artifacts were observed it was apparent that the relative position of any
artifact-bearing level with the datum point was of little significance. The
main vertical reference for any object or level of objects was its relative dis-
tance from the surface and the depth of the deposit for the particular block
in which it was found. For example, the debris deposits in the blocks of
division A, near the center of the shelter were as much as twice as deep
as the deposits in the blocks near the northeastern end, but the relative
position of cultural materials remained the same. With this information, all
levels of division A were converted into straight sequence from the measure-
ments made from the datum point, and all levels excavated in divisions B
and C were determined in sequence from the surface. This greatly simplified

Fig. 1. General view of Dixon Shelter before excavation

150

Annals of Carnegie Museum

vol. 35

the handling of the materials when the detailed study was begun.

All excavating was done by customary archeological methods. Trowels
were used for nearly all primary digging. Many objects were recovered as
the soil was loosened but all earth was passed through a screen as a safeguard
since available light did not permit proper viewing. Observations of soil
differences, features, and cultural materials were recorded for each level
within a block and, where necessary, drawings and photographs were made.
All open excavations were refilled when work was completed.

Fig. 2. Dixon Shelter. Division A excavations in progress

Each major area excavated presented its own peculiar problems. In the
following paragraphs the excavation in each division is discussed in detail.
Division A

Division A included all of the clear, excavatable area of the shelter north-
east of the collapsed ceiling. Digging began near the margin of the fallen
boulders and continued to the northeast end of the shelter, a distance of
approximately forty feet.

Immediately below the edge of the overhang and on the floor of the
shelter was a deposit of material broken from the ceiling and washed from
the surface above the outcrop. This deposit appeared as a ridge following
the outline of the overhang. Testing showed this area deposit to be stony
and relatively clear of cultural debris. Major digging was confined to the
floor area between this ridge of rubble and the rear wall of the shelter.

Ten whole or partial blocks were dug in division A. The cultural debris

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

151

in the blocks nearest the center of the shelter adjoining the collapsed ceiling
reached a maximum depth of 34 inches below the surface. Underneath the
debris-laden, dark-gray soil, was found a fine yellow gravel. No cultural
materials were found at any depth in this gravel but some objects were
definitely resting on its surface. The yellow gravel represents the floor of
the old stream bed at the time erosion was cutting away the soft sandstone
outcrop. Since there was no sterile layer of humus-laden soil between the
yellow gravel and the cultural deposits, the shelter was probably occupied
by people soon after the channel of the stream moved away from the stone
outcrop.

The debris in the blocks extending towards the northeastern end of the
shelter became progressively thinner in depth until in the last excavated
block it was at a maximum of 18 inches. In this area the yellow gravel was
thin and rested upon the sandstone bedrock which was closer to the surface
in this area because the formation dipped gradually downward to the
southwest.

All of the dark-gray soil of division A contained cultural debris. This
debris varied in kind and age from the surface to the bottom regardless
of the total depth. Tools of flint were common throughout this debris but
pottery was found only in the upper levels. In blocks with a deep deposit,
pottery was found to a depth of approximately eighteen inches below the
surface, but in the blocks with the thin deposits pottery was confined to
the upper six inches. From one-third to one-half of the lower depth of the
deposits was non-ceramic in content. The types of pottery and projectile
points found in the various levels and the cultural time periods to which
they belong will be discussed in later paragraphs dealing with these objects.
Suffice it to say here that definite cultural stratigraphy did exist. No clear-
cut soil differences could be correlated with the cultural changes. The debris
appears as a gradual build-up with no disturbances such as flood deposits
to interrupt the continuity.

All the features found in division A were poorly defined. The dark, dry,
loose soil does not cut along sharp lines but crumbles and breaks. Thus, a
pit dug into the soil did not maintain a sharp outline, and unless the yellow
gravel was reached there was little or no change in soil coloration. All of
the seven refuse pits recorded were small and shallow in depth. The largest
was 18 inches in diameter and the maximum depth was 12 inches. Refuse
deposited in these small pits consisted mainly of broken animal bones, flint
chips, and pottery sherds.

There were four areas in division A where fires of considerable intensity
had been located. None of these areas could be called fire-pits since the fires
had not been confined to a depression in the soil but had been placed
directly on the surface. Around these areas ranging in diameter from 18 to
30 inches was an accumulation of ash, small bits of charcoal, and reddened
earth. Small fires of short duration probably were represented by the scat-
tered ash and flakes of charcoal found throughout the debris in the shelter.
The remains of small fires would have easily become scattered in the
loose soil.

152

Annals of Carnegie Museum

vol. 35

Divisions B and C

The work done in division B will be considered in the overall picture
gained from the excavation in division C. Only one block adjoining division
C was dug in division B because most of the area was covered by large
boulders. Five whole blocks and a portion of a sixth were dug in division C.
All of the blocks were well under the protection of the overhang towards
the southwest end of the shelter. One large slab had collapsed from the
ceiling and covered a portion of the area. This fall appeared to be recent
since cultural materials and dark, debris-laden soil were found well under the
edge of the stone.

The dark, gray soil containing cultural debris extended to a depth of 30
to 34 inches below the surface in all the blocks of this area. Directly below
this debris was the yellow gravel as encountered in division A. There was
no sterile soil between the yellow gravel and the debris-laden soil. Some
objects were found directly on the gravel level but none was found below it.

The sequence of cultural materials was confused in a portion of this area
because of the presence of a large refuse pit approximately six feet in
diameter. This pit was made late in the occupation of the shelter as it began
near the surface and extended to a depth of nearly 28 inches. Its contents
consisted mainly of elk and deer bones mixed with pottery sherds repre-
sentative of various time levels. The margins of this pit, like those found
in division A, were indistinct and difficult to follow because of the loose soil.
There was the chance that some objects belonging to the pit would be in-
cluded in the undisturbed levels surrounding it.

All pottery was confined to the first 18 inches of the deposit except in
the disturbed area of the large pit where a few sherds were found to a depth
of 26 inches. Flint objects were found at all levels from the surface to the
yellow gravel.

The large pit was the only well defined feature found in divisions B and
C. Some thin lenses of ash were encountered but there was no evidence of
the rock-lined fire-pits common to open village sites.

POTTERY

Pottery was not abundant in the refuse of the Dixon Shelter but a suf-
ficient number of sherds was found to give a relatively clear picture of the
time sequence and types involved. From all the excavated blocks 681 sherds
were recovered. No whole or completely restorable vessels were found.

All sherds from the levels of each block were sorted and classified according
to “wares” based on the kind of temper. Using Carnegie Museum’s Upper
Ohio Valley Archeological Survey’s pottery nomenclature (Mayer-Oakes,
1955a), I identified four major “wares”— shell-tempered Monongahela, lime-
stone-tempered Watson, medium grit-tempered Mahoning, and coarse grit-
tempered Half-Moon. In the case of the Half-Moon ware, thickness and
surface treatment were important factors in separating it from the Mahoning
ware, also grit-tempered. Each of the ware groups was divided into its
component types with the following results:

Monongahela Cordmarked

Method of manufacture. Coiled; malleated with a cordwrapped paddle.

1958

Dr a goo: Archaic Hunters of the Upper Ohio Valley

153

Temper. Moderate to finely crushed particles of shell composing approxi-
mately 25% of the paste.

Texture. Finely laminated.

Hardness. 2.5 to 3.5.

Color. Exterior surfaces range from buff to black. Interiors usually are
gray to black.

Surface finish. Exterior surfaces covered by the impressions of a cord-
wrapped paddle. These impressions usually meet the lip at an oblique
angle. Some sherds show smoothing of the cord impressions. Interior
surfaces are smooth.

Decoration. Confined to the lip, small indentations of a cordwrapped
stick or paddle edge are common.

Form. Globular jars or bowls with flattened or rounded lips on a straight
or slightly outflaring rim. No appendages noted.

Thickness. 4 to 7 mm.

Diagnostic features. Paste, surface finish, rim form.

Geographic range .. This type is known from all major streams of the
Upper Ohio Valley. It occurs most abundantly on sites along the
Monongahela River and down the Ohio River proper where it overlaps
with similar types of Fort Ancient pottery.

Probable relationship. This pottery type is the most common and basic
shell-tempered type found in the Upper Ohio Valley. It occurs at villages
of the Late Prehistoric Period which in the Upper Ohio Valley appear
to be mixtures of Mississippi and Woodland cultures. This is known
as the Monongahela Complex and its closest relationship is with Fort
Ancient.

Size of sample. 102 sherds.

References. Butler (1939). Dragoo (1955). Mayer-Oakes (1955).

Watson Cordmarked

Method of manufacture. Coiled; malleated with a cordwrapped paddle.

Temper. Crushed fragments of limestone making up from 25 to 50%
of the paste.

Texture. Compact but often irregular because of the large size and the
angular shape of the limestone tempering.

Hardness. 2.5 to 3.5.

Color. Both surfaces usually a dull gray or tan.

Surface finish. Exterior surface impressed with the marks of a cord-
wrapped stick running at an oblique or right angle to the rim. The in-
teriors are generally smooth but faint cord impressions occur on a few
sherds.

Decoration. Confined to the lip or neck area were impressions of a cord-
wrapped stick or paddle edge applied at an oblique angle.

Form. Globular vessels with straight to slightly flaring rims.

Thickness. 4 to 8 mm.

Diagnostic features. Paste, surface finish, rim form.

Geographic range. This type is known at scattered sites throughout the
Upper Ohio Valley. It is best known from sites in the Ohio Valley

154

Annals of Carnegie Museum

vol. 35

proper where the type site, Watson, is located.

Probable relationship. This pottery type is the one most commonly asso-
ciated with Middle Woodland occupations in the Upper Ohio Valley.
At the type site it is associated with a stone mound burial complex
lying stratigraphically above an older Early Woodland occupation. It
is thus representative of the Upper Ohio Valley’s diluted Hopewellian
manifestations. Watson Cordmarked overlaps and shares many features
with the older Half-Moon type. It remained in use up until Late Pre-
historic times when it was replaced by the shell-tempered Monongahela
types.

Size of sample. 483 sherds.

References. Dragoo (1954, 1956). Mayer-Oakes (1955a).

Watson Plain

This type is similar to Watson Cordmarked in all known features ex-
cept for the smooth surface finish. It occurs as a minority companion
to the cordmarked type at all sites where it has been found. Its dis-
tribution and relationships parallel those of the cordmarked type.

Size of sample. 13 sherds.

References. Dragoo (1954, 1956). Mayer-Oakes (1955a).

Watson Incised

This type is not common. It is an incised decorated version of either
the Watson Cordmarked or Watson Plain types. The only two sherds
of this type at the Dixon Shelter had incised lines over smoothed cord-
marking.

References.. Dragoo (1954, 1956). Mayer-Oakes (1955a).

Mahoning Cordmarked

Method of manufacture. Coiled; malleated with a cordwrapped paddle.

Temper. Crushed fragments of rock, usually of igneous origin, and some-
times chert and quartz. Temper fragments are small to moderate in
size and make up 25 to 40% of the paste.

Texture. Compact and regular except for a few sherds containing large
temper fragments.

Hardness. 2.5 to 3.5.

Color. Ranges from dark gray to tan.

Surface finish. Exterior surface covered by the impressions of a cord-
wrapped paddle applied vertically to the rim. The interiors are smooth
except for a few sherds which have faint cord impressions.

Decoration. Confined to the lip and consisting of the impressions of a
cordwrapped paddle.

Form. Probably globular with a slightly flaring rim with square to
rounded lips.

Thickness. 4 to 7 mm.

Diagnostic features. Paste, temper size, and surface finish.

Geographic range. This type is known from sites throughout the Upper
Ohio Valley. It is concentrated on sites in the Beaver River and Ma-
honing River valleys but it occurs as a minority type on most sites
indicating extensive and long occupation.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

155

Probable relationship. This type covers a wide range of time and cul-
tural differences. It is known stratigraphically to be contemporary with
late Early Woodland types and it continued to be made until Late
Prehistoric times. As a time marker it is almost useless. It is the Upper
Ohio Valley’s contribution to the amorphous type often called “Wood-
land Cordmarked”.

Size of sample. 25 sherds.

References., Dragoo (1954, 1956). Mayer-Oakes (1955a).

Mahoning Plain

This type possesses, with the exception of the smooth surface, the
same general features, geographic range, and relationships as Mahoning
Cordmarked.

Size of sample. 15 sherds.

References. Dragoo (1956). Mayer-Oakes (1955a).

Half -Moon Cordmarked

Method of manufacture. Coiled; malleated with a cordwrapped pacldle.

Te?nper. Coarsely crushed fragments of igneous rock most common, but
chert, quartz, and limestone also used. Sometimes a combination of
tempering particles used. Temper fragments are large and make up
from 50 to 70% of the paste.

Texture. Rough and irregular.

Hardness. 2.5 to 3.5.

Color. From tan to light gray.

Surface finish. Exterior surface deeply marked by cord impressions. In-
teriors may be smooth or marked by cord impressions.

Decoration. No evidence.

Form. Large globular vessels indicated.

Thickness. 8 to 18 mm.

Diagnostic features. Paste, temper size, and thickness.

Geographic range. This type is known best from sites on the Ohio River
proper but some examples have been found at sites scattered throughout
the Upper Ohio Valley drainage area.

Probable relationships. This type is the earliest known pottery from the
Upper Ohio Valley. When found in stratified sites, it is always at the
lowest levels. It has been definitely associated with Adena burial mounds
and villages attributed to that culture. Half-Moon Cordmarked is closely
related to Baumer (Cole, 1951) and Crab Orchard (Maxwell, 1951) types
and the type Fayette Thick (Griffin, 1943) so common in the central
and lower Ohio Valley. It also shares basic similarities with Vinette I
of New York state (Ritchie and MacNeish, 1949).

Size of sample. 37 sherds.

References. Cole (1951). Dragoo (1954, 1956). Griffin (1943). Maxwell
(1951). Mayer-Oakes (1955a).

McKees Rocks Plain

This type closely resembles the paste features of the Half-Moon Cord-
marked type. The major difference lies in the smooth exterior and in-
terior surfaces. In the Upper Ohio Valley this type appears as the com-

156

Annals of Carnegie Museum

vol. 35

panion to the cordmarked variety in both time and cultural relation-
ships. It is most similar to “Adena Plain” (Griffin, 1942) of the central
Ohio Valley.

Size of sample. 4 sherds.

References. Griffin (1942). Mayer-Oakes (1955).

Scarem Plain

This type is represented by only two sherds at the Dixon Shelter.
Both sherds are untempered and are from small “toy” vessels. Sherds
of this type are common on Late Prehistoric sites of the Upper Ohio
Valley.

Size of sample. 2 sherds.

References. Dragoo (1953). Mayer-Oakes (1955a).

Pottery of division A. In the blocks of division A, 434 sherds were found.
The first level (0-6 inches) contained 267 sherds (61.6%), the second level
(6-12 inches) had 134 sherds (30.9%), the third level (12-18 inches) possessed
29 sherds (6.7%), and the fourth level (18-24 inches) accounted for only 4
sherds (0.8%) which were believed possibly to be in the bottom of an ill-
defined refuse pit.

There were 51 shell-tempered Monongahela Cordmarked sherds of which
84.3% (43 sherds) were found in the first level and the remaining 15.7%
(8 sherds) in the second level. All of the shell-tempered sherds recorded for
the second level were recovered from the blocks containing the deepest
debris. As the blocks of division A became thinner in depth toward the
northeast end of the shelter, the shell-tempered sherds were found only
on or just below the surface. (See Fig. 3, rows 1-2.)

The 347 Watson Cordmarked and Watson Plain sherds found in division
A were distributed 61.9% (215 sherds) in the first level, 32% (111 sherds)
in the second level, 4.9% (17 sherds) in the third level, and 1.2% (4 sherds)
in the fourth level. Thus, 38.1% of the limestone-tempered pottery was
found below the first level in contrast to only 15.7% for shell-tempered
sherds. (See Fig. 3, rows 3-5.)

Of the 22 Mahoning Cordmarked and Mahoning Plain sherds removed
from division A, there was 27.2% (6 sherds) in the first level, 36.4% (8
sherds) in the second level, and 36.4% (8 sherds) in the third level. There
was 72.8% of the sherds of this type evenly distributed in levels two and
three below the first level in comparison with 38.1% of limestone-tempered
and only 15.7% of shell-tempered sherds.

The meager samples of Half-Moon Cordmarked and McKees Rocks Plain
types were distributed 8.3% (1 sherd) to the first level, 58.3% (7 sherds) to
the second level, and 33.4% (4 sherds) to the third level. Four of the seven
sherds listed for the second level were from blocks where the debris was
thin. These sherds were at the bottom of the ceramic zone. (See Fig. 3,
rows 6-7.)

Pottery of divisions B and C. The contiguous blocks of divisions B and
C were considered as a single unit for pottery analysis since the nature and
depth of the debris appeared nearly identical in all blocks. The sequence of
deposition was interrupted only by the large refuse pit previously mentioned.
This large, vaguely defined pit made proper placement of a few sherds

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

157

Fig. 3. Dixon Shelter. Rows 1-2, Shell-tempered pottery. Rows 3-5, Lime-
stone-tempered pottery. Rows 6-7, Heavy grit-tempered pottery

158

Annals of Carnegie Museum

vol. 35

very difficult since it was almost impossible to draw the line between the
disturbed soil and the naturally accumulated soil. The same pottery types
found in division A were present in divisions B and C.

Of the 248 identifiable sherds found in divisions B and C, there were
175 sherds (70.6%) in the first level, 57 sherds (23.0%) in the second, 10
sherds (4.0%) in the third, and 6 sherds (2.4%) in the fourth level. All the
sherds tabulated for the fourth level were found in the immediate area of
the large refuse pit and may have been included in that disturbance.

There were 52 shell-tempered Monongahela Ware sherds of which 90.4%
(47 sherds) were in the first level and the remaining 9.6% (5 sherds) in the
second level.

Distribution of the 149 limestone-tempered Watson Ware sherds was
69.8% (104 sherds) in the first level, 24.8% (37 sherds) in the second level,
4.0% (6 sherds) in the third level, and 1.4% (2 sherds) in the fourth level.

There were only 18 grit-tempered Mahoning Ware sherds of which 33.4%
(6 sherds) were in the first level, 22.2% (4 sherds) in the second level, 22.2%
(4 sherds) in the third level, and 22.2% (4 sherds) in the fourth level.

The 29 coarse grit-tempered Half-Moon Ware sherds were confined to the
upper two levels. There was 62.0% (18 sherds) of the sherds in the first
level and 38.0% (11 sherds) in the second level. Nearly all of these sherds
were found in the area disturbed by the refuse pit and were, thus, out of
the natural deposition sequence.

POTTERY SUMMARY

Pottery types typical of all the major ceramic periods of the Upper Ohio
Valley were found in the Dixon Shelter. The pottery sample was small and
from a too shallow depth to be of the highest value for stratigraphic pur-
poses. Thus, a pottery sequence determined from the sherds at the Dixon
Shelter could be only suggestive; but when this material was placed in rela-
tion to sequences known from other sites in the Upper Ohio Valley, the
pattern was almost identical.

At all known stratified sites in the Upper Ohio Valley, the Monongahela
pottery types are found only in the top levels and are typical of the Late
Prehistoric period. In all areas dug at the Dixon Shelter, Monongahela
pottery was confined to the upper two levels with the highest percentage
of sherds in the first level near the surface. This pottery represented the
terminal Indian occupation.

In deep sites Watson Ware pottery occurs stratigraphically below the
Monongahela types and is representative of the Middle Woodland period.
At the Dixon Shelter, this pottery was the most abundant ware. It occurred
from the upper to the lowest pottery-bearing levels. On the basis of type
percentages per level, the Watson Ware was in sequence below the Monon-
gahela Ware. Its extensive vertical distribution would indicate a relatively
long period of its usage at the shelter.

Mahoning Ware pottery was distributed rather evenly from top to bot-
tom levels. A greater percentage of this type was found in the lower ceramic
levels than was the case with Watson Ware. The Mahoning Ware appears
as the surviving member of an early grit-tempered ceramic tradition. At

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

159

all known stratified sites in the Upper Ohio Valley, the Mahoning Ware
was used during the last phases of the early coarse grit-tempered tradition
and continued in use until replaced by shell-tempered Monongahela Ware.
It often occurs as a minority ware in association with the Watson Ware
(Dragoo, 1956).

Half-Moon Ware pottery is the earliest known from stratified sites in
the Upper Ohio Valley. It is the typical coarse grit-tempered pottery of the
Early Woodland period. The position of the Half-Moon pottery at the
Dixon Shelter was not clearly defined. In division A it was found in the
lower ceramic levels, but in divisions B and C the best examples of it came
from a disturbed area and in the upper levels. The few sherds found in
division A seem to indicate that Half-Moon Ware was also the earliest
pottery at the Dixon Shelter.

CHIPPED-STONE ARTIFACTS

The same problems encountered in the study of the pottery were present
in the tabulation of the chipped-stone materials. The varying depth of the
habitation debris made it impossible to correlate objects from all blocks on
a straight level basis. In order to approximate more closely the actual cul-
tural stratigraphy for study purposes, the blocks were divided into two
groups. All blocks containing four or more levels (most blocks had five
levels) were put in one group and all those containing only three levels
were placed in another. By doing this, it was possible to compare accurately
the sequence of chipped-stone materials for the entire excavation and their
relationship to the pottery-bearing levels.

Projectile points

The most abundant chipped-stone items recovered from the excavation
were projectile points. From all the blocks, 187 points were collected. Of
these, 180 could be sorted into generally recognized types. The seven re-
maining points were of unusual or undeterminable form.

Of the 134 points found in the blocks containing four or more levels, the
distribution was 27 (20.15%) in the first, 33 (24.62%) in the second, 34
(25.37%) in the third, 33 (24.62%) in the second, 34 (25.37%) in the third,
33 (24.62%) in the fourth, and 8 (5.97%) in the fifth level. In the blocks
containing only three levels there were 53 points distributed 12 (22.64%)
in the first, 23 (43.39%) in the second, and 18 (33.98%) in the third level.

Of the 180 classified points, there were 124 points (69.4%) found in
levels which also produced pottery and 55 points (30.6%) in levels where
pottery was absent. All of the 180 points were sorted into generally accepted
types (stemmed, side-notched, corner-notched, and triangular) and plotted
as to level distribution. In the discussion which follows, each type of point
and its subdivisions will be described and its position in relation to pottery
and non-pottery levels noted.

Stemmed points. These points constituted 56.11% (101 points) of the
classified sample of 180 points. During the sorting, two major variations of
the stem were encountered. The first variation consisted of points with a
parallel stem and a rounded or square base. The second and most numerous
variation was points with an expanded stem.

160

Annals of Carnegie Museum

vol. 35

The 36 parallel-stemmed points made up 35.64% of the total stemmed
point sample (101 points) and 20.0% of the total point sample. Of these
points, 77.78% (28 points) were found in levels where pottery also occurred,
while the remaining 22.22% (8 points) had no pottery association. The
stemmed points were the largest projectile points found in the excavation.
They range in length from 114 to 314 inches. In Fig. 4, no. 1-10, several of
the largest points of this type are shown. All of these points are of a
coarse-grained flint except for no. 10 which is of quartz. The chipping on
these points is crude, but the nature of the coarse material used in their
manufacture would have made fine chipping impossible. In spite of the
crude appearance of these points, they were, with the exception of two,
found distributed throughout levels containing pottery. The unique shape
and the materials from which they were made set them apart from all
others and produce speculation for their origin outside of the Ohio Valley.
Points of similar type are most common on many sites in the Potomac River
drainage. The only points comparable to these in the Ohio Valley occur
on the eastern fringe of the area in Garrett County, Md. (Corliss, 1954)
where they are commonly associated in an Archaic context more closely
related to materials found along the eastern seaboard.

A few points in the parallel-stemmed category have rounded bases. These
are shown in Fig. 4, no. 12-16. Only the bases of this type of point were
found. All are of fine-grained flint and nicely chipped. They are remi-
niscent of points found on Early Woodland sites, particularly Adena, in
the Upper Ohio Valley. Stemmed points were most common in the lower
pottery levels where they were associated with grit- and limestone-tempered
pottery.

The picture presented by the total distribution of the parallel-stemmed
points was one of long usage for this general type. It would appear that
from time to time distinctive modifications, as mentioned above, were made
by certain groups.

The most numerous points were those classified in the expanded-stem
category. They made up 36.11% (65 points) of the total sample of 180
points. Of these 65 points, 43 (66.15%) were found in levels containing
pottery and 22 (33.84%) were located in non-pottery levels.

These points varied in outline but each possessed a stem with a slightly
flaring base. The variation in size was considerable, but none was as large
as some in the parallel-stemmed group. Some points of this general type
could have been thrown into one of the notched-point categories but they
would have represented poor examples of such types. It may be said that
these points were characterized by their vagueness of form. The materials
used in their manufacture showed great diversity.

Expanded-stem points have been found at many sites in the Upper Ohio
Valley. They are known from shell-midden sites along the Ohio River
where they are associated with lanceolate and broad-stemmed types. They
are most abundant on sites of Archaic times (Stewart and Dragoo, 1954)
where they are the dominant type, but they are known to have persisted
in use up through Middle Woodland occupations (Dragoo, 1954).

Side-notched points. Points with side notches were second in abundance

1958 Dragoo: Archaic Hunters of the Upper Ohio Valley 161

%

* 0

# # # 1 1 ? - ’■ • «
^ (i § * I s i i #

#S#wt ‘ ‘

* ' , * , l :

I#44«lf 1 1 •

itiiliiiliu

** % !•« *# * k A i

, # ■ *

wj w

Fig. 4. Dixon Shelter. Various types of points found during excavation

162

Annals of Carnegie Museum

vol. 35

to those with expanded stems. The 58 points with side notches accounted for
32.21% of the sample of 180 points. During the sorting of this group it be-
came apparent that there were two main categories in which the points
could be placed. The first category consisted of points with varying out-
lines and notches distributed throughout all levels in the excavation. The
second category consisted of small, thin, slender points with deep, well
formed notches and a limited level distribution.

There were 39 points of a general side-notched type. See Fig. 4, no. 77-120
for examples. Of these points, 53.84% (21 points) were found in levels
containing pottery while 46.15% (18 points) were found in levels below the
pottery. Points of this type were found from the lowest to the top level
but their greatest concentration occurred in levels three and four in blocks
containing four to five levels and in levels two and three in blocks contain-
ing only three levels. Many of these points are small and closely parallel
the distribution of the expancled-stem type. Both of these types of point
appear to belong to the same cultural tradition. It was often difficult to
separate clearly the two types both in form and kinds of materials used
in their manufacture. Both types are typical of sites now classified as of
Archaic times in the Upper Ohio Valley.

The second category of side-notched points contained 19 points with
78.94% (15 points) in pottery levels and 21.05% (4 points) in non-pottery
levels. Their heaviest concentration occurred in levels two and three in
blocks containing four or five levels and in level two in blocks having only
three levels. The closest pottery association was with limestone-tempered
ware. This type of point does not appear to be associated with any estab-
lished Archaic manifestation in the Upper Ohio Valley. Its origin must be
sought among later pottery-producing cultures.

Corner-notched points. Points of this type do not compose a well de-
fined category at the Dixon Shelter. Only seven points, or 3.88% of the
total point sample, were tentatively assigned to the corner-notched type. Of
these points 57.14% (four points) were found in levels containing pottery
while 42.85% (three points) were below the pottery levels. These points
stand apart from the expanded-stemmed type only by their deeper and
more finely chipped notches, their greater width in relation to length, and
their generally thinner cross-section.

Triangular points. Small triangular points represented 7.77% (14 points)
of the total sample. All of these points were found in levels containing
pottery with 71.4% (10 points) recovered from the first level, 21.4% (three
points) in the second level, and the remaining 7.2% (one point) in the
third level. All of the points found in the first level were close to the
surface and closely associated with the shell-tempered pottery. The small
number of points found below the first level probably reached the lower
depth in areas of disturbance. All of the triangular points are of types
commonly associated with Late Prehistoric cultures.

Drills

Only 15 drills were found. They were distributed 6.7% (1 drill) in the
first level, 33.3% (5 drills) in the second level, 26.7% (4 drills) in the third
level, 20% in the fourth level, and 13.3% (2 drills) in the fifth level. Thus,

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

163

the second level contains the largest number of drills followed in diminish-
ing order by the third, fourth, and fifth levels. In relationship to other
materials, especially pottery, drills were more closely associated with lime-
stone-tempered and grit-tempered pottery and with the points below the
pottery levels than with shell-tempered pottery.

None of the drills is large, the length ranging from 3.8 centimeters to
5.2 centimeters. Nine of the 15 drills have slightly expanded bases, five
are straight, and the remaining one has a contracting stem. Those with
slighly expanded bases were more common in the second and third levels
while those with straight bases occurred more frequently in the fourth and
fifth levels. The materials used in the manufacture of the drills varied, but
fine grained cherts were most common. (See Fig. 5.)

Scrapers

Next to projectile points, scrapers were the second most abundant tool
found in the excavation. They occurred in all levels and are of various
types. Almost every flint chip found showed evidence of having been used
as a scraper. For the purpose of classification only well worked pieces with
obvious secondary chipping and evidence of repeated use were selected.
The well defined scrapers were divided into three types— side-scrapers, end-
scrapers, and gravers. Their characteristics and distribution were as follows:

Side-scrapers. Scrapers possessing one or more straight cutting edges were
common in all levels. They were usually made from irregular flakes of
medium to fine grained cherts. A total of 60 specimens was assigned to this
category. According to levels these 60 specimens were distributed 16.66% (10
scrapers) in the first level, 26.66% (16 scrapers) in the second level, 26.66%
(16 scrapers) in the third level, 21.66% (13 scrapers) in the fourth level,
and 8.33% (5 scrapers) in the fifth level. No basic distinction could be made
for side scrapers occurring in pottery or non-pottery levels.

End-scrapers. These scrapers are of the type commonly called “thumb-
nail” scrapers. They are generally small and nicely chipped with a rounded
cutting edge. Fine-grained cherts were most often used in their manufac-
ture. End-scrapers were not too common and they had a more restricted
distribution than the side-scrapers. The seven scrapers of this type were
found 42.85% (3 scrapers) in the third level, 42.85% (3 scrapers) in the
fourth level, and 14.28% (1 scraper) in the fifth level. End-scrapers were
found in only the lowest pottery level and in the non-pottery levels. They
would appear to be of definite Archaic context at this site. (See Fig. 5.)

Gravers. Only three objects were assigned to this category. They were
small with triangular outline and cross-section. They all had finely chipped,
sharp points and one or two sides worked as scraper edges. They were
made of fine-grained chert. Two of these tools were found in the second
level and one in the fourth level.

BONE ARTIFACTS

Animal bones of a refuse nature were found scattered throughout the
various levels of the site. John Guilday of the Museum staff has identified
these bones in the following report.

164

Annals of Carnegie Museum

vol. 35

Fig. 5. Dixon Shelter. Points, drills, scrapers, bone tools, and copper beads

found during excavation

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

165

Report on the bone refuse from the Dixon Rock Shelter (46 Pr 6)

The bone refuse, 2467 fragments in all, consisted of: unidentified mammal
fragments, 2116; identified mammal fragments, 332; bird 8; turtle 4; and
one fragment of human bone.

Ursus americanus Black bear: 4 bone fragments.

Procyon lotor Raccoon: 3 bone fragments.

Marmota monax Woodchuck: 2 bone fragments.

Tamias striatus Chipmunk: 1 bone fragment.

Sciurus carolinensis Gray squirrel: 3 bone fragments.

Castor canadensis Beaver: 1 tooth.

- Erethizon dorsatum Porcupine: 3 bone fragments.

Sylvilagus sp. Rabbit: 1 bone fragment.

Ceruus canadensis Elk: 30 bone fragments.

Odocoileus virginianus White-tailed deer: 283 bone fragments.
Unidentified mammal bone: 2116 fragments.

Colinus virginianus Quail: 1 bone fragment.
Meleagris gallopavo Wild turkey: 4 bone fragments.
Bird sp.: 3 bone fragments.

Turtle sp.: 4 bone fragments.

Man: One phalange, infant.

One fragment of turtle carapace showed signs of “sanding” on both
external and internal surfaces. Possibly a cup fragment.

Only a few tools made from bone were found in the excavation. None
could be called distinctive for any particular time period. Most common of
the bone tools were antler tip points, antler drifts or flakers, splinter bone
awls, and beads. Nearly all of these objects were broken, and some had been
partially burned. Two of the antler tips had been intentionally fire hard-
ened. All of the awls were made from crude splinters of bone with only
the tips carefully worked. The antler drifts showed more careful attention.
In Fig. 5, no. 111-122, the best of these items are illustrated.

One deer phalange had been modified into an artifact. The end had
been pierced and grooved. It was probably used as a tinkler. A fragment
of turtle carapace showed signs of sanding on both external and internal
surfaces. It was possibly a cup fragment.

The only unusual item of bone was a distal joint of a deer canon bone
1% inches in length which had been drilled for suspension. (See Fig. 5,
no. 110.)

The preservation of bone at the Dixon Shelter was not good. Bones found
in the lower levels were very fragmentary and bones in the upper level
were poorly preserved. The sample of worked objects does little to clarify
the cultural picture of the Dixon Shelter.

166

Annals of Carnegie Museum

vol. 35

METAL ARTIFACTS

With the exception of a historic pocket-knife and an iron ring found near
the surface only two objects of metal were found in the excavation. Both of
these objects are rolled copper beads from the third level in division A.
These beads were found a few inches apart and in association with grit-
tempered and limestone-tempered pottery.

DISCUSSION

The Dixon Shelter was occasionally occupied from Archaic times until
Late Prehistoric times. All major cultural time periods with the exception of
Paleo-Indian were represented by materials found in the shelter. Approxi-
mately one-third of the debris deposit of the shelter could be assigned to the
non-ceramic Archaic period while the upper two-thirds of the deposit con-
tained ceramic materials of the three major ceramic culture periods— Early
Woodland, Middle Woodland and Late Prehistoric.

The Dixon Shelter was a convenient station for small hunting parties. It
afforded shelter from rain and strong winds and was adjacent to fresh
water. Completely surrounding the shelter was an area containing a good
population of game.

In Archaic times, there were no large groups of people in the Upper
Ohio Valley. There were probably numerous small family groups inhabit-
ing, where available, small rock shelters. In later ceramic times, after the
introduction of agriculture, larger populations established permanent vil-
lages and rock shelters became stations for hunting parties and families on
collecting trips.

The materials from the Dixon Shelter present an interesting study of the
transition from Archaic into the later ceramic periods. The tool assemblage
of Archaic levels was characterized by the use of small notched and stemmed
points, side- and end-scrapers, and crude knives. The use of small notched
points began at the lowest level and continued in use up into the Early
Woodland and Middle Woodland periods. Large stemmed points were
limited to the upper levels of the Archaic deposit and were found occasion-
ally in the later culture periods.

With the advent of heavy, grit-tempered pottery came the introduction
of the distinctive parallel-stemmed point with rounded base. This type of
point was commonly associated with the Adena culture in the central Ohio
Valley and it appears to be a good time marker in the Upper Ohio Valley.
At this time level there appears a preference for finer materials for the manu-
facture of points. Some of these materials were imported from areas out-
side the Upper Ohio Valley, particularly from the Flint Ridge deposits
in Ohio.

The major change that appears during the Middle Woodland period is
the use of limestone-tempered pottery. There are no point types different
from those prior to that time. The medium-thick grit-tempered pottery also
remained in use.

The most abrupt change occurs with the advent of the Late Prehistoric
period. Major changes occurred in types of pottery and projectile points.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

167

The pottery became predominantly shell tempered and projectile points
were nearly always of the small triangular type.

Influences from the Potomac River area were at a minimum at the Dixon
Shelter. Only a few specimens were of materials generally associated with
coastal complexes, and some of this material is known from the local area.
The materials used in the manufacture of tools at the Dixon Shelter con-
form to those commonly found at sites in the Upper Ohio Valley. Most of
these materials are of local origin occurring either in pebbles or thin vein
deposits. A wide variety of materials was used, some of which were poorly
adapted to the Indian’s needs. The most distinctive import was small quan-
tities of Flint Ridge chert. From the standpoint of materials, cultural
movements spread up the Ohio Valley and into the East.

The distribution of points in the Dixon Shelter is indicated in Tables 1-3.

TABLE 1. DIXON SHELTER, DIVISION A. DISTRIBUTION OF POINTS

IN BLOCKS

HAVING

FIVE LEVELS

Level

T

CN ST

EXP-ST

SNT

SN

M

Total

1

5

3

2

1

11

2

2

1 1

3

4

5

1

17

3

4

4

5

1

14

4

3

1

2

3

9

5

2

1

3

Total

7

1 13

8

9

9

7

53

T = T riangular; CN — Corner-notched; ST = Stemmed; EXP-ST = Expanded-stem;
SNT = Side-notched thin; SN = Side-notched; M = Miscellaneous.

TABLE 2. DIXON SHELTER, DIVISION A. DISTRIBUTION OF POINTS IN

BLOCKS ]

HAVING THREE LEVELS

Level

T

CN ST

EXP-ST SNT

SN

M

Total

1

4

7

1

12

2

7

8 3

4

1

23

3

2 3

7

6

18

Total

0

2 14

22 3

11

1

53

T = Triangular;

CN = Corner-notched; ST = Stemmed;

EXP-ST :

= Expanded-stem;

SNT = Side-notched thin; SN = Side-notched; M = Miscellaneous.

TABLE 3. DIXON SHELTER, DIVISIONS B AND C. DISTRIBUTION OF

POINTS BY LEVELS

Level

T

CN

ST

EXP-ST

SNT

SN

Total

1

5

2

5

2

1

1

16

2

1

1

4

8

1

1

16

3

1

11

1

7

20

4

1

10

4

9

24

5

4

1

5

Total

7

4

9

35

7

19

81

T = Triangular; . CN = Corner- notched; ST — Stemmed; EXP-ST — Expanded-stem;
SNT = Side-notched thin; SN = Side-notched.

168

Annals of Carnegie Museum

vol. 35

THE ROHR ROCK SHELTER

During our excavation at the Dixon Shelter in August 1954, Albert Bauer
took me to a fine rock shelter close to the Cheat River near Rohr in Monon-
galia County, W. Va. This site was named the Rohr Shelter and given the
number 46 Mg 9 in the files of Carnegie Museum’s Upper Ohio Valley
Archeological Survey. By examining the surface and digging a small test
pit, we determined that this site contained cultural materials similar to
those at Dixon. The test also indicated greater depth to the deposit and a
heavier concentration of Archaic materials. This site presented an excellent
opportunity to cross check our findings at Dixon and to extend our knowl-
edge of the Archaic in time.

With the kind co-operation and permission of Edward Howard, the owner
of the site, plans were made for the excavation of the shelter in 1955. On
May 23, 1955, work was begun with the aid of Mr. and Mrs. Albert Bauer.
On May 28, John Leppla and Mrs. Dorothy Dragoo joined the crew. Major
excavation work was completed on June 6, but smaller areas have been
excavated since that time by Bauer and Leppla who have given much of
their time and energy to the project and without whose aid the work would
not have been completed.

NATURAL SETTING

The Rohr Shelter is located in low mountainous country at an elevation
of 2025 feet above sea-level. The surrounding hilltops reach an elevation
between 2100 and 2300 feet. The shelter is in a narrow valley through which
a small stream flows into the Cheat River about one mile to the east. This
stream now flows nearly fifty feet below the floor level of the shelter.

The overhanging ledge or roof of the shelter is composed of Pottsville
conglomerate, the basal formation of the Pennsylvanian period in the area.
It is a compact, grayish white sandstone with a dense inclusion of smooth
quartz pebbles. Underlying the conglomerate is a fine, brown sandstone
which has been eroded away to form the shelter. All of the major shelters I
have seen in the Cheat River area have been composed of these same
formations.

The vegetation is very much like that at the Dixon Shelter, and the animal
life in Prehistoric times was probably quite similar.

EXCAVATION

The overhanging roof of the Rohr Shelter extends 92 feet along the out-
crop in a line from northwest to southeast. From the front of the shelter to
the rear wall, the recess varies in depth from a few feet to as much as 45
feet. At the front of the shelter the ceiling ranges in height from three to six
feet above the present floor level. The ceiling slopes gently downward toward
the rear of the shelter in an area varying in depth from 10 to 15 feet; and
then lowers abruptly to a height of three feet which then lowers gently to
the rear of the shelter. Near the northwestern margin of the shelter is a
spring and an area of intermittent water flow from under the outcrop. No
cultural deposits were found in this area. Only a few stones had fallen from
the ceiling. All appear to be old falls as they were found to be resting on

1958

Bragoo: Archaic Hunters of the Upper Ohio Valley

169

the sterile floor and surrounded by the cultural bearing layers. On the
ground surface directly below the outer margin of the shelter was a ridge
of materials deposited by water action from the hill slope above the shelter.
Directly in front of the shelter was a dense growth of trees and mountain
laurel.

The area of the shelter was surveyed and a base-line was established
parallel with the front of the ceiling. From this base-line, a grid of blocks
10 feet square was marked for the entire habitable area of the shelter. A
datum point was established for the taking of elevations and the drawing
of profiles.

Each block in the grid was excavated in six-inch levels beginning at the
surface and numbered in sequence from top to bottom. All the materials
from each level of each block were bagged separately and catalogued. Any
materials from a feature within a block were catalogued separately from
materials at the general level.

The excavating was done by customary archeological methods. Trowels
were used for nearly all primary digging. Most objects were found during
the removal of the soil by trowel, but because of the general darkness of
the shelter it was necessary to screen all the soil to prevent loss of small items.

Observations of soil differences, features, and cultural materials were
made for each level within a block. Drawings and photographs were made
of all important features.

Initial excavation was concentrated in five blocks 10 feet square, running
in a line parallel with the front of the shelter and including the most
habitable area. The cultural debris in this area, 50 feet in length, varied
in depth from 4*4 feet near the center of the shelter in blocks two, three,
and four to a minimum depth of two feet in block one and three feet in
block five. (See Fig. 6.)

The debris layer was composed of three distinct layers of soil. The first
and topmost layer extended from the surface to a depth of approximately
six inches. The soil of this layer was gray. Most of the Late Prehistoric ma-
terials were found within the confines of this layer. Immediately below the
layer of gray soil was a layer of dark soil with scattered inclusions of yellow
sand. This layer varied in depth from 6 inches near the ends of the shelter
to 18 Inches at the center of the shelter. Throughout this layer pottery
sherds of the Early Woodland and Middle Woodland periods were found.
Below the layer of dark soil was a layer of very black soil. Scattered through-
out this layer were remains of many fires which had in many places turned
the soil a brick red. This layer varied in depth from one foot near the ends
of the shelter to 2]/> feet at the center. This very black soil was devoid of
pottery sherds. On the basis of the stone materials recovered, this layer
represents the major Archaic occupation of the shelter. (See Fig. 7.)

Below the very black soil was a yellow sand which appeared free from
archeological materials except for a few objects which occurred on or just
below its surface. Several tests dug into this yellow sand produced no evi-
dence of cultural materials or disturbance by man. However, in one small
test pit in block two a point and several flint chips were found in an area

170

Annals of Carnegie Museum

vol. 35

of reddish sand nearly six inches below the surface of the yellow sand. The
full extent of this reddish stained sand is yet to be determined. The evi-
dence is inconclusive as to whether the red sand, with the artifacts found in
it, was intrusive into the yellow sand or represents a very old occupation of
limited extent. It is hoped that future work will solve this problem.

Fig. 6. Rohr Shelter. General view of excavations with Albert Bauer re-
cording the features

There were no major features in the uppermost layer composed of gray
soil. A few thin areas containing darkened soil and small pieces of char-
coal were noted. No definite fire-pits were found as the fires had not been
confined to a depression in the soil but had been placed directly on the
surface. Around these areas of fire were a thin accumulation of ash, small
bits of charcoal, and slightly reddened earth. Only fires of short duration
were represented by these areas in the gray-soil layer. This situation would
lend support to the idea that rock shelters received only transitory use
during Late Prehistoric times.

A few accumulations of debris were found in the gray soil but no well
defined pits were discovered. The evidence indicates that refuse was de-
posited on the surface without benefit of pit disposal.

Beginning at approximately six inches below the surface the layer of
dark soil presented features of a well defined nature. Nine stone lined fire-
pits were found within the layer of dark soil from 6 to 18 inches below the
surface. All of these fire areas were similar in construction. Three to four
flat stones had been placed in a slight depression in the soil and then a

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

171

ring of stones placed around them. These fire-pits presented evidence of
long and intensive use as there were considerable concentrations of charcoal
and ash. The stones were cracked and discolored and the soil around the
stones was burned a brick red. None of the f. re-pits contained any quantities
of cultural materials. An occasional flint chip or point was found within
them. Several small areas, discolored by fire but not outlined by stones,
were also found in the dark layer. As with similar features found in the
gray soil above, these areas were thin and poorly defined.

Well defined refuse pits were also absent from the dark soil layer. There
were, however, concentrations of refuse thinly spread over small areas.

The very black soil immediately below the dark soil and above the yellow
sand contained a number of stone-lined fire-pits similar to those found in
the layer of dark soil. Fires had also been kindled among the few large
boulders which protruded several inches above the surface of the yellow
sand. The entire black soil layer, varying in thickness from 1 to 214 feet,
contained very extensive evidence of many fires. The black coloration of the
soil was caused by large accumulaions of charcoal and ash. Almost every
small stone found within the layer was discolored and cracked by heat. The
soil in several areas was burned a brick red.

Fig. 7. Rohr Shelter. Hearth at 18-inch level

The layer of black soil presented a picture of long occupation. Since the
cultural materials found in the black soil were of Archaic times, it is ap-
parent that the shelter provided an ideal home for small hunting and col-

172

Annals of Carnegie Museum

vol. 35

lecting bands of people. The shelter would have been at its maximum utility
during Archaic times as the ceiling would have been from two to nearly
four feet higher from the floor than it is at present; as the refuse and soil
accumulated, the height of the shelter gradually decreased. As well as the
diminishing desirability, in later times the changed economy and village
life of peoples living in the region made the shelter of only occasional use.
(See Fig. 8.)

POTTERY

Pottery was not abundant in the refuse of the Rohr Shelter but a suf-
ficient number of sherds was found to give a general picture of the sequence
and the types involved. All information is based upon the study of sherds
since no whole or completely restorable vessels were found.

All sherds from the levels of each block were sorted and classified ac-
cording to “wares” based on the kind of temper. Four major “wares”, as
defined in Carnegie Museum’s Upper Ohio Valley Archeological Survey’s
nomenclature, were represented— shell-tempered Monongahela, limestone-
tempered Watson, medium grit-tempered Mahoning, and coarse grit-tempered
Half-Moon. These are the same wares found at the Dixon Shelter and their
full cultural relationships were discussed in the previous section. Each of the
“ware” groups found at Rohr Shelter was sorted into its component types
with the following results.

Monongahela Cordmarked

This type is manufactured by coiling and the use of a cordwrapped
paddle. The temper is moderately to finely crushed particles of shell com-
posing approximately 25% of the paste. The texture is finely laminated and
the hardness ranges from 2.5 to 3.5. The color on the exterior surface ranges
from buff to black while interiors usually are gray to black. The exterior
is covered by the impressions of a cordwrapped paddle. These impressions
usually meet the lip at an oblique angle. On some sherds the cord impres-
sions have been smoothed. The interior surface is smooth. Decoration is
rare; when present, it is generally confined to the lip, and consists of small
indentations of a cordwrapped stick or the edge of a paddle. The most
common forms of this type as known from other sites are globular jars
and bowls with flattened or rounded lips on a straight or slightly flaring
rim. Thickness ranges from 3.5 to 7 mm. Only 65 sherds of this type were
found during the excavation of the Rohr Shelter.

Monongahela Plain

This pottery is identical with Monongahela Cordmarked with the excep-
tion of having a smooth, plain surface. Only 10 sherds of this type were
found in the Rohr sample.

Watson Cordmarked

Pottery of this type was manufactured by coiling and malleating with a
cordwrapped paddle. Crushed fragments of limestone making up from 25
to 50% of the paste were used for temper. The texture is compact but often
irregular because of the large size and angular shape of the limestone temper.
The hardness varies from 2.5 to 3.5. The color of both interior and exterior

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

173

surfaces is usually a dull gray or tan. The exterior surface is impressed with
the marks of a cordwrapped paddle. The interiors are generally smooth but
occasionally faint cord impressions occur. Decoration is rare and confined
to the lip or neck. The common form for this vessel is globular with slightly
flaring rim. Thickness ranges from four to eight mm. The Rohr sample in-
cludes a total of 130 sherds of this type.

Watson Plain

This type is similar in all features to Watson Cordmarked except for the
smooth surface finish. Eight sherds are in the sample from Rohr.

Mahoning Cordmarked

Pottery of this type was manufactured by coiling and the use of the
cordwrapped paddle. The temper is crushed fragments of igneous rock,
chert, or quartz of moderate to small size which make up 25 to 40% of the
paste. The texture is compact and regular and the hardness ranges from
2.5 to 3.5. The color ranges from dark gray to tan. Surface finish consists of

174

Annals of Carnegie Museum

vol. 35

cord impressions on the exterior while the interior surface is usually plain.
Decoration is confined to the lip and consists of the impressions of a cord-
wrapped paddle. The form was probably globular with a slightly flaring rim.
The thickness ranges from 4 to 7 mm. Only seven sherds of this type are
in the Rohr sample.

Mahoning Plain

Mahoning Plain shares all the general features of Mahoning Cordmarked
except for the smooth surface. The geographic and cultural relationships of
the two types appear identical. There were 1 1 sherds of Mahoning Plain
in the Rohr sample.

Half-Moon Cordmarked

Pottery of this type was manufactured by coiling and malleating with a
cordwrapped paddle. The temper is coarsely crushed fragments of igneous
rock, chert or quartz. A combination of these tempering materials also
occurs. Temper fragments are large and make up from 50 to 70% of the
paste. The texture is rough and irregular and the hardness ranges from
2.5 to 3.5. The color varies from tan to light gray. The exterior surface is
deeply marked by cord impressions. The interior surface is usually marked
by cord impressions but plain interiors also occur. No decoration occurred
on this type at Rohr. The form was apparently large globular vessels with
some vessels having flat bases. The thickness ranges from 8 to 16 mm.

POTTERY DISTRIBUTION

A total of 239 pottery sherds of all types was removed from the levels of
blocks A1 to A5 which constituted an area of 500 square feet and constituted
the main habitable area of the shelter. Of the 239 sherds from this area,
192 or 80.3% were found in the first six-inch level, 37 sherds or 15.4% in
the second level, 9 sherds or 3.8% in the third level, and 1 sherd or 0.4% in
the fourth level. Thus, 95.8% of all pottery found was concentrated within
the first two levels. The few sherds found in level three and the single sherd
from the fourth level may well represent sherds displaced from the upper
levels through disturbances.

The first-level pottery sample of 192 sherds contained examples of all
the major “wares” known from the site. Monongahela ware made up 33.33%
(64 sherds) of the sample. Watson ware was most prominent with 57.29%
(109 sherds). Mahoning ware composed 6.77% (13 sherds) of the sample.
Sparsely represented was 1.04% (2 sherds) of Half-Moon ware. The remain-
ing 2.08% (4 sherds) were unidentifiable because of their smallness and
fragile condition.

Of the 37 sherds recovered from the second level 29.72% (1 1 sherds) were
Monongahela ware, 62.16% (23 sherds) Watson ware, and 8.11% (3 sherds)
Mahoning ware. No Half-Moon ware was found in this level.

Only nine sherds were found in the third level. Watson ware accounted
ior 55.5% (5 sherds) of the sample. The remaining sherds were divided
evenly 22.2% (2 sherds) Mahoning ware and 22.2% (2 sherds) Half-Moon
ware. No shell-tempered pottery was found in this level.

One sherd of Watson ware was found in the fourth level of block A2.
This was the only sherd found below 18 inches in the entire excavation.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

175

All sherds in A1 and A5 were within the first 12 inches and those in A3
and A4 were contained within the first 18 inches of the deposit.

The shell-tempered pottery ware consisted of two major types— Monon-
gahela Cordmarked and Monongahela Plain. Of the total 75 shell-tempered
sherds 86.6% (65 sherds) were cordmarked and the remaining 13.3% (10
sherds) had a plain surface. The Monongahela Cordmarked type was dis-
tributed 84.6% (55 sherds) in the first level and 15.3% (10 sherds) in the
second level. Of the Monongahela Plain sherds 90% (9 sherds) were in the
first level and 10% (1 sherd) in the second level. Of the entire sample of
shell-tempered pottery, 85.3% (67 sherds) were found in the first level and
14.6% (11 sherds) in the second. From observations made during the excava-
tion and the sherd tabulations it was apparent that shell-tempered pottery
was within the last phase of pottery making at the shelter. Many of the
sherds recorded for the first level were found on or near the surface, and
those found in the second level were from blocks containing the greatest
depth of debris.

The limestone-tempered ware was of two types, Watson Cordmarked and
Watson Plain. Of the 138 limestone-tempered sherds 94.2% (130 sherds)
were cordmarked and 5.8% (8 sherds) were plain. Limestone-tempered pot-
tery was found from the first to the fourth level and constituted the most
abundant and longest pottery sequence found in the excavation. The 138
sherds of this sample were distributed 78.9% (109 sherds) in the first level,
16.6% (23 sherds) in the second level, 3.6% (5 sherds) in the third level,
and 0.7% (1 sherd) in the fourth level. Watson Cordmarked and Watson
Plain occurred together in all levels except the fourth which contained the
single cordmarked sherd. As with the shell-tempered pottery, the limestone-
tempered pottery shows its greatest concentration in the first level, but it
begins two levels below the first appearance of shell-tempered pottery and
is representative of Hopewellian and other so-called Middle Woodland mani-
festations. In deeply stratified sites it is always below the shell-tempered
types so common to Late Prehistoric times. Because of the greatest abundance
of pottery in the first level at the Rohr Shelter, it is difficult to separate
clearly limestone-tempered and shell-tempered pottery in time. The super-
ficial location of many shell-tempered sherds with triangular points would
suggest that the two wares were not contemporaneous and may not have
influenced each other in development.

The medium grit-tempered pottery was of two types— Mahoning Cord-
marked and Mahoning Plain. Of the sample of 18 sherds 38.8% (7 sherds)
were cordmarked and 61.1% (11 sherds) were plain. All of the cordmarked
sherds were found in the first level, but the plain sherds were distributed
54.5% (6 sherds) in the first level, 27.2% (3 sherds) in the second level, and
18.2% (2 sherds) in the third level. The distribution of the total medium
grit pottery was 72.2% (13 sherds) in the first level, 16.6% (3 sherds) in the
second level, and 11.1% (2 sherds) in the third level. The medium grit-
tempered pottery appears to have been an outgrowth of the early heavy
grit-tempered pottery tradition in the Upper Ohio Valley. It is found as a
minority ware on most sites which produce limestone-tempered pottery which
also appears to be related to the early heavy grit-tempered tradition.

176

Annals of Carnegie Museum

vol. 35

The heavy grit-tempered ware known as Half-Moon Cordmarked in the
Upper Ohio Valley is represented by only four sherds taken from the excava-
tion in the Rohr Shelter. Two of these sherds were found in the first level
and the other two in the third. Since sherds of this type are found only at
the lowest level in deeply stratified sites, the finding of two such sherds in
the first level at the Rohr Shelter could possibly indicate a disturbance not
noted during the excavation.

The general picture presented by the distribution of pottery at Rohr
was one of compactness. Although types representative of all major ceramic
periods were present, the greatest number of sherds for each type was
found in the first level. Below the first level the number of sherds dimin-
ished rapidly. The time span represented by all the pottery types found
within the first three levels would account for nearly 2500 years during
three major cultural periods. Below the pottery-bearing levels were three
to five levels, depending upon area, of cultural debris containing artifacts
from much earlier times.

CHIPPED-STONE ARTIFACTS
Projectile points

The most abundant chipped-stone artifacts recovered from the excava-
tion were projectile points, with a total of 288 points collected from all
blocks. Projectile points were found in all six of the debris-producing levels
of the excavation. Their distribution according to levels was 78 (27.08%) in
the first, 54 (18.75%) in the second, 75 (26.04%) in the third, 60 (20.83%)
in the fourth, 14 (4.86%) in the fifth, and 7 (2.43%) in the sixth.

All of the points were sorted into generally accepted types and plotted
as to distribution by levels. It was difficult to differentiate clearly many of
the small expanded-stem and side-notched points because of their crude-
ness and varied form. Many of these points possessed characteristics of both
types. As will be seen later in the discussion of these points and their dis-
tribution, the attempt to divide them into two types may be more artificial
than real. Distribution is indicated on page 186.

In the discussion which follows, each type of point and its variations will
be described, and its relationship to the pottery distribution noted. (See
Fig. 9-14.)

Triangular points. The 27 small triangular points represented 9.37%
of the total sample. All of these points were confined to the first two
levels, with the first level containing 26 points (96.29%) and the second
level having only 1 point (3.70%). In the first level there were 78 points of
various types. Triangular points were the dominant type, constituting
33.33% of the 78 points. Almost all of the triangular points were found
near the surface and closely associated with the shell-tempered pottery. The
one point found in the second level probably reached the lower depth
through disturbance of the soil. All of the triangular points are of the type
commonly associated with Late Prehistoric culture in the Upper Ohio Valley.

Corner-notched points. The points of this type do not compose a well
defined category at the Rohr Shelter. Only six points or 2.08% of the
total point sample were assigned to this group. Their distribution was one

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

177

Fig. 9. Rohr Shelter. Artifacts from level 1 (0"-6")

Fig. 10. Rohr Shelter. Artifacts from level 2 (6"- 12")

178

Annals of Carnegie Museum

vol. 35

Fig. 11. Rohr Shelter. Artifacts from level 3 (12"-18")

Fig. 12. Rohr Shelter. Artifacts from level 4 (18"-24")

•II

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

179

F.g. 14. Rohr Shelter. Artifacts from level 6 (30"-36") in row 1-3, and from
level 7 (36"-42") in row 4

180

Annals of Carnegie Museum

vol. 35

point (16.66%) in the first level, three points (50.00%) in the second level,
and two points (33.33%) in the third level. All of these points came from
levels which produced pottery. Corner-notched points stand apart from
the expanded-stem type only by their deeper and more finely chipped
notches, their greater width in relation to length, and their generally thinner
cross-section. Points such as these are similar to those known for late Early
Woodland and early Middle Woodland manifestations.

Stemmed, points. All the points assigned to this type possessed well de-
fined stems with parallel sides. The bases of the stems were either flat or
rounded. The majority of these points were four centimeters or more in
length and were the largest points found in the excavation. Most of the
stemmed points were made from crude materials which could not be finely
chipped. From looking at these points one could easily get the impression
that they should be old, but their distribution does not support this view.
Of the total 19 points of this type, nine points (47.39%) were found in the
first level, nine points (47.39%) in the second level, none in the third level,
and one (5.26%) in the fourth level. Only the point in the fourth level
was found in an area not containing pottery. Three of these stemmed points
with rounded bases and made of fine-grained chert are similar to those
associated with the Adena culture. The remaining 16 points appear closely
related typologically to some Ohio Valley late Archaic complexes.

Large expanded-stem points. During the sorting of the expanded-stem
points it became apparent that the large points of this group were found
more commonly in the upper levels of the excavation in contrast to the
greater depth distribution of small points with expanded-stems. Any point
of the expanded-stem group four centimeters or more in length was put
into the large expanded-stem type and all those below four centimeters in
length were classified as the small expanded-stem type. The large points
were more crude and usually made of a poorer variety of flint than the
small points.

The 10 points of the large expanded-stem type were distributed six points
(60%) in the first level, one point (10%) in the second level, two points
(20%) in the third level, and one point (10%) in the fourth level. Seven
of these points were found in levels containing no pottery. The one point
found in the fourth level was unusual in that it had broad serrations
along the cutting edges.

The large stemmed points closely approximate the distribution of the
parallel-stemmed points and are probably culturally related. They are also
part of a long tradition of expanded point making which began early in
Archaic times in the Upper Ohio Valley.

Small expanded-stem points , Of the 288 points found in the excavation,
134 or 46.52% were of the small expanded-stem type. All of the points in
this category were less than four centimeters in length. A great variety
of flints was used in their manufacture, but generally the material was
finer textured than that used for the large expanded-stem points.

Small stemmed points were distributed throughout all levels from the first
to the sixth with 22 (16.41%) in the first, 26 (19.4%) in the second, 46

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

181

(34.32%) in the third, 29 (21.64%) in the fourth, 9 (6.71%) in the fifth,
and 2 (1.49%) in the sixth. They are the dominant type of point in all
levels except the first, where triangular points are more abundant.

In the Upper Ohio Valley small stemmed points are found as the major
type of point on the majority of Archaic sites where they appear to have
remained in use over a long period of time.

Large side-notched, points. The side-notched sample of points was divided
into two groups according to size. Any point four centimeters or longer was
considered large while any point less than four centimeters was small. Un-
like the distribution of the large expanded-stemmed point, which was con-
fined to the upper levels, the large side-notched point was found through-
out all levels. Of the 20 points of this type there were five (25%) in the
first level, three (15%) in the second level, two (10%) in the third level,
eight (40%) in the fourth level, one (5%) in the fifth level and one (5%)
in the sixth level.

As was true of the large expanded-stem points, many of the large side-
notched points were of coarse material and not nearly so finely chipped as
the small type. The one large-stemmed point found in the sixth level was
unique because of the fine flint of which it was made and the excellent
workmanship.

Small side-notched points. Small side-notched points were second in abun-
dance at the Rohr Shelter. The sample of 65 points represented 22.56% of
the total 288 points. Small side-notched points were found in all point-
producing levels and presented the following distribution. Eight points
(12.3%) in the first level, 10 points (15.38%) in the second level, 22 points
(33.84%) in the third level, 19 points (29.23%) in the fourth level, 4 points
(6.15%) in the fifth level, and 2 points (3.07%) in the sixth level.

The small side-notched points and the small expanded-stem points had a
number of common features. Both types were wide in ratio to length and
usually rather crudely chipped. During the sorting of the points it was
often difficult to decide clearly in which type many of these points should
be placed. A wide variety of materials was used in the making of both types.

Bifurcated points. Only three points assigned to this type were found
at the Rohr Shelter. Such points are small, less than three centimeters long,
and relatively broad with a deeply notched base and well defined shoulders.
All of these points were found in the lower levels and were not associated
with pottery. Their distribution was as follows. One point (33.33%) in the
fourth level, and two points (66.66%) in the sixth level.

In the Upper Ohio Valley, bifurcated points occur in Archaic manifesta-
tions. As yet, their distribution is not well known as they occur only sparsely
on all sites from which they have been reported.

Miscellaneous points. Four points found during the excavations could
not be placed in the types just described.

In the first level was found a tapered-stem point which was quite differ-
ent from any belonging to the stemmed-point type. From weakly formed
shoulders the stem tapers abruptly to the base which is straight. The material
of which this point was made is a crude chert which could not be finely
chipped.

182

Annals of Carnegie Museum

vol. 35

The second point of interest in the miscellaneous group was a small
lozenge-shaped point found in the first level. The basal portion of this
point makes up approximately one-half the total length of the point. It
was 3 centimeters in length and 2.8 centimeters in width. The base is slightly
rounded.

In the third level was found a finely chipped lanceolate blade 4.5 centi-
meters in length and 1.9 centimeters in width. This point has a thinner
cross-section than any of the stemmed or notched points found in the ex-
cavation. The base is slightly rounded. The material of which the point
was made is a reddish, soft chert which has been weathered.

A crude lanceolate point 3.8 centimeters in length and 2 centimeters
in width was found in the fourth level. The sides of this point are parallel
and the base is straight. The material of this point is a coarse-grained chert.
There was no pottery associated with this point.

Projectile-point summary. The various types of projectile points and
their distribution at the Rohr Shelter present some interesting insights into
the cultural sequences of the Upper Ohio Valley. When information per-
taining to the point types is combined with that of the pottery types and
their distribution, there emerges evidence for four major cultural complexes—
Late Prehistoric, Middle Woodland, Early Woodland, and Archaic.

The small triangular points were found only in the uppermost levels
and usually on or very near the surface. Such points are well known from
all Late Prehistoric sites throughout the Ohio Valley. Their association
with shell-tempered pottery adequately demonstrates their cultural context
at the Rohr Shelter.

Corner-notched points were represented by only a few specimens. All of
these points were confined to levels containing pottery of predominantly
limestone or grit temper. Culturally they are generally considered to be
of late Early Woodland and Middle Woodland origin.

Points with parallel stems were basically concentrated within the first and
second levels where they were closely associated with grit-tempered and lime-
stone-tempered pottery. A few of the stemmed points with rounded bases
and made of fine-grained flint are similar to those known from Adena
sites. At the Rohr Shelter stemmed points probably were first used during
late Archaic times.

Large expanded-stem points were also concentrated within the upper-
most levels where pottery occurred, but a few of these points were found
in the upper non-pottery levels where they appear to be part of a late
Archaic complex.

In contrast to the large expanded-stem points, and parallel-stemmed
points, large side-notched points were found in all levels and were equally
divided between pottery and non-pottery levels. They appear to be the large
counterpart of the small side-notched points so common on many Archaic
sites and found in all levels at the Rohr Shelter.

Small expanded-stem and small side-notched points constituted 69.08%
of the entire point sample from the excavation. Both of these types were
found in all levels. The small expanded-stem point was the dominant type in

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

183

all levels except the first where the triangular type was more abundant. The
small side-notched point was second in abundance in all levels except the
first. Small points were basic items in the tool complex of many Archaic
sites in the Upper Ohio Valley. At the Rohr Shelter there is evidence that
they were used for a long time before the introduction of pottery and that
they persisted for some time after new cultures had entered the area.

The small bifurcated points were too few in number to be a reliable
sample, but their placement in the lowest levels would indicate that they
also belonged to an early Archaic tradition.

Scrapers and drills

There were a number of scrapers, drills, and other chipped-stone objects
found throughout the various levels of the excavation. An attempt was
made to classify these objects in generally accepted categories. In the follow-
ing paragraphs each of these categories will be described and the distribution
of such objects according to levels will be noted.

Side-scrapers. There were 68 side-scrapers found throughout the first
six levels of the excavation. The distribution was 6 (8.82%) in the first
level, 8 (11.76%) in the second level, 18 (26.47%) in the third level, 23
(33.82%) in the fourth level, 2 (2.94%) in the fifth level, and 11 (16.17%)
in the sixth level. Nearly 79.42% of the side-scrapers were found in levels
which did not contain pottery. As has just been noted, the greatest con-
centrations of side-scrapers occurred in the third, fourth, and sixth levels
which are considered to be Archaic in time. Side-scrapers varied in shape but
all possessed one or two straight cutting edges which in the majority of the
examples showed secondary chipping or retouching. Side-scrapers were used
by all peoples who inhabited the Rohr Shelter from early to late times, but
they appear to have been a more important item in the tool kit during
Archaic times.

End-scrapers. There were 25 end-scrapers found in the excavation. All
of these scrapers were found below the upper two levels in which pottery
was found in abundance. Their distribution was 5 (20%) in the third level,
6 (24%) in the fourth level, 12 (48%) in the fifth level, and 2 (8%) in the
sixth level. This distribution would indicate a definite Archaic context
for end-scrapers at the Rohr Shelter. All of these scrapers were well made
and exhibit in most examples fine secondary chipping. None of the end-
scrapers was large and many of them were very small. They ranged in
length from 3.5 centimeters to 1.5 centimeters. The material used for their
construction was usually a fine-grained chert. This contrasted with the wide
range of materials used for side-scrapers.

Crude scrapers.. This category included many small pieces of flint which
probably had been used only once and then discarded. Such pieces lacked the
definite form and secondary chipping common to the side- and end-scraper
categories. Nearly every piece of flint removed from the excavation showed in-
dications of use. There were 480 such pieces with the distribution 63 (15.44%)
in the first level, 83 (17.29%) in the second level, 96 (20%) in the third level,
105 (21.87%) in the fourth level, 105 (21.87%) in the fifth level, 21 (4.37%)
in the sixth level, and 7 (1.45%) in the seventh level. Such items were thus
distributed throughout all levels with the greatest number occurring in the

184

Annals of Carnegie Museum

vol. 35

fourth and fifth levels. As time and cultural markers these crude scrapers
were of little value.

Miscellaneous scrapers. There were seven specialized scrapers of various
types which could not be included in the above categories. In the first level
was found one small scraper shaped in general outline like an end-scraper
but with a concave, gouge-like cutting edge. One stemmed, bunt-like scraper
was found in the second level and one similar object was recovered from
the fourth level. In the fifth level were found two small gravers made of
fine-grained flint with triangular cross-section and sharp, cutting points.

In the seventh and lowest level producing artifacts an interesting chipped
stone gouge was found. This tool measured 3.2 centimeters in length by 1.8
centimeters in width. It was finely chipped from a light gray chert. The
lower surface was deeply grooved to produce a concave cutting edge. Also
from the seventh level was a crude chopper-like tool with evidence of having
been used as a scraper.

Drills. Thirteen drills or fragments of drills were found. All of these
drills appear to be either of the straight or slightly expanded base varieties.
Their distribution was five (38.46%) in the first level, four (30.76%) in the
second level, one (7.69%) in the third level, one (7.69%) in the fourth level,
and two (15.38%) in the fifth level. No drills were found in the sixth and
seventh levels, and 69.22% of them were found within the upper two levels
containing pottery.

Unidentified chipped-stone artifacts. There were 26 pieces of chipped
stone which could not be assigned to specific categories. Most of the objects
in this group were point tips or other portions of points not large enough
to permit definite placement as to type. Such fragments were found through-
out the first to the sixth level.

Polished-stone artifacts

Only a few items of polished stone were found in the excavation. A few
hammerstones were found near the fire-pits in the third and fourth levels.
Two pitted stones were recovered from the same levels. There were no
axes, celts, or adzes. In the second level was found a small piece of hematite
with a shallow groove cut into its surface. It was impossible to identify this
fragment as any specific object. In the first level was found a rim fragment
of a steatite bowl. The only stone objects with important cultural connations
were atlatl weights.

Atlatl weights. Five small atlatl weights were unearthed. One was found
in each level from the first to the fifth. All of these weights were of the
bar or rectangular type. None of the weights was whole. Each weight was
represented by a fragment comprising approximately % to ^2 of its total
length. The largest of these weights could have been no longer than M/i
inches by 13A inches in width. The remaining four weights were shorter
and narrower with the smallest one measuring about four inches in length
by one inch in width.

Fine-grained standstones were used for the manufacture of the bar weights.
Except for the one found in the first level, none of the bars was carefully
finished. They had been roughly pecked to shape and then rubbed to a
semi-smooth finish.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

185

Bar atlatl weights are commonly associated with Archaic manifestations
in the Ohio Valley. Webb (1950, p. 551) considers the bar weight as the first
type to be used in the Archaic shell-middens of Kentucky. Before the use
of bar weights, the atlatl was probably made entirely of wood.

BONE TOOLS

Bone refuse was common in the first to the fourth levels of the excava-
tion. Bone was not well preserved except in the first and second levels.
Below the second level the deterioration of bone became progressively
marked. The absence of bone below the fourth level may be due to the
slight acidity of the soil and the long time span represented by those levels.

Only 23 worked pieces of bone were found. In the first level were found
eight specimens consisting of two small pins, three fragments of awls, one
cut antler, and three antler tips. The second level produced eight objects
consisting of four fragments of awls, three antler tips, and one worked
antler base. In the third level were five specimens consisting of three frag-
ments of awls, one notched pin, and one worked fragment of deer skull. The
fourth level produced only one awl tip and one worked fragment unidenti-
fiable as to form.

All of the awls, with the exception of one turkey leg bone awl found in
the third level, were made of splinters of deer bone. The tips were usually
well formed but little attention had been given to the remainder of the tool.
None of the bone tools was culturally distinctive.

Table 4 gives the distribution of points, and Table 5 the distribution of
all the artifacts found during the excavations at the Rohr Shelter.

DISCUSSION

The materials found in the various levels of the Rohr Shelter present
some interesting insights into the cultural developments of the Upper Ohio
Valley. Compressed within the first and second levels of the site were objects
representative of three major cultural divisions— Early Woodland, Middle
Woodland, and Late Prehistoric. The pottery sample from the excavation
was not large, but distinctive types such as heavy grit-tempered sherds with
cordmarked exteriors and interiors belonging to Early Woodland, lime-
stone-tempered pottery of Middle Woodland times, and shell-tempered
sherds common to Late Prehistoric manifestations were all represented.

Stemmed projectile points with rounded bases were commonly associated
with the heavy grit-tempered pottery. A variety of notched and stemmed
points was in use during Middle Woodland times when limestone-tempered
pottery was in vogue. Small triangular points found near the surface at
Rohr Shelter belonged to the users of shell-tempered pottery during the
Late Prehistoric period.

The ceramic development represented in the Rohr Shelter covers a time
span of nearly 2500 years. Pottery-using groups accounted for only about
one-third of the debris deposit. Beginning at the third level and continuing
through the seventh level were remains of non-pottery-using peoples who
may have inhabited the shelter over a period nearly twice as long as that
of the pottery makers. The distribution of projectile point types indicates
that perhaps for a considerable time peoples of both the ceramic and non-

TABLE 4. ROHR SHELTER. DISTRIBUTION OF POINTS BY LEVELS

186

Annals of Carnegie Museum

vol. 35

£ oc ^ mo ® m f
C !>■ mo CO i— '

h

s

oo H
O0 c/0
CM '

X

w

c/o

£ CM — <
o ®

CO 00 CM »■
C OO

(_ GO (M CM CM

2

2 £ C 00 00
r c o
H PQ *-*

<

CM 050 OJ

be

X i2 'ac

W - ‘qj

X £
X

X

X

OOOCMOOrjiCMxO1-^
rH bl - CO ...

C ^

gl

^ °

00 mo CO CM —I

CM OO O CO CO
CM CM ^ CM

t/3 as

jj 8

G C

S w

cb — .
Oh g j
X y
y .2
y S

[3"

J S

ifO SO CM 00 I-H 1—1 O
CM

c/o

pit <o

X

a

OO CO

X

n -h oo w

o g

X b

W Ch

J £

C w
y jr
vo b

H §

^ y
.h-c
® •£
y

b w
c £
c c/o

tJ II
cX
Sx

CO

^ -H ^

X y

2 c

2

Dd ,

h-v

o/O C

w

H «

i—l 3

W c/o

E

c/o

&

X

o

c4

<*0 CO CO

ob

lf5 in rH

CO 00 CO O O CM

in C CM CM

0-, oo OO 00 ®0 CM i—i

Ob HH 1M -H

y

> ,—i CM SO 'bH
y

-1

I- •h’C

CM b ^

9 a

c^Oh
.2 qj

2 ^
o

*H II Cb

( c/o

w ^

J y cm t- co

PQ S CO oo

< c -

h *

y

> ~ CM SO

y

Total 939 288 68 25 480 15 5 26 5 23 1172

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

187

ceramic traditions were using the shelter. It is entirely possible that there
was little actual change in population; only the addition of new ideas
concerning the making of pottery and new styles of points.

The materials from the Rohr Shelter have given a number of insights
into the development of Archaic culture in the Upper Ohio Valley. No
other site now known in the area presents such a long temporal series.

In the lowest levels of the site there was a predominant use of small
side-notched and expanded-stem points. Occasionally a large, slender side-
notched point was used. Small bifurcated-base points also appear to be
part of the early Archaic complex. All of these point types, except those
having the bifurcated base, continued to be used until after the introduction
of pottery.

The most apparent changes occurred in the late stages of the Archaic
when large stemmed points became a common item of the complex. Such
points were of both the parallel- and expanded-stem types. These types
with modifications and refinements were carried over into the early pottery-
using cultures.

Small end-scrapers of the “thumbnail” type were important items in the
tool kit throughout the Archaic, but they were absent from the major
pottery levels. Side-scrapers were used from early Archaic through ceramic
times. Drills were more abundant in levels containing pottery but they
make their appearance midway in the non-pottery levels.

Of particular interest are the atlatl weights found at the Rohr Shelter.
All were of the bar type which has been considered an early form. None of
the more elaborate winged types occurred.

The Rohr Shelter undoubtedly presents only a limited picture of the
total Archaic complex. In other areas of the eastern United States much
richer Archaic complexes have been found. Unfortunately, in the Upper
Ohio Valley no such sites have as yet been reported. As of this date it has
been necessary to piece together the story of Archaic times from a number
of small sites which have produced only a limited number of artifacts.

Here no mention has been made of the relationship of the Rohr Shelter
materials to those known from other sites in the Upper Ohio Valley and
the eastern United States. In the following section of this article the Rohr
materials have been correlated with those found at the Dixon Shelter and
at other sites in the Upper Ohio Valley.

RELATIONSHIP OF DIXON AND ROHR SEQUENCE
IN THE UPPER OHIO VALLEY ARCHAIC

In both the Dixon and Rohr shelters it was possible to trace a sequence of
cultural events beginning in Archaic times and extending into the Late
Prehistoric. Both shelters produced pottery sequences beginning with heavy
grit-tempered cordmarked sherds which gave way to medium-thick grit- or
limestone-tempered sherds which in turn disappeared with the introduction
of shell-tempered pottery. It was, however, in the non-pottery producing
levels that new information for understanding the cultural development in
the Upper Ohio Valley was found. The sequence of stone artifacts, particu-

188

Annals of Carnegie Museum

vol. 35

larly points and atlatl weights, has given new insights into the formation
of the Archaic.

The Archaic occupation at the Rohr Shelter seems to have begun at an
earlier time than at the Dixon Shelter. At Rohr, the Archaic debris ac-
counted for approximately one-half of the cultural deposit, while at Dixon
it composed only about one-third of the total deposit.

The Archaic picture at each shelter begins with the use of side-notched
and expanded-stem points. Many of these points are very small. The
presence of small bar atlatl weights at the Rohr Shelter indicates that throw-
ing sticks were used to propel spears. These were probably spears with
compound shafts like those used by the Eskimo in historic times. Consider-
able conjecture has centered around the way in which these small Archaic
points were used. Ritchie (1944, p. 321) believed that the many small points
of the Archaic suggested the use of the bow. After working with the Shriver
Site materials (Stewart and Dragoo, 1954, p. 110) which produced hun-
dreds of small points but no atlatl weights among its many artifacts, I was
inclined to agree with Ritchie on the early introduction of the bow. How-
ever, the finding of bar weights at Rohr cast serious doubts upon this
belief. In recent conversations with me, Mr. Ritchie expressed the view
that the small points were used on compound spear shafts rather than on
arrows as he had previously conjectured.

Small points remained in use at both shelters well up into ceramic periods.
Near the upper margin of the non-ceramic levels larger stemmed points be-
came more prominent in the inventory of artifacts. Previous to the Dixon
and Rohr excavations, large, crude stemmed points were generally con-
sidered as being somewhat older than small notched and expanded-stem
points. Their placement near the end of the Archaic at the Dixon and
Rohr shelters calls for re-examination of some Archaic sites in the Upper
Ohio Valley which were considered early because of the presence of such
points. Another factor involved is the possible intermingling of two distinct
cultural traditions. This point will be more clearly shown later in this
section when other Archaic sites in the Upper Ohio Valley are discussed.

At both the Dixon and Rohr shelters, small, finely worked end-scrapers
were important items in the Archaic tool assemblage. They remained in
use throughout the Archaic, but appeared in greater abundance during the
early phases of the Archaic at Dixon and Rohr. Small end-scrapers have a
long history beginning with Paleo-Indian cultures and continuing into Late
Prehistoric times.

Along with end-scrapers there were a number of less distinct types of
scrapers found at Dixon and Rohr. Most noticeable among these are finely
worked side-scrapers and small triangular cross-sectioned gravers or “burins.”
Nearly every piece of flint found in the Dixon and Rohr excavations showed
evidence of having been utilized for scraping purposes. A common feature of
most Archaic sites in the Upper Ohio Valley is the great abundance of
scrapers. It is my belief that such tools were used for extensive wood and
bone working. If these perishable objects were available for study, the pic-
ture of Archaic life would undoubtedly be enriched.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

189

So far in this study data from only two stratified sites have been presented.
The picture presented is of limited scope in understanding the Archaic in
all its ramifications. Only small family groups and hunting parties occupied
the shelters. The full cultural inventory of larger groups living in open
villages was not present. Certain tools perhaps were not needed by the people
during their stay in the shelters and are missing from our lists. The most
important aspect of the shelter data is the chronological framework it
provides to which we can attach specific artifact types. With such a frame-
work it is possible to arrange other Archaic sites and materials in a logical
sequence. Thus, new insights are obtained regarding the growth of specific
items, and local developments and contacts can be more clearly traced. In
the paragraphs which follow, attention will be focused upon the several
Archaic manifestations known in the Upper Ohio Valley. (See Fig. 15.)

Archaic sites in the southern Upper Ohio Valley

Archaic materials in surface collections are known from several sites in
the areas surrounding the Dixon and Rohr shelters, but only one of these
sites is well enough known to present a picture of Archaic life. This site,
known as the Gay Shriver site (36Grll), was discovered and worked by Dr.
Paul R. Stewart of Waynesburg College. A preliminary report on the site
was made by Stewart and Dragoo (1954).

The Gay Shriver site is situated almost one mile southeast of Waynesburg
in Greene County, Pa. It is situated on a flat-topped hill probably repre-
senting a remnant of the so-called Harrisburg erosion surface, and commands
a fine view of the surrounding area. It lies above one of the finest flowing
springs in the county. The hill is so sandy that many years ago it was
called the “Sand Hill”, and large quantities of sand were dug for use in
road building. This sand comes from the weathering of sandstone which is
located stratigraphically about half way from the base to the top of the
Washington Formation and is of local significance only, and has been given
no particular geological name.

Several years ago the hill had a number of large productive chestnut
trees upon it. On the slope of the hill still exists a small virgin forest of
black and scarlet oak with some hickory. Forests of this type are generally
called “Oak-Hickory Climaxes” (Illick, 1928). Before the extinction of chest-
nuts, oak-hickory climaxes were often interspersed with chestnut trees. Such
climaxes renew themselves for many thousand years and there is no reason
to cloubt that such forests covered this hill during Archaic times. The forest
floor has an association of blueberries ( Vaccinium ) and other acid-loving
plants. Soil tests in the portion of the site covered by virgin forest showed
strong acidity (pH4.5-5).

The site proper covers less than an acre on the hilltop. Dr. Stewart over a
period of many years searched for artifacts when portions of the site were
under cultivation, and in 1951 he made a number of test pits. No artifacts
were found more than two feet below the surface, and practically all speci-
mens occur within plowshare depth. Approximately 1200 identifiable arti-
facts plus several thousand crude pieces of flint were deposited in the
museum at Waynesburg College.

190

Annals of Carnegie Museum

vol. 35

Fig. 15. Major Upper Ohio Valley Archaic sites

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

191

The majority of the projectile points are of the corner-notched, corner-
removed, or expanded-stem varieties. It is difficult to classify these points
into well defined categories because of their crudeness and inconsistent
shape. For example, a point may be corner-notched on one edge and side-
notched on the other. Any one point could easily be classified in any one
of the above mentioned varieties depending upon the inclination of the
classifier. The largest points are no more than 2 inches long and most are
less than 1^2 inches. It is not uncommon to find points less than one inch
long. Many of these small points are quite blunt while others have well
defined tips. Most are rather thick and nearly as wide as long. (See Fig. 16-17.)

Among the minor point types the side-notched points are most numerous
followed by those with stems. Pentagonal, bifurcated, and basal-notched
points occur but are quite rare. There are only eight broad-stemmed and
lanceolate points in the collection. (See Fig. 18-19.)

A wide variety of material was used in the manufacture of the projectile
points. In fact, it is notable that the chalcedony, chert, jasper, quartzite, etc.,
cover a far wider range of materials than those found on several Late Pre-
historic sites located in the same general area. A number of the artifacts
are made of poor quality dark chert found near Waynesburg. Outside of
this local flint, no large number of artifacts is of any one type of material.
A great variety of materials was noted in the Dixon and Rohr excavations
and small surface collections from other Archaic sites manifest the same
heterogeneity.

Some drills and perforators are known from the Shriver site. Of the 16
unbroken drills, 10 have the notched base. Each of the remaining six drills
has a straight to slightly expanded base. Both side and corner notches occur
on the notched drills. Most of the drills are crudely fashioned. (See Fig. 20.)

Scrapers and flake knives are abundant. Unspecialized chips and flakes
were most commonly used in the manufacture of these scrapers and knives.
Many of the flake knives resemble those used in Hopewellian times, but
they are generally smaller and less finely made. Many of the scrapers show
evidence of having been retouched several times. In the flint samples that
have been gathered from the site, there are several hundred crude scrapers.
Some of these show considerable use while others were apparently used
only once and then tossed aside. The small end-scraper so common in the
Archaic levels at Dixon and Rohr is not well represented in the Shriver
collection. (See Fig. 20.)

A number of small leaf-shaped blanks have been found at the Shriver
site. The largest of these blanks are little more than two inches long while
some are less than one inch long.

Among the heavy polished stone tools are two full-grooved axes, two
small grooved adzes, and three broken ax or adz fragments. All of these tools
are badly battered and broken. Considering the number of specimens that
have been found on the Shriver site axes and adzes were not an abundant
tool. No axes or adzes were found in the Dixon and Rohr excavations.
(See Fig. 21.)

Pitted stones and hammerstones are common in the Shriver materials. The

192

Annals of Carnegie Museum

vol. 35

Fig. 17. Shriver site. Typical small expanded-stem and notched points

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

193

Fig. 18. Shriver site. Stemmed and side-notched points

Fig. 19. Shriver site. Blanks, wide-stemmed points, and points of weathered

materials

194

Annals of Carnegie Museum

vol. 35

Fig. 20. Shriver site. Scrapers and drills

Fig. 21. Shriver site. Polished-stone tools

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

195

pitted stones differ from those found on many later sites in that they were
made from angular blocks of sandstone rather than from smooth pebbles.
Some of the hammerstones are crudely shaped spheres while others are only
utilized pebbles.

No pottery, bone implements, or bone refuse were found; and only a
few tiny fragments of mussel shells were discovered. With the exception of
five slate and three hematite fragments, no whole ornaments were found.
Two of the slate fragments appear to be of Laurentian origin. One is a por-
tion of a tube or possibly of an atlatl weight; the other could be a knife.
The other three fragments are of local carbonaceous slate.

The several excavations made at the Shriver site failed to uncover any
true fire-pits or other sub-surface features. Scattered over almost the entire
site, however, are stones showing evidence of firing. Burials are also unknown
at the site. The high acidity of the soil has probably destroyed all human
remains as well as all other bone refuse.

With some conjecture we can reconstruct the village and portray the life
of the inhabitants of the Shriver site. This site, although comparatively level,
probably has been lowered a few inches by weathering, moving the artifacts
downward a short distance. Chamberlin and Salisbury (1921, p. 84) estimated
that the entire Mississippi Valley has been lowered approximately one-half
foot in the past 5000 years. Granting that the top few inches of the site have
disappeared since the Archaic occupation, so far no deep refuse pits, shelters,
or burials have been found. Shelters were probably constructed of a light
pole framework covered by bark or skins similar to the lean-to. Fires for
cooking were built on the surface of the ground and banked by the stones
which litter the site today. Refuse was not abundant and consisted mainly
of bone which has been destroyed by the acid soil. Life was simple and re-
volved around the gathering and hunting of food. There was little time for
frills and the cultural elaboration which came in later times with the intro-
duction of agriculture.

There is a great possibility that the inhabitants did not live on this site
during the winter season. The absence of evidence of substantial houses
and deep fire-pits indicates that these people did not subject themselves to
the cold winds which batter this hilltop during the winter. Very likely the
inhabitants chose this locality for a summer and fall residence because it
was near a fine spring and upon a hill of sour sand where underbrush was
not rank. The chestnut trees along this and other ridges furnished them
with considerable food in the fall. Wild pigeons and turkeys used such
chestnut ridges as their roosting places and areas from which they gathered
most of their provender. There was abundant small game as well as deer,
elk, and bear to add to the menu. During the months of late summer and
early fall there were various types of berries and nuts to spice their food
with variety. With the beginning of the cool days of autumn, the people
began their movement to more sheltered areas. It was during the winter
months that rock shelters such as Dixon and Rohr saw their greatest occupa-
tion. Remains of large and intense fires are a common feature of the rock
shelter.

196

Annals of Carnegie Museum

vol. 35

The material culture of the Shriver site shares many similarities with
the materials found at the Dixon and Rohr shelters. This is particularly true
of point types. The small notched and expanded-base points of Shriver are
very much like those found in the lower and middle levels of the Dixon
and Rohr Archaic levels. The broad-stemmed points considered as late in the
Dixon and Rohr sequence are very rare at Shriver. On this basis I am now
inclined to consider Shriver as representative of Middle Archaic times in
the Upper Ohio Valley and not Late Archaic as I did in the earlier report
on this site (Stewart and Dragoo, 1954, p. 110). The problem of the broad-
stemmed point as a time marker will come up again later in this section
with discussion of certain sites in the Ohio Valley proper where they have
been considered as Early Archaic (Mayer-Oakes, 1955b, p. 19).

The absence of atlatl weights at the Shriver site was mentioned in con-
nection with the possible use of the bow in Archaic times. Considering
the large number of artifacts found at Shriver over several years of collect-
ing by Dr. Stewart, the only logical conclusions that can now be drawn
are that either the weight and atlatl were not used or that the atlatl was
constructed entirely of perishable material. I believe the latter to be more
nearly correct. At the Rohr Shelter several crude bar weights were found
throughout the middle and upper Archaic levels. Only in the lower levels
were they missing. Since the bar weight has been considered as the earliest
form of weight (Webb, 1950, p. 351), the absence of weights at Shriver may
indicate an earlier Archaic manifestation than has been admitted in the past.

Similiarities of the Shriver site materials with the Brewerton Focus of New
York state (Ritchie, 1944) have been stated previously (Stewart and Dragoo,
1954). These similarities are particularly evident when the flint objects are
compared. The thick, broad-bladed, notched and expanded-stem points;
bifurcated base points; pentagonal points; notched-base drills; flake scrapers
and knives; and small leaf-shaped blanks are nearly identical in form with
those found at the Oberlander and Robinson sites by Ritchie (1944, p.
235-249). There are, however, important differences which will be discussed
more thoroughly in the next section of this article.

A number of small Archaic sites in southwestern Pennsylvania similar to
Shriver were recorded by Carnegie Museum’s Upper Ohio Valley Archeo-
logical Survey. None of these sites is well known. The materials in small
surface collections from them are similar to those found at Shriver and in
the Dixon and Rohr excavations. Some temporal differences may become
evident when these sites are more thoroughly investigated and larger collec-
tions of materials for study are available.

Approximately twenty miles east of the Dixon Shelter in Garrett County,
Md., a number of Archaic sites were reported by Frank R. Corliss, Jr.
(1951, 1954) and recorded by Carnegie Museum. Garrett County is geo-
graphically on the dividing line between the Ohio and Potomac watersheds.
The major stream in the area is the Youghiogheny River of the Ohio drain-
age. Most sites known for the area are located around the shores of Deep
Creek Lake on the headwaters of the Youghiogheny.

This lake, of artificial origin, has eroded into and removed the topsoil
from nearly forty small sites along its shores. None of these sites has pro-

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

197

duced pottery. At 18Gal Corliss (1951, p. 1) found side- and corner-notched
points made of local chert obtained from a quarry site about six miles away.
Along with the points were found scrapers, drills, cache blades, hammer-
stones, knives, and many flint flakes.

An important feature of 18Gal was the finding of three cache deposits
of artifacts. One of these caches contained 10 objects consisting of three
corner-notched (expanded-stem) finished points, one turtle-back blank, and
six roughly chipped large leaf-shaped blanks. The material for all these
objects was a gray chert of local origin.

When the Garrett County sites are considered as a group they present
considerable variations in types of points. Represented in the collections
are small expanded-stem, side-notched, and bifurcated points similar to
those found at Dixon and Rohr. Some of the sites, particularly 18Ga31, pro-
duced crude, tapered, stemmed points which are wide, with a thick cross-
section. These points are not generally known from other Archaic mani-
festations in the Upper Ohio Valley. At both Dixon and Rohr shelters a few
points with wide, tapered stems were found in the upper levels of the
ceramic zones. Thus, at Dixon and Rohr such points are late in the Archaic
sequence and perhaps attest to contacts with the Garrett County area.

Corliss (1954, p. 9) believed the large tapered-stem points to be representa-
tive of Early Archaic. In light of the Dixon and Rohr evidence, combined
with the finding of a winged bannerstone at 18Ga2 (Mayer-Oakes, 1955,
p. 91) and the presence of cache blades, 1 am inclined to consider the
materials as representative of the Late Archaic. In many respects the Gar-
rett County materials are similar to those of Witthoft’s (1953) Transitional
Period Cultures of eastern Pennsylvania and the coastal region. Broad
points with contracting stem, and the presence of cache blades are common
features of the Transitional Period Cultures. The broad, contracting-stem
point tradition is generally absent from the Ohio Valley except on its ex-
treme eastern fringes. The occurrence of such points in Garrett County,
Md., and in the Dixon and Rohr sequence presents important evidence of
the overlapping of two distinct cultural traditions, in eastern Pennsylvania
and adjacent areas, the Transitional Period Cultures appear as a base
upon which later Adena and Hopewell-like cultures developed in those areas.

Except at the sites already mentioned, little is known about the Archaic
in much of the Monongahela, Youghiogheny, and Cheat River drainages.
Numerous small sites were located but materials from them are scarce. The
few local collections checked by Carnegie Museum workers contain evidence
of a wide distribution of the small notched and expancled-stem points found
in the Archaic levels of the Dixon and Rohr shelters.

Ohio Valley proper Archaic sites

Our attention will now turn from the southern extensions of the Upper
Ohio Valley to the Ohio River proper in western Pennsylvania, eastern
Ohio, and northern West Virginia. Most of our information comes from
three small shell-middens and several non-shell-bearing sites.

Prior to 1953 almost the only well known Archaic site in the Upper Ohio
Valley was the East Steubenville site (46Br31) on the east bank of the Ohio

198

Annals of Carnegie Museum

vol. 35

River opposite Steubenville, Ohio (Mayer-Oakes, 1955a, p. 132). This site
is a small shell-midden on a shelf projecting from a steep hillside more than
a hundred feet above the present water-level and approximately 250 yards
from the river. Excavation of the site was begun in 1938 by Walter Caldwell
of Wellsburg, W. Va., and Wilbur Harris of Short Creek, W. Va., aided
on several occasions by Walter Singer of Wellsburg and Elmer Fetzer of
Weirton, W. Va. Notes taken by Fetzer and much of the material in the
various collections were made available to Carnegie Museum for study.

The most conspicuous feature of the site is the accumulation of refuse
from the shellfish which probably constituted a major item of the people’s
diet. The hunting of other animals was attested to by the finding of bones
of deer, elk, bear, turkey, turtle, and fish. The only features other than the
refuse accumulation, known for the site are a flexed human burial and two
dog burials. The site appears to have been occupied only during Archaic
times by a small band of people.

The most numerous artifacts in the collections from East Steubenville
were chipped-stone tools consisting mainly of projectile points. Two major
types of projectile points from the site have been named by Mayer-Oakes
(1955a, p. 140). One type, the “Steubenville Lanceolate”, is a crude lanceo-
late form usually with a square base. The other type, “Steubenville Stemmed”,
is a square-based point with a broad, parallel stem. The lanceolate type is
not as well defined as is the stemmed type. Several of the lanceolate type
(Mayer-Oakes, 1955a, p. 138) appeared to be weakly shouldered variations
of the stemmed form. Most of the lanceolates are crude and have the general
appearance of blanks rather than of finished tools. Those that are finished
possess usage marks similar to those found on blades used as knives. The
“Steubenville Stemmed” type is well defined in characteristics and seems to
have a specific cultural and spatial distribution. (See Fig. 22.)

A few small notched points as well as straight and expanded-base drills,
scrapers, crude leaf-shaped blanks, and choppers are also part of the chipped-
stone inventory. The notched points are similar to those found in the
Archaic levels of Dixon and Rohr and at the Shriver site. (See Fig. 22.)

Among the polished-stone tools found at East Steubenville are three-
quarter-grooved adzes with pointed polls, plain adz, three-quarter-grooved
ax, bi-pitted stones, hammerstones, and notched pebble net-sinkers. One
crude bar-shaped stone may have been used as an atlatl weight. (See Fig. 23.)

A number of bone items including various forms of awls, notched pins,
conical projectile points, conical and cylindrical antler drifts, cut and per-
forated bear jaw, perforated stemmed bone point (harpoon), and an un-
usual split and perforated cylindrical awl were found. Almost all bone objects
known for the Archaic in the LTpper Ohio Valley have come from shell refuse
deposits where the alkalinity of the soil was maintained at a proper level
by the presence of the shells. (See Fig. 23.)

The second shell mound for which there is some information (Mayer-
Oakes, 1955a, p. 132) is the New Cumberland site (46Hkl). It is a small
shell-midden on top of the bluff overlooking the Ohio River just north of
New Cumberland, W. Va. The site is now about 310 feet above the present
river level and nearly one-half mile away. This site is known only from

1958 Dragoo: Archaic Hunters of the Upper Ohio Valley 199

lililslll

iilii II

\

j v

i i i k &

ii

v.

t I 4

□

4 1 5 16 1 7

Fig. 22. East Steubenville site. “Steubenville Stemmed” and “Steubenville
Lanceolate” points, notched points, drills, and blanks

200

Annals of Carnegie Museum

vol. 35

Fig. 23. East Steubenville site. Bone and polished-stone tools

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

201

surface collections gathered by A. Michael, the land owner, and Elmer
Fetzer of Weirton, W. Va. A tabulation was made of both collections and
some of the items in the Fetzer collection were illustrated by Mayer-Oakes
(1955a, p. 133).

Of the 20 points tabulated from the New Cumberland site, 11 or 55.5%
are of the broad-stemmed “Steubenville Stemmed” type. The remaining
points are smaller notched and expanded-stem forms typical of the Archaic
sites already discussed in this article. A straight drill, scrapers, and crude
choppers of chipped stone are also known for the site.

Included among the polished-stone tools of the New Cumberland site are
one full-grooved ax, two crescent bannerstones, one bar atlatl weight, and
two notched pebble net-sinkers.

No bone implements, burials, or subsurface features are known for the
site. The present disturbed condition of the site seems, to preclude the pos-
sibility of future work adding materially to our knowledge of the site.

The third shell-midden, and the only one for which there is available
information derived from controlled excavations, is the Globe Hill site
(46Hk34-l) excavated by Carnegie Museum in 1953 and reported upon by
Mayer-Oakes (1955b).

The Globe Hill site is on the east bank of the Ohio River opposite the
north end of Blacks Island in Hancock County, W. Va. The site lies on a
hilltop nearly 150 feet above the present river bank. The main feature
of the site is. a shell-midden about 80 feet in length and 50 feet in width.
The deepest deposit of shell is only about 20 inches near the south end of
the midden. The entire shell deposit thins out toward the margins. Mixed
in with the shell are various artifacts of chipped and polished stone, heat-
cracked rocks, and bone refuse. Some artifacts have been scattered on the
surface surrounding the midden but there is no great concentration of
such items.

Although the site was carefully excavated by levels, no important differ-
ence in the artifacts was noted from the top to the bottom levels. No features
such as pits or burials were found during the excavation in 1953, but frag-
ments of human bone have been found recently by amateurs digging in an
area not worked at that time. The site appears to represent a single occupa-
tion by a small band of people who depended basically upon the shellfish
for subsistence.

Chipped-stone objects were abundant in the shell debris at Globe Hill.
The most common objects were projectile points. Broad-stemmed points of
the “Steubenville Stemmed” type made up 15% of the sample and lanceolate
points of the “Steubenville Lanceolate” type accounted for 14% (Mayer-
Oakes, 1955b, p. 21.). The remaining 71% of the points consisted of 38%
corner-notched or expanded-stem points, 15% side-notched, 1% single-
shouldered stemmed points, 1% triangular points, and 16% crude lanceo-
late points. Again, as was noted for the East Steubenville and New Cumber-
land sites, the lanceolate points are not a well defined category. Some of the
points classified as lanceolate by Mayer-Oakes (1955b, p. 9) appear to be
weakly shouldered stemmed points while those considered as crude lanceo-

202

Annals of Carnegie Museum

vol. 35

late are more like blanks than true projectile points. The ‘‘Steubenville
Stemmed”, however, seems to be a well defined type. (See Fig. 24.)

The corner-notched or expanded-stem points and the side-notched points
are similar to those recorded for Dixon and Rohr and the Shriver site.
(See Fig. 24.)

A number of scrapers and a fewT drills were found at Globe Hill. The
most common form was the side-scraper usually of the flake variety. Three
spoke-shave scrapers and several thin flake knives were also recorded. No
end-scrapers were noted at the site. Only one expanded-base and one straight-
base drill were found during the excavation, but one slightly expanded-base
drill from the site is part of the Elmer Fetzer collection.

A number of interesting polished-stone items are known for the Globe
Hill site. These include plain adz, beveled adz, grooved adz, celt, crescent
bannerstone, bar atlatl weight, notched pebble net-sinker, stone disk, bi-
pitted stone, hammerstone, polished hematite concretion, sandstone shaft
smoother, worked cannel coal, slate reel-shaped gorget fragment, and small
hand-pestle. A three-quarter-grooved ax was found about one hundred yards
from the midden deposit. This item may not belong to the complex. The
grooved adzes are crude examples of the type. They do not have well de-
fined raised ridges along the groove. The bar weights are rectangular pieces
of crudely smoothed stone. The crescent bannerstone is the most distinctive
spearthrower weight. (See Fig. 25.)

Among 3376 bone fragments found at Globe Hill there are a few worked
pieces including such items as a cylindrical drift, antler flakes, awls, spatula,
perforated deer astragalus and cut dog jaw. The most interesting item is
the perforated deer astragalus which is similar to items known in Lamoka
in New York (Ritchie, 1944) and some Tennessee Valley sites (Kneberg and
Lewis, 1955). Worth mention, also, is the presence of dogs. Globe Hill and
East Steubenville are the only two Archaic sites in the Upper Ohio Valley
for which dogs have been reported. (See Fig. 25.)

A few small shell-midden deposits are known for the hilltop areas sur-
rounding the Globe Hill site. No materials are known for these sites. Only
the three shell-middens just discussed have produced enough information
to make comparative studies possible. On the basis of these three shell-
middens and selected materials from other sites Mayer-Oakes (1955a, 1955b)
formulated a complex which he called the “Panhandle Archaic.” He be-
lieved the “Steubenville Lanceolate” and “Steubenville Stemmed” points
were the end product of the Paleo-lndian period and the earliest Archaic
points in the Upper Ohio Valley. These points were derived from late
Paleo-lndian “Scottsbluff” and “Starved Rock Lanceolate” types, which
came into the area from the west at about the time of the post-glacial
climatic maximum. Thus, to Mayer-Oakes, the shell-midden sites represent
the adoption of new food resources based on the collecting of shellfish and
the lessening of hunting activities formerly predominant in the Paleo-lndian
period.

Since stemmed and lanceolate points represent the bulk of the known col-
lection for East Steubenville, Mayer-Oakes (1955b, p. 20) considered the

1958 Dragoo: Archaic Hunters of the Upper Ohio Valley 203

ft*

«•

«

:

il.ai mi

: I .

****4#*4#46

#♦> * m

*14111

4 § H 'm t -j.

^ /■

Pi #

4*

fSl

m

MifiP

Rl

1

2

ZJ-

■ L L_

8 I? |l0 |

Fig. 24. Globe Hill site. “Steubenville Stemmed” and “Steubenville
Lanceolate” points, blanks, and notched points

204

Annals of Carnegie Museum

vol. 35

Fig. 25. Globe Hill site. Bone ancl polished-stone tools

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

205

East Steubenville site as older than Globe Hill where notched and expanded-
stem points made up a greater proportion of the sample. On this basis the
Shriver site was considered as quite late, as only a few stemmed and
lanceolate points were present. To support his chronological sequence based
on point types, Mayer-Oakes cited the presence of the grooved adz as the
major ground-stone tool at East Steubenville and equated it with the
presence in Lamoka of the beveled adz as the primary tool there. He assumed
that western New York Lamoka was the origin of the beveled adz in Early
Archaic times; therefore, the presence of a beveled adz at Globe Hill
and not at East Steubenville became support for the later relative date
for Globe Hill. The presence of the crescent-shapecl bannerstone at Globe
Hill and not at East Steubenville was cited as additional evidence for a
later date.

There are several elements in the “Panhandle Archaic” complex which
tend to confuse the picture both culturally and chronologically. The “Steu-
benville Stemmed” and “Steubenville Lanceolate” points are distinctive for
the “Panhandle Complex.” They occur in only minor numbers or not at
all at other Archaic sites. If such points are considered as the end product
of the Paleo-Indian period, we must assume that they remained in use for
a period covering several thousands of years during the Archaic. The Globe
Hill site which produced several of these points also possesses traits, such
as the crescent-shaped bannerstone, which are generally considered as late.
Both Globe Hill and East Steubenville possess a well developed ground-
stone complex. If we accept the assumption that ground-stone and polished-
stone tools are not common until late Archaic times, then we are faced
with the problem of a supposedly old point complex associated with much
later polished-stone items.

The small notched and expanded-stem points found at Globe Hill have
been considered as an indication of lateness (Mayer-Oakes, 1955b). The
Dixon and Rohr sequence, however, presents evidence that these small
points were in use in the Upper Ohio Valley for a very long period of time
and that their presence at a site does not necessarily indicate lateness. The
Dixon and Rohr sequence is of little aid in the chronological placement of
“Steubenville Stemmed” and “Steubenville Lanceolate” types. No typical
points of these types were found. All of the wide-stemmed points which could
be remotely associated with such types were confined to the upper Archaic
and lower ceramic levels.

The presence of beveled adzes at Globe Hill and not at East Steubenville
has entered into the picture as a factor in the chronological ordering of the
two sites. Mayer-Oakes (1955b, p. 22) has assumed that the East Steubenville
site is earlier because it does not have the beveled adzes. In the Lamoka and
Laurentian cultures of New York the beveled adz was found in early levels
and continued in use over a considerable span in time. Since the beveled
adz is known to be early in Archaic contexts, we must conjecture that the
grooved adz which occurs at East Steubenville is earlier than the beveled
adz. Thus, the temporal, spatial, and cultural relationships of the grooved
adz become of crucial importance in the determination of the time and cul-

206

Annals of Carnegie Museum

vol. 35

tural affiliations of the “Steubenville Stemmed” and “Steubenville Lanceo-
late” points.

The grooved adz is best known from sites in New England, New Jersey,
and a few scattered sites in eastern New York and eastern Pennsylvania.
The grooved adz was present in the late Archaic Coens-Crispin complex of
New Jersey. (Cross, 1941). In eastern Pennsylvania grooved adzes have been
found in the Philadelphia area and the Schuylkill Valley where they are
perhaps related to the late Coens-Crispin complex of New Jersey (Witthoft,
1957, personal communication). According to Rouse (1957, personal com-
munication) the grooved adz is common in collections from Connecticut, and
perhaps belongs to the Sebonac Focus of Late Woodland. In eastern New
York this tool is more gouge -like than adz-like, having a more or less
concave lip and trough-shaped surface. The grooved gouge occurs sparingly
in the Hudson Valley and in the Lake Champlain area. Ritchie (1957, per-
sonal communication) relates this item to the Vosburg complex which he
believes to extend from middle to late Archaic.

The grooved adz is an unknown tool in the Archaic of the lower Ohio
Valley and the Tennessee Valley. Thus, its development took place in the
Northeast, and from available evidence this occurred in late Archaic times.
Whether it developed first in the Upper Ohio Valley or in New England is
as yet unanswerable.

Archaic materials similar to those known for East Steubenville and Globe
Hill have been found at several sites along the Ohio River. There are only
four sites, however, for which sufficient information is available to make
comparisons. These sites are the Half-Moon site (46Br29), Watson site
(46Hk34), McKees Rocks (36A1 16), and the Wheeling College site (460h22).
No definite shell-midden deposits are known at these sites, but they all
produce Archaic materials as well as items from later manifestations.

The Half-Moon site (46Br29) is an extensive bottom-land site in Brooke
County, W. Va., about four miles southwest of Weirton (Fetzer and Mayer-
Oakes, 1951. Mayer-Oakes, 1955). The site once consisted of four small
Adena burial mounds and several areas containing culture debris. Recent
construction work has destroyed much of the site including all but one of
the mounds.

Most of the materials from this site which can be assigned to the Archaic
on a typological basis were found on the surface by Elmer Fetzer of Weirton.
Included among these objects are the crescent-shaped bannerstone, three-
quarter grooved adz, full-grooved ax, “Steubenville Stemmed” and “Steuben-
ville Lanceolate” points, net-sinkers, and pitted stones. All of these materials
are similar to those found at the East Steubenville and Globe Hill shell-
middens. The Half-Moon site, however, is best known for its Adena mounds
and associated objects such as reel-shaped gorgets, tubular pipes, celts,
hematite hemispheres, copper beads, and Half-Moon ware (Fayette Thick)
pottery.

The McKees Rocks site (36A1 16) is similar to the Half-Moon site. This
site is on the west side of the Ohio River at the mouth of Chartiers Creek
approximately three miles northwest of Pittsburgh. It is a large site situated
on high ground consisting of an Adena mound and areas producing village

1958 Dragoo: Archaic Hunters of the Upper Ohio Valley 207

debris of several time periods. The mound was partially excavated by Car-
negie Museum in 1896 and the artifacts have been made available to several
workers for study. The first major report on the site was made by Carpenter
(1951) and two subsequent analyses have been made by Mayer-Oakes (1955a)
and M (Michael (1956).

Included among the many artifacts excavated in 1896 at the McKees Rocks
mound were several “Steubenville Stemmed,” “Steubenville Lanceolate,”
and small - notched and expanded-stem points which typologically can be
assigned to the Archaic. Also present are the grooved adz and the beveled
adz. The presence of this material in the mound fill has been considered as
due to the scraping of a nearby Archaic village for earth (Mayer-Oakes,
1955a, p. 152). Many of the objects from the mound are of Adena types and
include reel-shaped gorgets, hematite cones, celts, shell beads, bone awls,
and stemmed points of typical Adena form. Thus, in the McKees Rocks
mound materials, there were representative items from three known com-
plexes—Panhandle Archaic, Laurentian, and Adena. A fourth complex, repre-
sented in surface collections from the site and one intrusive burial in the
top of the mound, is the Late Prehistoric. This last mentioned material,
however, is quite late in the occupation of McKees Rocks. (See Fig. 26-27.)

On the terrace along the Ohio River just north of the Globe Hill site in
Hancock County, W. Va., is the large Watson site (Dragoo, 1956). The main
features of this site belong to the Middle Woodland period and include at
least two stone mounds, one of which was excavated by Carnegie Museum
in 1953. In the lowest levels of this excavation were found several pottery
sherds of Half-Moon ware which is representative of Early Woodland.
Stemmed points of typical Adena form and a few “Steubenville Stemmed,”
and “Steubenville Lanceolate” points have been found on the surface along
with a fragment of a grooved adz. A large number of points of the notched
and expanded-stem types also have been collected from the surface. Al-
though the Archaic elements of the Watson site are only superficially known,
there is a similarity in the contents and the sequence (from Panhandle
Archaic to Early Woodland) with the Half-Moon and McKees Rocks sites.

Several sites in Beaver County, Pa., lying in the Ohio Valley between
the Watson site and the McKees Rocks site have produced Archaic materials.
None of these sites, however, is well known and they are represented only
by surface materials. Site 36Bv22 has produced two grooved adzes, a beveled
adz, five plain adzes, a crescent bannerstone, broad-stemmed points, and
notched points. At site 36Bv52 a grooved adz and several “Steubenville
Stemmed” and “Steubenville Lanceolate” points have been found (Mayer-
Oakes, 1955a, p. 141). From site 36Bv3Q, currently under investigation by
Carnegie Museum, a wide range of Archaic and Early Woodland materials
has been collected. Among these materials are grooved adzes, full-grooved
axes, celts, crude choppers, many scrapers, including burin-like tools, broad-
stemmed and lanceolate points, notched points, crescent bannerstones, and
Early Woodland pottery. Future study of these sites should add much to
our knowledge of the Archaic.

In marked contrast to the sites discussed above is the Wheeling College
site (460h22) at Wheeling, W. Va., only a few miles down the Ohio River

208

Annals of Carnegie Museum

vol. 35

Fig. 26. McKees Rocks Mound. Stemmed and lanceolate points and blanks

Fig. 27. McKees Rocks Mound. Polished-stone tools

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

209

from the East Steubenville site. The debris of this site covers an extensive
area of a hilltop overlooking Wheeling Creek. No mounds or shell-middens
are known on or near the site. Over a period of many years several local
collectors gathered artifacts from the area. A recent report by Clifford M.
Lewis (1955) gives the contents of these collections and other pertinent data
concerning the site.

The range of materials known for the Wheeling College site indicates
that more than one culture was represented, but the great bulk of the
materials consists of small notched and expanded-stem points similar to
those found in the Dixon and Rohr excavations and common on the
Greene County, Pa., sites such as Shriver. Only a few points could possibly
be considered as examples of the “Steubenville Stemmed” or “Steubenville
Lanceolate” types.

Missing from the list of items from the Wheeling College site are the
grooved adz and the crescent bannerstone common to the sites of the “Pan-
handle Archaic.” Among the ground- and polished-stone tools from the
site are full-grooved axes, crude choppers, hammerstones, pitted stones,
notched pebble net-sinker, and smoothed pieces of hematite. Other items
including the three-quarter-grooved axes, celts, pottery sherds, and gorget
fragments have been found on portions of the site, but such items do not
seem to be associated with the notched and expanded-stem points from
the hilltop area of the site. Lewis (1955, p. 7) noted that pottery and
artifacts generally believed to be later in time than the Archaic were found
on the bottom-land.

On the basis of present evidence, the Archaic materials from the Wheeling
College site are more closely related to the lower Archaic levels of Dixon
and Rohr and to sites in Greene County, Pa., than to the “Panhandle
Archaic” as represented at East Steubenville and Globe Hill. Thus, the
materials from the Wheeling College site can be related to Archaic mani-
festations found over most of the Upper Ohio Valley and shown to be old
in the Dixon and Rohr sequence.

Returning now to the overall picture of the Archaic in the Upper Ohio
Valley proper, we are faced with the problem of two major Archaic com-
plexes (Laurentian and Panhandle Archaic) inhabiting the same area and
both supposedly having considerable antiquity.

The Laurentian culture, as manifested at the Shriver site, Wheeling College
site, and Dixon and Rohr shelters, has a wide distribution throughout the
Upper Ohio Valley and can be closely correlated with the Laurentian of
New York state where radiocarbon dates have established its presence there
at around 3500 B.C. (Ritchie, 1955). The small notched and expanded-stem
points so common to Laurentian culture were found beginning at the
lowest levels in the Dixon and Rohr sequence where they were also asso-
ciated with the simple bar atlatl weight. The full-grooved ax is the common
polished-stone tool present at the southern Laurentian sites while the plain
or beveled adz is the prevalent tool on the northern Upper Ohio Valley
and New York sites. The grooved adz is not common and when present is con-
fined to sites near the “Panhandle Archaic” sites. Also associated with the

210

Annals of Carnegie Museum

vol. 35

grooved adz are stemmed points similar to the “Steubenville Stemmed.”
Such points in the Dixon and Rohr sequence occur in the late Archaic
levels and early ceramic levels.

The position of Laurentian materials in the lowest levels of stratified
deposits, the distribution of such materials over much of the Ohio Valley
and Northeast, the simple form of certain objects such as the bar atlatl
weight at some sites, and the presence of local developments within the
general Laurentian framework, all indicate that the Laurentian was an old
and well established culture which in its late periods came into contact
with new cultural traditions. Such contracts account for the variable artifact
inventories present at some widely scattered Laurentian sites.

Considerable antiquity also has been claimed for the “Panhandle Archaic”
by Mayer-Oakes (1955b, p. 19) who has considered the “Steubenville
Stemmed” and “Steubenville Lanceolate” points as the end product of
Paleo-Indian culture surviving into Archaic times. On the basis of the in-
formation presented in preceding paragraphs of this article, l am unable
to see such antiquity for the “Panhandle Archaic.” My reasons for this belief
can be summarized as follows: 1. The “Panhandle Archaic” can be related
to Paleo-Indian culture only by a typological resemblance of the “Steuben-
ville Stemmed” and “Steubenville Lanceolate” points to similar types such
as “Scottsbluff” in the Plains. 2. Similar points were found in the upper
levels of the Dixon and Rohr shelters where they were considered as
transitional forms from Late Archaic to Early Woodland. 3. Typologically
the “Steubenville Stemmed” and “Steubenville Lanceolate” points also can
be considered as variations of stemmed points and lanceolate points which
occur at late shell-middens in Kentucky (Webb, 1946) and on Early Wood-
land sites in the Upper Ohio Valley which appear to be representative of
Adena. 4. The distribution of “Panhandle Archaic” sites is confined to or
is in close proximity to the Ohio Valley proper and closely parallels the
distribution of Early Woodland or more specifically, Adena sites. 5. The
presence of the crescent bannerstone and grooved adz as typical traits of
the “Panhandle Archaic” also indicates lateness for the complex.

At a site currently under investigation by Dr. Paul R. Stewart of Waynes-
burg College, additional evidence for the lateness of the “Panhandle Archaic”
is being discovered. On this site, known as the Sproat site (36Grl6), in
Greene County, Pa., several “Steubenville Stemmed” and “Steubenville
Lanceolate” points have been found in association with fragments of tube
pipes and gorgets. Some of the stemmed points are similar to those found on
Adena sites. A few notched and expanded-stem points were found at the
site but they are in the minority. Future work at this site may do much
to clarify the chronological position of the “Panhandle Archaic.”

The oldest and most widespread Archaic culture in the Upper Ohio
Valley proper was the Laurentian. The strategic position of sites in the
Ohio Valley near the confluence of several major tributaries entering the
Ohio River from all directions permitted contacts with distant areas. These
contacts led to regional specializations, particularly in polished-stone tools.
For example, grooved axes are most common on southern sites while the

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

211

plain and beveled adzes are typical of the northern sites. The grooved adz
and associated stemmed points represent a new cultural thrust into the
Upper Ohio Valley from probably the northeast in Late Archaic times. This
new culture blended with the resident Laurentian to form the “Panhandle
Archaic” which became the base upon which the Early Woodland developed.

ARCHAIC SITES IN THE NORTHERN UPPER OHIO VALLEY

The northern Upper Ohio Valley is drained by two major rivers and
their tributaries. The eastern and northernmost portions of the area are
drained by the Allegheny River and its major tributaries— the Kiskiminitas
and Conemaugh rivers, the Clarion River, and French Creek. The western
area is drained by the Beaver River and its main tributaries, the Shenango
and Mahoning rivers.

Prior to Carnegie Museum’s Upper Ohio Valley Archeological Survey
program, started in 1950, little was known about the cultures represented in
the northern portion of the Upper Ohio Valley. Large areas still remain to
be investigated. Most of the information pertaining to the Archaic was
obtained from the surface of sites by the above survey and by recording
and photographing private collections. No major excavation has been con-
ducted on any Archaic site in the area.

Butler (1936, p. 27) recorded the presence of materials of a hunting and
fishing people belonging to the Archaic in Erie County, Pa. Her informa-
tion was based on the presence of beveled adzes and “narrow-bladed” points
in collections of surface finds. Carpenter (1942) reported on the results
of a site survey in northwestern Pennsylvania. He indicated the presence
of several small sites in which Laurentian materials had been found. Con-
sidering the beveled adz as the main Archaic trait. Carpenter (1942, p. 23)
stated:

“Unfortunately, all site collections that yielded beveled adzes were so intermixed
with the implements of other cultures that the former artifact represented the
sole trait for the Archaic Pattern. Beveled adzes have been found in collections
from Tionesta, West Hickory, Sugar Run, Tracy Run, Hodge Run, and Kiantone.
At no one place were they plentiful, and the largest number from any site was
three. The characteristic choppers and antler-tine pendants were absent, but the
thin celt-like blade or scraper characteristic of Archaic is found associated with,
and greater numbers than, the beveled adze.”

Carnegie Museum’s survey in the Allegheny drainage produced similar
information concerning the presence of narrow-bladed stemmed and notched
points, plain and beveled adzes, and crude chopping tools (Mayer-Oakes,
1955a, p. 50). Sites 8 and 9 in Chautauqua County, N. Y., and sites 1, 2, 3,
7 and 9 in Forest County, Pa., contained a high proportion of these artifacts.
In the Conemaugh drainage, sites 36In7, 36In8, 36In9; 36Cb2 and 36Cb3 pro-
duced similar materials.

Most of the information about the Archaic in the Allegheny Valley comes
from the Siggins site (36Fol) and the County Home site (36Fo3) reported by
Mayer-Oakes (1955a, p. 52). Both are large sites on high terraces along the
Allegheny River. Much of the material from both sites appears to be of
Archaic origin, but some items represent later Woodland occupations.

212

Annals of Carnegie Museum

vol. 35

Typical Archaic objects from the Siggins site (36Fol) are beveled adzes,
plain adzes, crude choppers, full-grooved axes, bi-pitted stones, hammer-
stones, bannerstone fragments, notched pebbles, stemmed points, notched
points, small bifurcated points, straight and expanded-base drills, end-
scrapers, and side-scrapers. Also included in the collection from this site
are several tube pipe fragments, elbow pipes, gorget and pendant frag-
ments, celts, and steatite sherds. Most of the notched and expanded-stem
points are similar to Laurentian, but the parallel-stemmed points are more
like Early Woodland forms and are probably associated with the pipe and
gorget fragments.

On the County Home site (36Fo3) have been found such items as beveled
adzes, plain adzes, full-grooved axes, notched and grooved pebbles, hammer-
stones, winged bannerstones, notched and stemmed points, many end-
scrapers, and several steatite sherds. Thus, this site presents a mixture of
Archaic and Early Woodland traits.

The presence of steatite vessel fragments at 36Fol and 36Fo3 is suggestive
of the Transitional Culture of eastern Pennsylvania described by Witthoft
(1953). Missing are the wide-stemmed points so common to that complex.
Because of this factor, Mayer-Oakes (1955a, p. 53) hesitated to postulate a
transitional period for the above sites, but considered the use of steatite
vessels to be part of the general late Archaic complex. Due to the close
proximity of these sites to the Susquehanna drainage where the Transi-
tional Culture is present, contacts between the Upper Ohio Valley and
the East in late Archaic times seems most likely. A similar situation was
previously mentioned in this article for the materials from Garrett County,
Md., and the presence of broad-stemmed points and sherds of steatite vessels
in the transitional (Archaic and Early Woodland zones) of the Dixon and
Rohr sequence.

The Boyer’s Run Shelter (36Ve6) in Venango County, Pa., has produced
a few Archaic artifacts at its lowest levels (Thompson, 1955). Although most
of the items found in the excavation of this shelter by A. C. Thompson be-
longed to ceramic periods, a few Laurentian-like notched and expanded-
stem points are represented in the collection.

The Allegheny Valley Archaic materials are basically similar to those of
the Laurentian in New York. Unfortunately, only materials from late sites
are represented in the collections that have been studied. Until further
work is done in the area, it is impossible to establish a valid Archaic sequence.

In the Beaver Valley many small sites located on benches and terraces
along both major and minor streams and producing Archaic materials
have been found. Only a few of these sites have been investigated and none
has been excavated.

During Carnegie Museum’s Upper Ohio Valley Archeological Survey,
several Archaic sites and local collections of artifacts from them were re-
corded for the Shenango Reservoir area in Trumbull County, Ohio, and
Mercer County, Pa. (Mayer-Oakes, 1953). On five sites (33Tr4, 36Mel6,
36Mel7, 36Me26, and 36Me28) were found both Archaic and Early Wood-
land materials. On four of these sites Middle Woodland items were also

1958 Dragoo: Archaic Hunters of the Upper Ohio Valley 215

present. Thus, our knowledge of these cultures, confined to sorting of col-
lections on a typological basis, is not clear because of this mixing of artifacts.

Included among the Archaic artifacts from the above sites are beveled
adzes, plain adzes, full-groovecl axes, several types of bannerstones (winged,
pick-shaped, rectangular with oval cross-section, and crescent), bi-pitted stones,
hammers tones, end-scrapers, drills, bifurcated-base points, stemmed points,
and notched points. The bifurcated-base and notched points are similar to
those found in the Archaic levels of the Dixon and Rohr sequence and other
Laurentian-like sites previously mentioned in this report. The stemmed
points are like those from Early Woodland sites and are probably asso-
ciated with the gorgets, pendants, fire-clay tubes, and cache blades also
found on these sites.

Other important sites in the Beaver Valley were reported by Mayer-Oakes
(1955a). Site 33Mh4 in Mahoning County, Ohio appears to have been occu-
pied both in Archaic and Early Woodland times or was perhaps a transi-
tional site. Mayer-Oakes (1955a, p. 74) considered the plain and beveled
adz, bannerstones, scrapers, and “Steubenville Stemmed” points found at
this site as typical Archaic artifacts. In his summary of Beaver Valley sites
Mayer-Oakes (1955a, p. 78) stated:

‘The overall impression one gets for this area is of an abundance of small
camp sites on benches or terraces along both major and minor streams. Most of
these sites represent very brief and probably intermittent occupations by wander-
ing hunters. Surface materials are of chipped and ground stone with the lanceolate
projectile points representing the continuation of late Paleo-Indian styles into
early Archaic times. The fuller complex of adzes, axes, bannerstones, and “Steuben-
ville Stemmed” points indicates an early period within the Archaic. A later
Laurentian-like unit possibly represents a developed Archaic period.”

Again, as was the common occurrence in the Ohio Valley proper, it is
important to note that the “Steubenville Stemmed” points are found on
sites with late Archaic and Early Woodland items. Thus, their associations
are with transitional sites late in Archaic times and not with materials
that reflect connections with the Paleo-Indian period.

The grooved adz, which occurs in association with the “Steubenville
Stemmed” points in the Ohio Valley proper, is a rare item in the Beaver
Valley; but it has been recorded for site 36Me9 where stemmed points were
also present. The plain adz, beveled adz, and full-grooved ax are common
items from the Beaver Valley sites.

At a number of small sites now under investigation by Carnegie Museum
in the Beaver Valley, typical Laurentian notched, bifurcated and expanded-
stem points occur in association with a simple ground stone complex con-
sisting of plain adzes, grooved axes, and hammerstones. Missing from many
sites are the more elaborate forms of bannerstones and the wide stemmed
points. The varied inventories from these sites indicate that they represent
occupations scattered over a long time during the Archaic.

SUMMARY

The Archaic complexes of the Upper Ohio Valley may be summarized as
follows:

E The earliest and longest Archaic occupation of the Upper Ohio

214

Annals of Carnegie Museum

vol. 35

Valley was by peoples possessing a Laurentian-like culture. Their subsist-
ence was based upon a hunting, fishing, and gathering economy. Most groups
depended upon game from the forests, but some peoples living in the Ohio
Valley proper also collected shellfish. The smallness of the villages indicates
that during most of the Archaic the population was never large. Typical
manifestations of the early Archaic are found in the lower levels of the
Dixon and Rohr shelters and at numerous small surface sites scattered over
the entire Upper Ohio Valley. Basic items in their tool kits included small
notched, bifurcated, and expanded-stem points, a variety of scrapers in-
cluding end-scrapers, full-grooved axes, plain adzes, bar atlatl weights, bi-
pitted stones, and hammerstones. In late Archaic times new items such as
the beveled adz and elaborate forms of the bannerstone, such as the crescent
and winged types, were added to the basic items. Many Laurentian-like
Archaic groups may have persisted in the more isolated areas of the Upper
Ohio Valley well into Early Woodland times.

2. During late Archaic times along the eastern and southern portions of
the Ohio Valley in close contact with the Potomac and Susquehanna River
drainages, influences, and perhaps actual groups of people, of the Transi-
tional Cultures of eastern Pennsylvania and the seaboard introduced to the
Archaic new items including broad, tapered-stemmed points, cache blades,
and steatite vessels.

3. In the Upper Ohio Valley proper at the late Archaic level, new
peoples using broad-stemmed points, lanceolate blades, grooved adzes, and
crescent bannerstones mingled with resident Laurentian-like groups to form
the “Panhandle Archaic,” which in turn appears to be the base upon which
Early Woodland (particularly Adena) developed in the area. The distribution
of objects and sites associated with the “Panhandle Archaic” closely follows
the distribution of Early Woodland, and materials of both complexes are
commonly found on the same sites. Although the major occupation of the
“Panhandle Archaic” occurs in the Ohio Valley proper, several sites in the
Beaver River drainage show a mixture of this complex with Early Woodland
materials.

4. A number of local developments and specializations of certain tools
occurred in the various areas of the Upper Ohio Valley during the Archaic.
On sites in the northern Upper Ohio Valley the common polished-stone
tool is the plain and beveled adz which is the most common Archaic tool in
the adjoining area of New York. Adzes occur on the southern sites but are
less common than the full-grooved ax which is widely distributed on
sites down the Ohio Valley and common on the shell-middens in Ken-
tucky. The full-grooved ax is probably the oldest polished-stone, chop-
ping tool. The grooved adz was used mainly in the Ohio Valley proper
during late Archaic times. Its origin and relationships to other areas are
not yet clear.

5. No positive evidence can be shown for direct contact between the
Archaic and the earlier Paleo-Indian cultures. This may be due either to
inadequate sampling, or a true chasm actually may have existed between

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

215

the two periods in the Upper Ohio Valley. Only continued and intensive
research can solve this problem.

Having set forth the major tenets of the Archaic in the Upper Ohio
Valley, attention must be turned now to the relationship of this material
to surrounding areas and in general to the Archaic as known in eastern
North America.

ARCHAIC MANIFESTATIONS OF EASTERN NORTH AMERICA

The Upper Ohio Valley has been the path for movements of many peoples
over long periods of time. It is the natural gateway and most direct route
to the northeast from the Mississippi and Ohio Valleys. Thus, Archaic assem-
blages of the Upper Ohio Valley have important relationships with mani-
festations in surrounding areas in eastern North America. In the following
paragraphs these relationships and their implications for understanding
the origin and spread of the Archaic are investigated. (See Fig. 28.)

The closest area to the Upper Ohio Valley for which detailed comparative
information exists on the Archaic is New York. The recognition and elucida-
tion of several diversified Archaic manifestations in New York developed
through the excavations and publications of William A. Ritchie under the
auspices of the Rochester Museum of Arts and Sciences between 1925 and
1940, and since then in his capacity as New York State Archeologist. Ritchie
(1955) defined an Archaic period culture sequence involving two distinctive
traditions which he named Lamoka and Laurentian. While Lamoka was
given priority in time, its later interaction with a phase of Laurentian was
believed to have occurred at Frontenac Island. Ritchie (1955, p. 3) envisioned
the Frontenac complex as a product of an acculturational situation where
various recombinations of older items and new traits occurred.

The Lamoka tradition (Ritchie, 1932, 1955) was characterized by a primary
concentration on a bone and antler industry. Rough and polished stone
tools were moderately emphasized, while chipped-stone tools received only
weak expression.

Most Lamoka bone work was of a utilitarian nature relating to hunting
and fishing activities. A few items, however, were ornamental forms, the
most outstanding being a unique and diagnostic class of pendant-like ob-
jects often decorated with red ocher (Ritchie, 1932, p. 106). Fishing equip-
ment of bone included both hooks and gorges. There were bone awls of
various styles but those made from deer scapula, and turkey radius were
particularly numerous and distinctive. Other objects of bone included antler
punches, beaver incisor chisels, bone whistles, and perforated deer astragali.

The Lamoka polished-stone inventory included crudely executed celts,
thin celt-like scrapers, plano-convex adzes, and beveled adzes. This last item
Ritchie (1955, p. 5) considered as diagnostic of Lamoka and as having
originated from the plano-convex adz in western New York where it has
been found in large numbers.

Other rough stone items in Lamoka included hammerstones, bi-pitted
stones, anvil stones, crude choppers, notched net-sinkers, multi-pitted stones,
mu Hers, cylindrical pestles, flat grinding stones, and basin-shaped and
trough-shaped mortars (Ritchie, 1955, p. 5). These last items coupled with

216

Annals of Carnegie Museum

vol. 35

Fig. 28. Major central and northeast Archaic manifestation

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

217

the finding of acorns and other plant remains demonstrated the importance
of wild plants in Lamoka economy.

Ritchie (1932, 1955) stated that the typical Lamoka projectile point was
crudely percussion-chipped, small, weakly side-notched or stemmed, slender-
bladed and with a thick base. Such points were used on spears rather than
on arrows.

According to Ritchie (1955, p. 5) ceramics, copper or shell objects, har-
poons, barbed bone points, bannerstones, grooved axes, gouges, and polished
slate articles were unknown to Lamoka.

The burials associated with the Lamoka complex were of the flexed
type (Ritchie, 1932). Grave goods were absent or scanty. Physically the
Lamoka people belonged to a dolichocephalic group.

Lamoka occupation was heaviest in western, central and southern New
York. Lamoka-type points have been found in the lower levels of Hudson
Valley sites and on certain sites farther east in lower New England (Ritchie,
1955). The main components of Lamoka are in western New York at
Lamoka Lake, Geneva, and Scottsville.

Ritchie’s (1944, 1955) second major division of the Archaic was the
Laurentian which he believed to be later than Lamoka in the Northeast.
In contrast to Lamoka he characterized Laurentian as a culture in which
chipped-stone industries predominated over bone work. Since there was
a wide range of projectile point types in Laurentian, Ritchie (1955, p. 6)
believed that the fusion of discrete “local strains” probably occurred prior
to the appearance of several focal complexes encompassed within the
Laurentian tradition in the Northeast.

Among the chipped-stone tools of Laurentian were several types of
scrapers, bi-facially chipped knives, retouched flakes, and points which were
prevailingly broad bladed, thick, with straight stem, or side or corner notches
(Ritchie, 1955). The first trianguloid points supposedly made their appear-
ance during Laurentian.

A number of polished-stone tools have been attributed to Laurentian
(Ritchie, 1955, p. 6). Included were celts, grooved axes, plano-convex adzes,
beveled adzes, and gouges. The gouge came into use after the departure of
the beveled adz and was considered as late in Laurentian (Ritchie, 1955, p. 6).
Also included in the Laurentian inventory are plummets and atlatl weights
in a variety of forms including notched, grooved, rectangular, and per-
forated styles. Ritchie (1955, p. 6) believed that the bannerstone and the
grooved ax diffused northward from the southeastern Archaic.

Laurentian burials were often made in village refuse dumps but extended
inhumations and cremations occurred along with flexed burials (Ritchie
1955, p. 7). Some grave goods, including red ocher, were often placed with
the burial. Intentional burials of two breeds of dogs also occurred in
Laurentian.

The contents, distribution, and relationships of Laurentian to other
groups in New York are complex. Ritchie’s (1944, 1955) attempts to draw
together the traits of Laurentian tossed perhaps too many disparate things
into one category. The excavated sites which produced Laurentian materials

218

Annals of Carnegie Museum

vol. 35

also contained materials from later occupations. It was apparently not
possible clearly to separate these different manifestations. Although most
of the later materials were confined to upper levels, some late items were
found mixed with Laurentian objects in lower levels. Most of the detailed
information on Laurentian comes from three major sites— Robinson, Ober-
lander (Ritchie, 1940) and Frontenac Island (Ritchie, 1944). All of these
sites showed a mixture of materials. Several of the artifacts, such as winged
bannerstones and gouges, that were assigned to Laurentian are typologically
late in the Archaic. The presence of such objects at the excavated sites has
done much to tag Laurentian as a late Archaic complex. This is a contention
that must be more closely examined.

MacNeish (1952, p. 47) stated that stratigraphic data were such that the
Lamoka, Frontenac, and Brewerton foci could be arranged in chronological
order and their genetic relationships could be determined. MacNeish (1952,
p. 47) stated:

“At the Lamoka Lake Site, Brewerton Focus artifacts appear near the top of
the midden and above most of the Lamoka Focus materials (as well as under
Vinette 1 pottery). Since the Frontenac Culture has traits of both Brewerton and
Lamoka Lake, it is thought of by Ritchie as being chronologically between the
Brewerton Focus and the Lamoka Focus.”

Ritchie (1955, p. 3) stated that the priority of Lamoka over Laurentian
had been stratigraphically established. In a search of the original site re-
ports and subsequent publications, I have been unable to find demonstrable
proof for such stratigraphy. Ritchie’s recent field work, yet unpublished,
may hold the solution to this problem. Concerning the presence of Brewer-
ton artifacts over Lamoka materials at Lamoka Lake just quoted above
from MacNeish, Ritchie (1944, p. 307) was uncertain of the upper materials
and stated that “the upper level complex from Lamoka Lake is, on the
whole, still an enigma.’’ It is unlikely that Ritchie would have made such a
statement if the identification of the upper level complex was positive for
Brewerton. Ritchie (1944, p. 303) actually suggested relationships of the
upper occupation materials with the “Early” Focus of the Coastal Aspect,
but he also drew attention to the fact that the physical type differed.

The Frontenac Focus was established on typological grounds as chrono-
logically between Lamoka and Brewerton (Ritchie, 1944, p. 292. MacNeish,
1952, p. 47). No positive stratigraphy for the Lamoka and Brewerton ma-
terials could be determined. Lamoka traits found at Frontenac included
narrow-bladed stemmed and side-notched points, beveled adzes, choppers,
antler pendants, notched pendant-like objects, and other more general
traits. Brewerton traits at Frontenac included broad-bladed stemmed and
side-notched points, plummets, bannerstones, perforated single-barbed har-
poons, gouges, copper awls, and ground slate objects. Frontenac traits not
common to Lamoka or Brewerton included elongated pendants of shell,
receptacles of shell, new types of bone whistles, bird effigy combs, worked
wolf mandibles, perforated elk and bear teeth, and new varieties of barbed
harpoons. These last items are typologically late and bear some resemblance
to objects in later Point Peninsula and Iroquois of New York. MacNeish
(1952, p. 47) remarked that the presence of these items did not strengthen

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

219

the hypothesis of an early Brewerton group which moved into western
New York, amalgamating with and replacing Lamoka. The sophisticated
elements of Frontenac are more indicative of a late Archaic manifesta-
tion than an early one. Therefore, it is difficult to see the Frontenac site
as representing the contact of early Brewerton with the supposedly older
Lamoka. If, however, Lamoka does not possess the antiquity and priority
over Laurentian that has been attributed to it, the Frontenac situation of
typological complexity and lateness becomes more understandable. The
age of Lamoka becomes a crucial factor in northeast Archaic development
and chronology.

Since positive stratigraphic evidence for the priority of Lamoka over
Laurentian has not been demonstrated and the typological picture is in-
conclusive, it is necessary to turn to other types of evidence. Until many
more Lamoka and Laurentian sites are investigated and reported in New
York, this evidence must come from other areas.

Frequent mention has been made of the resemblance of Lamoka to the
Shell Mound culture of Kentucky (Ritchie, 1944, p. 322. MacNeish, 1952,
p. 47). Unfortunately, these references have been vague and without sup-
porting evidence, in May, 1956, with the co-operation of Dr. William S.
Webb, I selected Lamoka and Laurentian-like points at random from In-
dian Knoll and Annis Mound. The depth at which each of these points
was found in the midden was then determined from the card catalogue
and plotted on a graph. The results of this preliminary study reveal im-
portant new information on the relationship of Lamoka and Laurentian
to the Shell Mound culture.

The Annis Mound is one of the oldest of the shell-middens. Materials
from early to late Archaic have been found there. Laurentian-like small
notched and expanded-stem points selected at random from this site were
distributed in all levels from top to bottom. Of the 147 points in the sample,
43.53% were found above four feet, and 56.46% were found in the levels
between four and eight feet. They were most abundant in the 6- and
6.5-foot levels. This distribution presents a picture of long and continued
use of Laurentian-like points. Since such points were present at the low-
est level at Annis Mound, considerable antiquity is established for the
Laurentian-like points.

The distribution of Lamoka-like points at Annis Mound contrasts sharply
with the Laurentian distribution. Except for burial inclusions, where the
digging for the burial had destroyed the natural stratigraphy, the Lamoka
points are virtually all in the upper fourth of the midden. Of the 37
Lamoka points in the sample, 72.97% were found above the 1.5-foot level
and 83.78% were above the 3.5-foot level and there were no Lamoka points
in the lowest levels of the Annis Mound.

At Indian Knoll the entire random sample of 21 Lamoka points was
found above the four-foot level. Not one Lamoka point was found below
four feet to the bottom of the midden at seven feet. Of the points present
in the upper levels, 47.61% was found above the one-foot level.

The distribution of Lamoka points at both Indian Knoll and Annis

220

Annals of Carnegie Museum

vol. 35

Mound is most significant. Their placement in the upper levels at both
of these can not be an accident. Careful field notes and records of depth
for these points were made by Webb and his associates, and the methods
and techniques of excavation were of the finest. The conclusion that
Lamoka points are late in the Shell Mound culture of Kentucky seems in-
escapable. MacNeish’s (1952, p. 47) statement that Lamoka has relationships
with “the lower levels of the Annis Mound” does not agree with the facts.

The relationship of Laurentian to Shell Mound culture has been ignored.
Yet it is the Laurentian materials that are most abundant and widely dis-
tributed within the Shell Mound culture. The presence of Laurentian
points at the lowest levels at Annis Mound indicates that a Laurentian-like
tradition was a basic element in Shell Mound culture. No such evidence
has been presented for Lamoka.

The radiocarbon dates for the Archaic in New York and Kentucky are
of little aid in clarifying the chronology. Three dates have been obtained
for Lamoka (Ritchie, 1951). Two of these dates were taken from charcoal
samples at Lamoka Lake, while the third sample was removed from a
hearth at Frontenac presumed to be of Lamoka origin. Sample 288 from
Lamoka gave a date of 4369±200 years. As this sample was believed to be
contaminated by recent carbon, another sample from a different hearth
was dated at 5383±250 years. Ritchie (1951) accepted this later date as
more nearly accurate for the Lamoka site. The sample from Frontenac
Island which was presumed to pertain to the initial contact of Lamoka
with Laurentian was dated at 4930±260 years. A date for Point Peninsula,
which followed the Archaic in New York, has been given as 4400±260 years
ago (Libby, 1952). If the Point Peninsula date is correct, the dated Archaic
manifestations appear to be late as no great interval of time separates the
two periods.

Two recent dates from the Smith Farm site near Ellsworth Falls, Me.,
aid in the establishment of late Archaic and early ceramic contacts (Byers
and Hadlock, 1955). One sample taken from middle levels considered as
of Archaic origin was dated 3959±310 years. The other sample, taken from
the uppermost level with Archaic materials, was dated 3350±4Q0 years.
Immediately above the Archaic materials was a deposit comprising floors
in which sherds of Vinette I pottery and later Point Peninsula pottery types
were found. These dates would place the contact of the Archaic with the
Point Peninsula culture 1000 years or more later in Maine than New York.

The dates for the Shell Mound sites are even more perplexing than those
for New York. The first dates for the Annis Mound were 4900 ±250,
5149±300 and 7374±500 years ago. (Webb, 1951). The earliest date, how-
ever, was from a level above the two later dates. The shell sample dated
from the three-foot level showed an age 2000 years greater than the shell
sample from the six-foot level. Since the two samples were carefully collected
from the same profile, Webb (1951, p. 30) believed that possibly the two
samples had become interchanged in the laboratory. The date of 4900±250
was obtained from deer antler found at the 6.5-foot level. This date agreed
favorably with the one of 5149±300 years for the same level. Two additional

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

221

dates were obtained later from the Annis Mound (Libby, 1953). Both of
these dates were from composite samples of deer antler. The first was
dated 4289 ±300 and was from levels 1.5, 2, and 2.5. The second sample
was dated 4 3 33 ±4 50 and was from levels 5.5, 6, 6.5, and 7. While the first
group of samples from Annis Mound was similar to those for New York
Archaic, the second group was several hundred years later.

The dates from the Indian Knoll site were also inconsistent (Webb, 1951.
Libby, 1953). The first sample of antler tested from the one-foot level
was dated 53O2±30O years (Webb, 1951). Later, another antler sample
from the same one-foot level was dated 428 2 ±250 years (Libby, 1953).

The inconsistencies of the dates from the Kentucky Shell mounds make
them of little value and cast serious doubts upon the reliability of shell
and antler for dating purposes. Differences of 1000 years for dates from
the same level of a site make chronological comparisons of that site with
more distant sites extremely difficult. Until many more dates can be ob-
tained from sites and various levels within specific sites over wide areas of
the Northeast, carbon dates can be only suggestive and not conclusive evi-
dence for the age of a manifestation. Actual stratigraphy and comparison of
materials associated with specific levels still remain the best methods of
correlating chronology and relationships.

In the Upper Ohio Valley there are no well defined manifestations which
can be attributed to Lamoka. Stemmed points, plain adzes and beveled
adzes similar to Lamoka have been found on many sites, but such items are
usually associated with a predominantly Laurentian complex. There was
no indication of the Lamoka complex at either the Dixon or Rohr sites.
All materials from the lower Archaic levels at both of these sites are similar
to those attributed to Laurentian. The small notched points, expanded-
stem points, and bar atlatl weights found at Rohr are very similar to items
found in the lower levels of the Annis Mound in Kentucky.

The closest relationships of Lamoka to the Upper Ohio Valley are found
in the “Panhandle Archaic” complex. Present in both of these complexes
are numerous similar bone tools and plain and beveled adzes. The presence
of beveled adzes in the “Panhandle Archaic” is the most distinctive link
with Lamoka. Both Lamoka and “Panhandle Archaic” have individual traits
which differ. Both, however, have relationships and contacts with Laurentian
which is widespread over the Northeast. I believe that Lamoka and “Pan-
handle Archaic” are distinctive regional complexes which developed upon a
Laurentian base, subjected to differing influences from surrounding areas.
Some of these influences were probably carried by new peoples. As shown
by the presence of Lamoka-like points in the upper levels of the Kentucky
Shell mounds, the typological similarities of “Panhandle Archaic” items
with early ceramic manifestations, and the bone and polished stone traits
shared by Lamoka and “Panhandle Archaic,” both complexes appeared late
in the Archaic of the Upper Ohio Valley.

Since several objections have been stated against Lamoka and the “Pan-
handle Archaic” as representative of early Archaic in the Upper Ohio
Valley, the origin, distribution, and relationships of Laurentian become of
great importance.

222

Annals of Carnegie Museum

vol. 35

Ritchie (1955, p. 8) did not believe that the Lamoka complex had de-
veloped in situ from a Paleo-lndian base. The known Paleo-Indian assem-
blages in the Northeast, from the Shoop (Witthoft, 1952), Reagen (Ritchie,
1953), and Bull Brook (Byers, 1954) sites represent a chipped-stone industry
with fluted points, end- and side-scrapers based on trimmed flakes, gravers,
and some knife forms usually from retouched flakes. None of these items
has a counterpart in Lamoka. Conversely, Ritchie (1955, p. 8) also stated that
“we seek in vain for a single Lamoka trait in the Paleo-Indian inventory.”

The Laurentian, however, does possess traits that are similar to Paleo-
Indian. Ritchie (1955, p. 8) stated that “in the Brewerton Focus of the
Laurentian, on the other hand, a few elements of possible Paleo-Indian
derivation occur, viz., uniface flake, end and side scrapers and knives, and
a few lanceolate blades with parallel flaking.” All of these items, plus
gravers, have been found on sites producing Laurentian materials in the
Upper Ohio Valley. Gravers, end-scrapers, and side-scrapers were associated
with the Laurentian points found in the lower levels of the Dixon and
Rohr shelters.

Two fluted points were found at the base of the Annis Mound in Ken-
tucky (Webb, 1950). Although these points were believed to have been left
on the site before the dwellers of the shell-midden arrived, there are traits
within the midden which may have been derived from Paleo-Indian. Such
traits are thumbnail end-scrapers, gravers, and end-scrapers. As noted pre-
viously in this report, it is also within the lower levels of the Annis Mound
that Laurentian-like points were most numerous.

The Parrish site in Hopkins County, Ky., possessed an interesting mixture
of Paleo-Indian and Archaic materials (Webb, 1951). During the excavation
of this village a number of fluted points as well as typical Archaic artifacts
were found. Items of particular interest are crude flake side-scrapers, thumb-
nail end-scrapers, gravers, utilized flakes, and choppers. Webb (1951, p. 433)
assigned these objects along with the fluted points to Paleo-Indian; but most
of these items are also common in the Laurentian and occur on many of
the Upper Ohio Valley sites assigned to that complex.

The relationship of Laurentian to Paleo-Indian is far from being clearly
understood. It seems unlikely that Laurentian was a direct development
from Paleo-Indian. That the two cultures were in contact near the close of
Paleo-Indian times is indicated by the traits they share and certain sites
which possess remains of both. I believe the shared traits of Paleo-Indian
and Laurentian give added support to the priority of Laurentian over
Lamoka and other local complexes in the Northeast to which considerable
age has been attributed.

The origin of Laurentian has been sought by some workers in the sub-
arctic zone of North America. This position was expressed by Spaulding
(1955, p. 17) who stated:

“From our standpoint, the important feature of the Laurentian is that it cannot
be derived satisfactorily from the preceding cultures. It represents a new complex
of elements associated with a new physical type in the northeastern part of our
area. The gouges, ground slate points, and other elements point to a northern
origin, and an actual movement of people from the Canadian forest seems to

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

223

be the most likely explanation. The ultimate origin of the Laurentian complex
need not concern us here, but it is surely in some sense a representative of the
north European-Siberian-North American cultural tradition of the boreal forest
zone, and its appearance in the northeastern fringe of our area is a reflection of
the proximity of that fringe to the northern forests. Indeed, certain parts of the
area of Laurentian occupancy such as the Upper Peninsula of Michigan can
be considered essentially of boreal forest type from the standpoint of ecology.”

Ritchie (1955, p. 8) pointed out that it is not easy to find sites over the
immense span of northern forests of Canada between the northeastern
United States and the Alaskan bridgehead. Also, the recent views of Quimby
(1953, p. 235-326) concerning the presence of ice of the Wisconsin glaciation
in central and eastern Canada between 4500 B.C. and 2500 B.C. appear
to rule out that area as a convenient passage for the Laurentian invasion
of the Northeast. Since the Laurentian was already established in New
York during that time, Ritchie (1955, p. 8) suggested that “we should seek
a west to east route through the deciduous forest belt bordering the Great
Lakes in Altithermal times.”

I am in complete agreement with Ritchie’s suggestion that a west to east
passage below the Great Lakes should be sought for Laurentian. Present
evidence on the distribution of Laurentian material is indicative of such a
passage. That this route was used by Laurentian is most clearly seen when
we investigate the contents and relationships of local manifestations.

Ritchie (1951, 1955, p. 6) remarked on the peculiar distribution within
Laurentian of a dual category of ground slate tools comprising stemmed
spear or lance heads, knives, and the ulo or semilunar knife. These forms
occur separately or in various combinations on certain sites. Such items are
known particularly from the Brewerton Focus, and the Vosburg Focus
in New York, and the Vergennes Focus in Vermont (Ritchie, 1951. Bailey,
1939). The presence of slate items in Laurentian was considered to be by
Martin, Quimby, and Collier (1947, p. 505) indicative of influences from
the Dorset culture. Ritchie (1951, p. 50) suggested that the early occupation
of the coniferous forest zone of North America by nomadic hunting bands
from a similar Old World milieu was responsible for the establishment of
these traits, probably first in Laurentian and later in Eskimo culture. This
“proto-Laurentian” cultural stratum then underwent various regional elabo-
rations. In Maine, for example, there was a rich efflorescence of ground
slate points, gouges, adzes, and plummets which reach their apogee in the
Red Paint Culture (Ritchie, 1951).

In the Upper Ohio Valley slate objects are rare or absent on sites attrib-
utable to Laurentian. When such items are found, they are usually present
on sites adjacent to New York. Since slate objects are not common to the
Laurentian sites of the Ohio Valley, it seems reasonable to consider such
items as a specialization in the Northeast which came about through contacts
with early Eskimo cultures. If such items were part of the early Laurentian
when it passed through the Ohio Valley below the Great Lakes, it becomes
difficult to explain their absence on the sites in that region. Willey and
Phillips (1955, p. 751) summarized the Eskimo-Archaic problem as follows:

“We have already pointed out the difficulty of applying our classificatory

224 Annals of Carnegie Museum vol. 35

criteria to cultural developments in the Arctic. Here we are dealing with cultural
traditions so unique in total configuration as to make comparisons appear a trifle
ridiculous. Nevertheless, with the possible exception of the Denbigh complex, as
already noted, it is surprising how many of the criteria of our definition of
Archaic are applicable to Eskimo cultures of the far north. This is particularly
noticeable in the Dorset tradition of the eastern Arctic (Jenness 1930; Rowley
1940; Leechman 1943; Harp n.d.), whose relationship with northeastern Archaic is
a perpetual subject of discussion (Hoffman 1952). But even such climatic Eskimo
phases as Old Bering Sea (Collins 1937) and Ipiutak (Larsen and Rainey 1948)
would, if subjected to the same unfavorable conditions for preservation, appear to
be on the same general level of development as the well-known Archaic cultures
further south. There is not only a basic conformity in stone and bone technologies
but a great many shared elements of specific nature. Until fairly recently the
tendency was to attribute the presence of such traits in Archaic contexts to
‘Eskimoan’ influence. Now, with the far greater time depth allowed to North
American Archaic cultures, the problem of interchange between them and con-
temporaneous Eskimo cultures has been immensely complicated, and it would be
a rash student who would attempt to disentangle them.”

The presence of copper tools in some New York Laurentian sites indi-
cates important relationships with the Old Copper culture of Wisconsin.
Typologically the copper gouge, adz, celt, and awl of Laurentian are similar
to those of Old Copper (Ritchie, 1955). Until recently the Old Copper cul-
ture was little more than a technological tradition. Work at the Osceola site
(Ritzenthaler, 1946) and at the Oconto site (Ritzenthaler and Wittry, 1952)
in Wisconsin has produced excavated specimens, and Wittry (1951) con-
ducted a typological and distributional study of Old Copper artifacts in
collections in Wisconsin. Popham and Emerson (1954) did a similar distri-
butional study for Ontario and described the discovery of a cache of copper
tools at Farquar Lake, Ontario. Two summaries of the Old Copper culture
by Miles (1951) and by Wittry and Ritzenthaler (1956) appeared recently in
American Antiquity.

Wittry and Ritzenthaler (1956) noted the resemblance of many Old Copper
traits to those of the Kentucky shell-middens and to those of the Brewerton
and Frontenac foci of New York. The one bone whistle found with a burial
at the Oconto site has a parallel in the Annis Mound in Kentucky. The
burial complex of the Oconto site, however, has closer parallels with the
New York Laurentian manifestations where cremations also are present,
but there are differences in physical type between these early populations.
Cremations are lacking in Lamoka (Ritchie, 1944, p. 394) and also in the
Kentucky shell-middens, but they do occur in the lower levels of the Lauder-
dale Focus in Alabama (Webb and Dejarnette, 1942).

The radiocarbon dates of 5600±600 years and 7510±600 years (Libby,
1954, p. 740) for the Oconto site makes the Old Copper culture one of the
earliest dated Archaic manifestations in North America and also indicates
the earliest use of metals in the Americas and one of the earliest anywhere
in the world. The large number and diversity of forms of the copper im-
plements indicate that the Old Copper complex must have existed for a
long time. Some Old Copper sites such as Osceola (Wittry and Ritzenthaler,
1956) possess lithic traits that show contact with Woodland groups.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

225

Close resemblances of Laurentian and Old Copper are to be found not
only in the use of copper tools and cremations but also in the use of certain
lithic items. Although the chipped-stone complex of Old Copper is not
well known, the few points from the early Oconto site are very similar in
size and form to those of Laurentian. It seems entirely possible that the
underlying base of Old Copper and Laurentian was the same. In Wisconsin,
a specialized development of the use of copper tools began very early as a
regional expression of this early Archaic base. At a later time the use of
copper tools spread eastward around the Great Lakes into Ontario and
New York and became a part of Laurentian.

The relationships of the early lithic materials found in Canada to the
Archaic of the Northeast are not well known. Lee (1955a, p. 4) stated that
Lamoka traits for early Archaic are “not at all applicable to Ontario.” At
the Sheguiandah site on Manitoulin Island, Lee (1955b, p. 71) noted the
presence of materials which he considered to be an early variant of Lauren-
tian. Ridley (1954, p. 43) made the following correlation for materials at
the Frank Bay site, Ontario:

“Partial correlation of the Frank Bay Mattawan complex is geographically wide-
spread but exact duplication is unknown. Archaic material similar to that from
western New York is found elsewhere on the perimeter of Lake Nipissing, is re-
ported at several stations many miles north of Lake Nipissing, and is well repre-
sented by sites and surface collections over much of southern Ontario. Nevertheless,
though Mattawan and eastern Archaic share many features, differences are also
great, and the two are clearly not the same. Neither does the Mattawan complex
appear to be proto-Archaic, nor a regional variation of eastern Archaic, but seems
to be considerably older than this general phase. Similarly, there are suggestive
parallels with Paleo-Indian sites; for example Reagan (Ritchie 1953), Shoop
(Witthoft 1952) and Brohm (MacNeish 1952) with materials from Quebec (Rogers
and Rogers 1950), and with arctic stations, for example, Dorset (de Laguna 1946;
Wintemberg 1939). Once again differences are obvious, and in any event too little
is known of the Mattawan complex to justify interpretation at this time. Recently
Paul Sweetman and I discovered a site 20 feet above the shore of Lake Temis-
kaming, well to the north of Lake Nipissing. In a dark top stratum, artifacts
tentatively identified as Archaic, resembling Atlantic states Archaic, and Indian
Knoll, overlay a deep deposit of yellow sand and alluvial gravel. In the sand,
and well below the gravel, were small hearths and 8 Mattawan type artifacts.”

In the MacKenzie Valley (MacNeish, 1954, p. 234-253), stemmed and
notched projectile points comparable with Eastern Archaic forms have
been found in association with specialized tools such as burins, microblades,
and polyhedral cores. Wettlaufer (1955, p. 91-113) illustrated many points
from the Mortlach site in Saskatchewan which resembled those typical of
Laurentian.

There can be no doubt that typologically similar points occur both at the
Laurentian sites of the Northeast and Ohio Valley and at several sites in
Canada. Until many early Archaic sites are known in the Northeast and Ohio
Valley, it will be difficult to demonstrate what relationships may exist be-
tween these sites and those in Canada.

One of the most interesting Archaic sites in the Ohio Valley is the McCain
site in Dubois County, Ind. The work at this site was conducted by Rex K.

226

Annals of Carnegie Museum

vol. 35

Miller (1941) from 1929 through 1934. This work and the excellent report
which followed were completed before the Archaic manifestations in Ken-
tucky and the Northeast were well known. As a result, important relation-
ships of the McCain site materials with Laurentian have been overlooked.
Through the courtesy of Glenn A. Black of the Indiana Historical Society,
all the lithic specimens from the McCain site were made available to me for
study and revaluation. The information which follows is based on my
recent study of the materials as well as Miller’s (1941) report.

The McCain site is located on an ancient terrace along the East Fork of
White River in Boon Township, Dubois County, Indiana. This terrace has
been heavily cut by erosion and the river has shifted its channel until now
it is some distance away from the site.

Miller’s (1941, p. 10-16) excavations revealed that the site was covered in
many places by an occupation debris of humus-laden soil containing ash,
charcoal, bone, shell, burned stones, and artifacts to a depth varying from
2lA to 314 feet. Along the slopes of the terrace were refuse dumps as deep
as 414 feet. These dumps contained great quantities of mussel shells. No evi-
dence of house sites, permanent fireplaces, storage or refuse pits were
found. Large quantities of ash and burned stones indicating temporary fires
were found scattered throughout the debris. Artifacts were found in both
the village debris and the dump areas. Twenty flexed burials were found
in shallow pits in the debris.

Many fine bone objects were found at the McCain site. Their state of
preservation, due to the large amounts of shell in the debris, was excellent.
Included among the bone items are antler punches and drifts, awls of deer
and bird bone, bone scrapers and knives, and fish-hooks. However, the most
interesting bone objects are a number of bone pins with incised geometric
decorations. Some also have notches and many have a perforation at one
end. These and the other bone objects were illustrated by Miller (1941,
p. 35-47). The decorated bone pins are similar to those from Lamoka
(Ritchie, 1932, p. 106), and they are especially similar to the incised bone
items from the Annis Mound in Kentucky (Webb, 1950, p. 325).

Among the polished-stone tools of the McCain site, the 31 full-grooved
axes were the most important and the only large cutting tool except for
four crude, pecked-stone choppers. Three of the full-grooved axes were
made of limonite. Other items of stone included 56 hammerstones and 13
mullers. No atlatl weights were found. Most of the polished-stone tools,
particularly the full-grooved axes, are similar to those of the Kentucky
Archaic and other scattered Archaic sites throughout the Ohio Valley.

The chipped-stone tools of the McCain site present many interesting
similarities to tools from other Archaic sites in the Ohio Valley. Since
Miller (1941) touched only lightly on these items in his report, they are
described more fully here. All the chipped-stone tools were separated into
types. Some items were difficult to classify because of their indefinite form,
and could logically be placed in more than one category. This material may
be summarized as follows:

Small side-notched points. The 85 small side-notched points are variable

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

227

in outline but all possess crudely chipped side notches. Many of these
points have concave bases, and some have projecting ears. Basal thinning
occurs on several specimens. The flaking is crude and the material is local
cherts. They range in length from % inch to 2 inches. These small points are
similar to those found on the Laurentian sites of the Upper Ohio Valley
and Northeast. They are especially like those found at the Dixon and Rohr
shelters and at the Shriver site (Stewart and Dragoo, 1954). A close resem-
blance can be seen in these points to those from the Annis Mound in Ken-
tucky (Webb, 1950) and from the Faulkner site in Illinois (Cole, 1951,
p. 216). (See Fig. 29.)

Stemmed points. There are 51 points with stems. Most of these points
have a weak shoulder and short, parallel-sided, rounded base. They range
in length from \x/% inches to 3]4 inches. The larger of these points are long
and slender, while the smaller ones tend to be more broad in relation to
length. Chipping is generally crude except for a few specimens which were
finely executed. Many of these points have bases which are thinned. The
material is variable with local coarse-grained cherts predominating. The best
made and largest of these points are of Harrison County, Ind., flint. In
form many of these points are similar to those in the Archaic levels of the
Dixon and Rohr sequence. Not only do they resemble the Lamoka points
of New York (Ritchie, 1944, p. 299), but also similar points are found on
Laurentian sites in the Upper Ohio Valley and Northeast. (See Fig. 30.)

Expanded-stem points. There are 39 points with stems formed by cutting
crude side or corner notches into the point. These notches, however, are
not well defined as in the points assigned to the side-notched and corner-
notched categories. These points resemble the stemmed points both in the
crude chipping and the materials used. Some show basal thinning. They
range in length from l1/^ inches to 3 inches. These points are very similar
to those found on the Upper Ohio Valley Laurentian sites. (See Fig. 31.)

Large side-notched points. There are 19 side-notched points ranging in
length from 2 inches to 23A inches. These points are similar to the small
side-notched points except that they tend to be more slender, and finer
flint, such as that from Harrison County, Ind., was more commonly used.
(See Fig. 32.)

Corner-notched points. There are 22 corner-notched points ranging in
length from one inch to three inches. The notches in these points are deeply
cut, and the width of the point is wide relative to the length. The chipping
was finely executed and flint from Harrison County, Ind., and Flint Ridge,
Ohio, was most commonly used. Some local materials, however, were used
for a few of the specimens. Although corner-notched points do occur in the
Archaic, they are commonly associated with the later Woodland cultures.
(See Fig. 33.)

Miscellaneous points. This group was composed of a small number of
points from the McCain site which do not fit well into the above categories.
Among these points are two large side-notched blades with deeply cut
notches and wide, slightly rounded bases which have been ground smooth.
The cutting edges are beveled. There are also one large and two small side-

228

Annals of Carnegie Museum

vol. 35

Fig. 29. McCain site. Small side-notched points

Fig. 30. McCain site. Large stemmed points

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

229

Fig. 32. McCain site. Side-notched points and knives

Fig. 31. McCain site. Expanded-stem points

1

230

Annals of Carnegie Museum

vol. 35

notched points with deep notches and straight rectangular bases. There is
also one large, wide, corner-notched point with deep notches and serrated
edges. All of the above points are common to Middle Woodland sites. Their
distinctiveness at McCain suggested to me that a later component may be
present. (See Fig. 33.)

The remaining miscellaneous items included 2 lozenge-shaped blades, 1
knife-like blade, 3 large blades with shallow notches, 2 small triangular
points, and 18 broken points.

Drills.. There are 19 complete drills and 10 fragments of drills in the
McCain site collection. These drills were classified as follows. Ten expanded-
base drills made of varied materials, two drills with side-notched bases, one
straight drill, one expanded-center drill, and five crudely chipped drills
with triangular cross-section and expanded bases. The 10 fragments con-
sisted of tips and were unclassifiable as to type. All of these drills are
similar to those found in Archaic manifestations both in the Northeast
and Kentucky. (See Fig. 33.)

Scrapers. The most distinctive scrapers in the McCain collection are the
71 end-scrapers. The scrapers of this group are either stemmed, side notched,
or corner notched. Only one scraper is of the thumbnail type. Many of
these scrapers appear to be reworked points. The material varies from
crude local cherts to fine Harrison County, Ind., flint. These scrapers are
similar to those from Indian Knoll in Kentucky (Webb, 1946, p. 263) and
Faulker in Illinois (Cole, 1950, p. 216). (See Fig. 34.)

There are several side-scrapers of an unspecialized form. Some of the
crude blanks were also used as scrapers.

Blanks. There are many roughly percussion-chipped blanks of various
sizes in the McCain collection. Most of these show no secondary usage,
but a few specimens have chipped and battered edges.

The importance of the McCain site lies in its similarities to several Archaic
manifestations from widely separated areas of the Ohio Valley. The many
expanded-stem and side-notched points are particularly similar to those of
the -Laurentian sites of the Northeast and the Upper Ohio Valley. The
full-grooved axes, stemmed points and end-scrapers, however, are more closely
related to the Kentucky Archaic and Faulkner of Illinois. The flexed burials
and the quantities of shell in the village follow the pattern of the Kentucky
Archaic. Again in the McCain site as in the Annis Mound there is evidence
of an early Laurentian lithic tradition in the Ohio Valley.

In Indiana there are many sites on which Archaic materials are found.
The best known of these sites are the shell-middens along the Ohio River
and the lower Wabash River (Adams, 1949. Dragoo, 1955. Kellar, 1956).
Only surface collections are known for most of these sites. In general,
they are similar to the Kentucky Archaic sites, but many of the point types
from these sites are also similar to those of the Upper Ohio Valley and
northeast Laurentian sites. The chronological placement of most of these
sites within the Archaic is unknown. Black’s (1946) report on the Cato
site in Pike County, Inch, indicated the interaction of a late Archaic mani-
festation with Adena. The McKinley site near Noblesville, Inch, recently

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

231

Fig. 33. McCain site. Drills and corner-notched points

Fig. 34. McCain site. End-scrapers

232

Annals of Carnegie Museum

vol. 35

excavated by Downey Raibourn for the Indiana Historical Society, appears
to be of an earlier date.

Less well known than the large shell-middens, are many small Archaic
sites scattered over the entire state of Indiana. These small sites usually
do not contain shell debris, but their lithic inventory is similar to that
of the shell-midden sites. Many of these sites are located on terraces: which
are now at a considerable distance from the nearest streams. One of the
pressing needs of midwestern archeology is the thorough investigation of
these small Archaic sites which have often been ignored or mislabeled
“Woodland Campsite” in the reports.

The Archaic manifestations of Ohio are even less well known than those
of Indiana. Scant attention is given to the Archaic since so many productive
burial mounds are present to lure the archeologists and collectors. Morgan
(1952, p. 84) pointed out the presence in Ohio of materials similar to those
of Indian Knoll, and suggested that the area was occupied by related peoples
during the Archaic. Laurentian-like materials are also present in many of
the collections from northern Ohio (Smith 1954, p. 9-11). Surface collections
recently made by Charles Sofsky of Warren, Ohio, and John Zakucia of
Youngstown, Ohio, from sites in northeastern Ohio contain many Laurentian
items. Investigations are now underway at many of these sites and reports
of the findings should appear in the near future.

The materials from the Faulkner site in southern Illinois bear many re-
semblances not only to Laurentian but also to the Kentucky Archaic (Cole
and others, 1951). Many of the notched and expanded-stem points, drills,
and end-scrapers are similar to those of Laurentian sites in both the Upper
Ohio Valley and New York. The notched and stemmed scrapers and the
straight-stemmed points are similar to those of Indian Knoll (Webb, 1940)
and the McCain site (Miller, 1941).

The concave-based points of Faulkner are within the range of some Paleo-
Indian fluted points, and their specific resemblance to points of the Par-
rish site in Kentucky was noted by Cole (1951, p. 216). Again, as with both
Laurentian and the Kentucky Archaic, there are elements in Faulkner which
may have been derived from the earlier Paleo-Indian.

The work conducted by the Illinois State Museum at the Modoc Rock
Shelter in southern Illinois added much to our knowledge of the age of cer-
tain tools associated with the Archaic (Fowler and Winters, 1956). The
radiocarbon dates obtained from charcoal found at this site range from
B.C. 2665±300 years to B.C. 7922±392 years for Archaic deposits.

The major cultural manifestation at the Modoc Rock Shelter was Archaic
and was found from a depth of 5 feet down to 21.5 feet. The upper five
feet contained pottery that indicated middle to late Woodland occupation
of the site. Within the Archaic zone there are changes through time that in-
dicate the increasing complexity of the culture. Among the changes were:
1. a change from predominantly side-notched, slender projectile points to a
variety of forms that are predominantly stemmed, notched, and broad
bladed, and, 2. an increase in the number and variety of chopping and cut-
ting tools and polished-stone artifacts. Fowler and Winters (1956, p. 44)

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

233

tentatively set apart the lowest zone of the site (21.5 to 26.5 feet) from the
higher layers. They believed the one broad tapered-stemmed projectile point
found in this lowest zone was similar to those found in the lowest portions
of the Hidden Valley Shelter in Missouri (Adams, 1941). Also, these points
are similar to the Gypsum Cave points of Nevada (Harrington, 1933) that
were associated with extinct fauna and have been dated at a time contem-
porary with the lowest zone of the Modoc Rock Shelter.

The projectile points from the Archaic zone of the Modoc Rock Shelter
possess many resemblances to points of the Kentucky Archaic and of
Laurentian. The small side-notched points (Fowler and Winters, 1956, p. 18,
22) are especially similar to those of the Upper Ohio Valley Laurentian as
found in the Dixon and Rohr sequence. The choppers, side-scrapers, and
end-scrapers are also to be found in Laurentian manifestations in the
Upper Ohio Valley. The large stemmed points and notched end-scrapers of
Modoc are more at home in the Kentucky Archaic and at sites like McCain
in Indiana (Miller, 1941).

The radiocarbon dates for the Archaic at Modoc Rock Shelter push cul-
tures of this period back to at least 6200 B.C. Such a date is indicative of
the antiquity that can be assigned to other Archaic manifestations in the
Ohio Valley. The finding of Laurentian-like items at Modoc is again evidence
of the presence of an early basic Archaic culture below the Great Lakes
upon which Laurentian developed in the Upper Ohio Valley and the
Northeast.

Outside of the Ohio Valley and the Northeast, the relationships of
Laurentian with other Archaic manifestations of the eastern United States
are not well known. In the Carolina Piedmont, Coe (1952, p. 306) considered
the Badin culture as similar to the Indian Knoll culture of Kentucky. He
considered the correlation of traits so great that a direct connection between
the two groups was postulated. Since the Badin culture appeared in the
Piedmont about the same time the Kentucky Archaic was being displaced
by Adena, it is late in the Archaic sequence. Within the Badin culture are
some items such as projectile points, drills, and scrapers (Coe, 1952, fig. 163)
which are similar to Laurentian forms, but these similarities probably rep-
resent the basic underlying relationships of Kentucky Archaic and Lauren-
tian. The Early Archaic of the Piedmont, as manifested by the Guilford
Focus (Coe 1952, p. 304) appears to be unrelated either to the Badin culture,
Kentucky Archaic, or Laurentian.

In the lower Susquehanna Valley and in the Potomac Valley, the
Laurentian did not gain a major foothold. Some minor thrusts of Lauren-
tian into these areas did occur. A few Laurentian materials are present in
private collections from sites along the upper Potomac and the upper Susque-
hanna. The early Archaic manifestations of these areas are not well known.
Recent work by John Witthoft of the Pennsylvania State Museum indicated
the presence of an early chopping-tool assemblage which may have consider-
able antiquity. The late Archaic is best represented by the broad spearpoint
tradition and the use of steatite pots (Witthoft, 1953, p. 4-31). Contact of
the “Transitional Cultures” with Laurentian did occur on the eastern

234

Annals of Carnegie Museum

vol. 35

margins of the Upper Ohio Valley. Witthoft (1953, p. 25) suggested that
the beginnings of pottery in the East was perhaps derived from the soap-
stone vessel tradition of the “Transitional Cultures.”

Graham Cave (Logan, 1952), located in Logan County, Missouri, is a
stratified site that was occupied primarily during the Archaic, but earlier
Paleo-Indian materials were found in the lowest levels. A distinctive feature
of the site is the presence of a series of lanceolate projectile points. There
are two major varieties of these points. The first is characterized by short
channel flakes on one or both faces, by grinding on the sides which pro-
duced a slight expansion at the base, and by a concave base which has
slightly flaring “ears.” The second variety has a flat base and is propor-
tionately shorter and broader. While both varieties are more abundant in
the lowest level, they are in later levels. Logan (1952, p. 67-68) interpreting
this material as a fusion of cultural traditions, stated:

“The Archaic horizon of Central Missouri is not the product of one stream of
cultural influence. Traits from several different sources have fused to produce the
Archaic culture seen in Graham Cave. The Archaic horizon in level 6 is a series
of Eastern Archaic elements injected into a previously existing complex which has
the appearance of having had a western origin. This original complex included
the concave based points with flutes and basal grinding reminiscent of the Clovis
Fluted points. Since a Pleistocene or near Pleistocene antiquity for this level seems
improbable, there must have been an antecedent to Graham Cave elsewhere in the
region.”

Logan (1952, p. 68-69) also pointed out that “the Archaic complex of
Graham Cave especially from levels 4 and 5 on is essentially a complex
with an eastern orientation” and that “from level 5 up, Graham Cave is
comparable to the southeastern material.” Implied in Logan’s remarks was
the belief that a Paleo-Indian horizon was basic in this region and that
an Archaic complex from the East interacted with and finally completely
absorbed this Paleo-Indian tradition. Baerreis (1955, p. 7) presented an
alternative hypothesis stating that “a Paleo-Indian tradition was involved in
a pattern of acculturation with an already resident archaic tradition.” This
hypothesis not only has merit for the Graham Cave situation, but also for
other areas in the Eastern United States. Since the bulk of our information
pertains to sites with already well developed Archaic complexes, one of the
basic needs for the understanding of the origin and spread of the Archaic is
the discovery and definition of early Archaic manifestations.

The lower time limits and relationships of the Archaic with Paleo-Indian
cultures remain one of the most interesting problems in American archeol-
ogy. Direct contacts for Archaic and Paleo-Indian occurred in both the
Kentucky Archaic and at Graham Cave. The presence of scrapers and
other forms in Laurentian similar to those known for Paleo-Indian are
important evidence of early contacts. The small bifurcated points found in
the bottom levels of the Rohr Shelter and on many Archaic sites are sug-
gestive of points associated with the Pinto Basin Culture in the West. Many
of the crude choppers, scrapers, and gravers of the Archaic may be related
to an early lithic complex different from Folsom Paleo-Indian, but with
as great, if not greater, antiquity. Until extensive work over large areas is

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

235

concentrated on early sites, the beginnings of the Archaic will remain only
speculative.

SUMMARY AND CONCLUSIONS

The Archaic cultures of the Upper Ohio Valley and their relationships
to surrounding Archaic manifestations in the Eastern United States are
summarized as follows:

1. In the lowest levels of the Dixon and Rohr shelters was found an
early Archaic culture which had as its major tools small expanded-stem and
notched points, small end-scrapers, side-scrapers, utilized flakes, small bar
atlatl weights, crude hammerstones, and pitted stones. Many small open
sites, usually located on prominent hilltops near streams or springs such as
the Shriver site and the Wheeling College site, had a similar chipped-stone
complex plus a more extensive inventory of polished-stone tools such as
full-grooved axes and adzes.

The economy of the people living on these sites was based on hunting,
fishing, and gathering. Deer, elk, and certain small game were major items
on the menu. The presence of pitted and abraded stones on many of these
sites indicated that wild vegetal foods were also an important part of the
diet. In contrast to the lower Ohio Valley and Tennessee Valley Archaic
sites where shellfish were of great importance as a source of food, the
Upper Ohio Valley Archaic sites only rarely contain shellfish remains.

The rock shelter sites were probably used most extensively during the
winter months while the open sites were used during months of good
weather. No evidence of substantial dwellings of any type has been found
on any of the Upper Ohio Valley Archaic sites.

2. The material culture of these early Archaic sites has been compared
most often with the Lauren tian of the Northeast. The materials are similar
but important differences do exist. The early Upper Ohio Valley Archaic
sites contain items, such as grooved axes and bar atlatl weights, which appear
to be earlier than the adzes and winged atlatl weights associated with New
York Laurentian. The materials from the lowest levels of the Dixon and
Rohr sequence are representative of an early Archaic tradition which de-
veloped upon an early base similar to, and probably ancestral to, the
Laurentian of the Northeast. Tentative observation of the materials from
Archaic sites in the Upper Ohio Valley has indicated that the materials tend
to be progressively later from south to north with the more northerly Ohio
Valley sites most closely related to the Laurentian of New York.

Since the term Laurentian has been in the literature for several years, its
use has been extended in this article to cover related Archaic manifesta-
tions, some of which are much earlier than the New York Laurentian, in
areas surrounding the Northeast. Some of my colleagues may object to this
extension of Laurentian to cover these new manifestations; however, I be-
lieve at the present time the use of Laurentian or Laurentian-like to be
preferable to the addition of a new term to the literature.

3. The Upper Ohio Valley Laurentian manifestations are related to an
early Archaic culture which spread over much of the Ohio Valley and into
the Northeast. Many Laurentian elements are found in the early develop-

236

Annals of Carnegie Museum

vol. 35

ment of the Kentucky Archaic. At the Annis Mound and Indian Knoll of
the Kentucky Archaic, Laurentian-like points were found from the lowest
to the highest levels. Important basic elements in Laurentian are also
present in such sites as McCain in Indiana, Faulkner in Illinois, and in the
Old Copper complex of Wisconsin. Laurentian-like materials are present in
many surface collections from Archaic sites scattered throughout the Ohio
Valley, southern and northwestern Canada, and portions of the Mississippi
Valley.

4. The early basic Laurentian culture entered the Ohio Valley from
northwest of the Great Lakes, spread throughout the area below the Great
Lakes and into the Northeast while glaciers still covered much of central
Canada. With the glacial retreat, Laurentian spread northward into central
and eastern Canada. Some ancestral Laurentian groups were contemporane-
ous with late Paleo-Indian peoples and contacts between these two cultures
undoubtedly occurred. (See Fig. 35.)

5. Having established itself in the Ohio Valley and the Northeast, the
basic Laurentian culture underwent many local modifications. In the North-
east, plain and beveled adzes and slate tools became part of the complex.
The use of copper for tools was developed early by peoples possessing a
Laurentian-like culture in Wisconsin. In Kentucky and surrounding areas,
Laurentian elements were important early components in the formation of
the extensive shell-mound Archaic. Time, isolation, dissimilar natural en-
vironments, different contacts with other groups, and varied responses to
their problems by the people in the widely separated areas were important
factors in the development of distinctive complexes in late Archaic times.

6. The Upper Ohio Valley Laurentian sites during late Archaic times
had varied contents depending upon their nearness to other major Archaic
complexes in surrounding areas. The northern sites have traits, such as
plain adzes, gouges, and winged atlatl weights, which are common in the
New York Laurentian. The southern sites have traits, such as the full-
grooved ax and bar atlatl weights, which link these sites with the Kentucky
Archaic. The sites on the northern and eastern fringes of the Upper Ohio
Valley near the Susquehanna and Potomac drainages were greatly influenced
by the “Transitional Culture”. Found on these sites are large tapered-stem
blades, semi-lozenge points, and steatite vessel fragments. Studies now in
progress indicate that the “Transitional Culture” was more important on
the late Archaic scene in the Upper Ohio Valley than was previously sus-
pected.

7. Beginning perhaps during middle Archaic and extending into late
Archaic times in the Upper Ohio Valley proper, the “Panhandle Archaic”
developed. Among the most important tools of the “Panhandle Archaic”
are the grooved adz, stemmed point, lanceolate blade, and crescent banner-
stone. Outside of this area, the grooved adz is best known from late Archaic
sites in New England and along the east coast. What connections, if any,
existed between these widely separated areas remain unknown. The bone
items of the “Panhandle Archaic” are similar to those of Lamoka of New
York and also to Kentucky Archaic. The sites of this complex are the only

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

237

Fig. 35. Probable route of Laurentian culture into the eastern United States

238

Annals of Carnegie Museum

vol. 35

sites with shell-midden deposits in the Upper Ohio Valley. The peoples
responsible for the “Panhandle Archaic” were certainly in contact with
Laurentian culture and may have participated in the development of Early
Woodland cultures in the same area.

8. The Lamoka complex of New York presents questions difficult to
answer at this time. Its existence early in west-central New York is un-
deniable, but outside of that area the Lamoka as a total complex does not
seem to exist. Individual traits such as the beveled adz and occasionally the
typical narrow, stemmed points of Lamoka occur on sites in the Upper Ohio
Valley, but these sites always possess predominantly late Laurentian traits.
Some Lamoka-like points have been found in the Kentucky Archaic at Annis
Mound and at Indian Knoll. At both these sites the Lamoka points were
found only in the upper levels in contrast to the bottom-to-top distribution
of Laurentian points. Only continued research will tell us whether the
Lamoka is only a regional specialization of the basic Archaic culture which
was also ancestral to Laurentian in that area, or whether Lamoka had en-
tirely different origins.

9. The Adena peoples of the Early Woodland period seem to have
pushed Archaic peoples with Laurentian culture into less accessible areas
away from the Ohio River Valley proper. Within the Middle Woodland
complexes of the Upper Ohio Valley are a number of Laurentian traits, espe-
cially point types, which indicate that Laurentian influences persisted long
after the introduction of pottery and different economic pursuits.

CARBON-14 DATES FROM ROHR SHELTER
While this report was in press the following radiocarbon dates for the
Rohr Shelter were received from Dr. Edward S. Deevey, Yale University,
Geochronometric Laboratory, New Haven, Connecticut (Personal communi-
cation from Deevey to Dragoo, April 18, 1958):

Charcoal from Rohr Shelter (39°35' N. Lat., 79°50/ W. Long.), Monon-
galia County, West Virginia. The samples span the history of Archaic
culture at the site, where it is represented by the lower two-thirds of the
deposit, overlain by pottery bearing layers. The underlying yellow sand is
intrpreted as fluviatile, laid down when Bee Run, now many feet below the
shelter, was cutting the rock formation in which the shelter is found. Col-
lected in 1955 and submitted by Don W. Dragoo and Albert Bauer.

Y484. Rohr Shelter, Late Archaic 2200 ±60

From hearth in block A3, 18 inches below surface, sample 46 Mg
Cl, in transition zone from Late Archaic to Early Woodland.

Y485. Rohr Shelter, Middle Archaic 1940±70

From hearth in block A3, 24-26 inches below surface, sample 46 Mg
9-C2, about midway in Archaic deposit; the culture materials are
similar to those of Laurentian culture in New York state.

Y486. Rohr Shelter, Early Archaic 5310±90

From hearth in block A2, 32-36 inches below surface, sample 46 Mg
9-C3, just above sterile yellow sand; the artifacts are typologically
early Archaic.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

239

Comments: The difference between the ages of Y484 and Y485 is not sig-

nificant. The three samples give a satisfactory dating of Archaic culture in
the Upper Ohio Valley, from its inception through its transition to the
pottery making Early Woodland culture.

The difference between the ages of Y484 and Y485 may best be explained
on the basis of contamination with recent carbon from fine rootlets which
permeated all the samples from the shelter. As many of these rootlets as
possible were removed from the samples but because of their smallness
many still remained. The presence of more recent carbon in sample Y485
than in sample Y484 may account for the reversed order of these dates.
The presence of modern carbon in all the samples would tend to make
the dates more recent than perhaps they are in actuality.

The date of 5310±90 for the low level at the Rohr Shelter is very close
to the early dates obtained for the Archaic shell mounds along the Green
River in Kentucky and for the early Lamoka culture in New York state as
mentioned in the preceding section of this report. The 5310±90 date may
be interpreted as a conservative date for the presence of proto-Laurentian
culture in the Upper Ohio Valley and would indicate that Laurentian was
present in this area earlier than its appearance in the Northeast.

REFERENCES

Adams, Robert McCormick

1941. Archaeological investigations in Jefferson County, Missouri,

1939-40. Transactions of the Academy of Science of St. Louis,
v. 30, p. 151-221.

Adams, William R.

1949. Archaeological notes on Posey County, Indiana. Indiana His-
torical Bureau, 81 p. Indianapolis.

Anonymous

1943. First archaeological conferences on the Woodland pattern.

American Antiquity, v. 8, no. 4, p. 393-400, Menasha, Wis.

Baerreis, David A.

1955. The Archaic as seen from the southwestern periphery.

Mimeographed paper presented to Society for American Archaeol-
ogy conference, Bloomington, Ind., 1955.

Bailey, John H.

1939. A ground slate producing site near Vergennes, Vermont. Bulle-
tin of Champlain Valley Archaeological Society, v. 1, no. 2,
p. 1-29. Fort Ticonderoga, New York.

Bennett, Wendell C.

1948. A reappraisal of Peruvian archaeology. Memoirs of the Society
for American Archaeology, v. 8, no. 4, p. 1-7. Menasha, Wis.
1953. New world cultural history; South America. In “Anthropology
today,” University of Chicago, p. 211-225.

Black, Glenn A.

1946. The Cato site, Pike County, Indiana. Proceedings of the Indiana
Academy of Science, v. 55, Indianapolis.

240

Annals of Carnegie Museum

vol. 35

Butler, Mary

1936. Archaeological problems in Erie County, Pennsylvania. Penn-
sylvania Archaeologist, v. 6, no. 2, p. 27-30. Milton, Pa.

1939. Three archaeological sites in Somerset County, Pennsylvania.

Pennsylvania Historical Commission, Bulletin no. 753, p. 1-79.
Harrisburg.

Byers, Douglas

1954. Bull Brook— a fluted point site in Ipswich, Massachusetts.
American Antiquity, v. 29, no. 4, p. 343-351. Salt Lake City.

Carpenter, Edmond S.

1942. Archaeological reconnaissance in the Upper Ohio Valley. Penn-
sylvania Archaeologist, v. 12, no. 1, p. 20-23. Milton, Pa.

1951. Tumuli in southwestern Pennsylvania. American Antiquity,
v. 16, no. 4, p. 329-346. Menasha, Wis.

Chamberlain, Thomas C., and Rollin D. Salisbury

1921. Introductory geology. United States Geological Survey. Wash-
ington. (Water supply paper 234.)

Coe, Joffre

1952. The cultural sequence of the Carolina Piedmont. In “Archeol-
ogy of eastern United States,” p. 301-311. University of Chicago
Press, Chicago.

Cole, Fay-Cooper, and others

1951. Kincaid; a prehistoric Illinois metropolis. 385 p. University

of Chicago' Press, Chicago.

Collins, Henry B.

1937. Archaeology of St. Lawrence Island, Alaska. Smithsonian Mis-
cellaneous Collections, v. 96, no. 1, Washington.

Corliss, Frank R., Jr.

1951. The finding of an unusual cache at an Archaic site. Archeolog-
ical Newsletter, no. 4, p. 1. Carnegie Museum, Pittsburgh.

1954. Archeology in Garrett County, Maryland. Archeological News-
letter, no. 9, p. 6-12. Carnegie Museum, Pittsburgh.

Cross, Dorothy

1941. Archeology of New Jersey. Archeological Society of New Jersey,
v. 1, 271 p. Trenton.

deLaguna, Fredrica

1946. The importance of the Eskimo in Northeastern archeology. In
“Man in northeastern North America.” Papers of Robert S.
Peabody Foundation for Archeology, v. 3, p. 106-142. Andover.

Dragoo, Don W.

1953. Excavations at Watson site, Hancock County, West Virginia.
Bulletin Eastern States Archeological Federation, no. 13, p. 9.

1955a. Excavations at Johnston site, Indiana County, Pennsylvania.

Pennsylvania Archaeologist, Memoir no. 1, p. 1-56, Milton, Pa.

1955b. Archaeological survey of Gibson County, Indiana, p. 1-39. In-
diana Historical Bureau, Indianapolis.

1956. Excavations at the Watson site, 46Hk34, Hancock County, West

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

241

Virginia. Pennsylvania Archaeologist, v. 16, no. 2, p. 59-88.
Milton, Pa.

Fairbanks, Charles

1942. The taxonomic position of Stallings Island, Georgia. American
Antiquity, v. 7, no. 3, p. 223-231.

Fetzer, Elmer, and William J. Mayer-Oakes

1951. Excavation of an Adena burial mound at the Half-Moon site.
West Virginia Archeologist, no. 4, p. 1-25. Moundsville.

Fowler, Melvin L., and Howard Winters

1956. Modoc Rock Shelter. Illinois State Museum Report of Investi-
gations, no. 4, Springfield.

Griffin, James B.

1942. Adena pottery. American Antiquity, v. 7, no. 4 p. 334-358.
Menasha, Wis.

1943. Adena village site pottery from Fayette County, Kentucky. Uni-
versity of Kentucky Reports in Anthropology and Archaeology,
v. 5, no. 7, p. 666-672. Lexington.

1946. Cultural changes and continuity in eastern United States

archeology. In “Man in northeastern North America.” Papers
of Robert S. Peabody Foundation for Archaeology, v. 3, p. 37-95.
Andover, Mass.

1952. Culture periods in Eastern United States archeology. In “Arche-
ology of eastern United States,” p. 352-364. University of

Chicago Press, Chicago.

Hadlock, Wendell S., and Douglas Byers

1955. Radiocarbon dates from Ellsworth Falls, Maine. American An-
tiquity, v. 21, no. 4, p. 419-420. Salt Lake City.

Harp, Elmer, Jr.

1952. Cultural affinities of the Newfoundland Dorset Eskimo.

Ph.D. Dissertation at Harvard University, Cambridge, Mass.

Harrington, M. R.

1924. The Ozark Bluff-Dwellers. American Anthropologist, v. 26,

no. 1, p. 1-21. Menasha, Wis.

Hennen, Ray, and David B. Reger

1914. West Virginia Geological Survey— Preston County. Wheeling

News Litho. Co. Wheeling.

Hoffman, Bernard G.

1952. Implications of radiocarbon datings for the origin of the

Dorset culture. American Antiquity, v. 18, no. 1, p. 15-17.
Menasha, Wis.

Jenness, Diamond

1930. A new Eskimo culture in Hudson Bay. Geographical Review,
v. 15, p. 428-437. New York.

Kellar, James H.

1956. An archaeological survey of Spencer County, Indiana. Indiana
Historical Bureau, 68 p. Indianapolis.

242

Annals of Carnegie Museum

vol. 35

Kneberg, Madeline, and T. M. N. Lewis

1955. The Archaic in western Tennessee.

Mimeographed paper presented to Society for American Archaeol-
ogy conference, Bloomington, Ind., 1955.

Kroeber, Alfred

1944. Peruvian archeology in 1949. Viking Fund Publications in An-
thropology, no. 4, New York.

Larsen, Helge, and Froelich Rainey

1948. Ipiutak and the Arctic whale hunting culture. Anthropological
Papers of the American Museum of Natural History, v. 42,
p. 1-279. New York.

Lee, Thomas E.

1954. The first Sheguiandah expedition, Manitoulin Island, Ontario.
American Antiquity, v. 20, no. 2, p. 101-111. Salt Lake City.

1955a. The Archaic as seen from Ontario.

Mimeographed paper presented to Society for American Archaeol-
ogy conference, Bloomington, Ind., 1955.

1955b. The second Sheguiandah expedition, Manitoulin Island, On-
tario. American Antiquity, v. 21, no. 1, p. 63-71. Salt Lake City.

Leechman, Douglas

1943. Two new Cape Dorset sites. American Antiquity, v. 8, no. 4,
p. 363-375. Menasha, Wis.

Lewis, Clifford M.

1955. The Wheeling College site (460h22). The West Virginia Arche-
ologist, no. 7, p. 1-8. Moundsville.

Lewis, T. M. N., and Madeline Kneberg

1947. The Archaic horizon in western Tennessee. Tennessee Anthro-
pology Papers, no. 2, Knoxville.

Libby, Willard F.

1952. Chicago radiocarbon dates, III. Science, v. 116, p. 673-681.

Washington.

1954a. Chicago radiocarbon dates, IV. Science, v. 119, p. 135-140.

Washington.

1954b. Chicago radiocarbon dates, V. Science, v. 120, p. 735-742.

Washington.

Logan, Wilfred D.

1952. Graham Cave, an Archaic site in Montgomery County, Mis-

souri. Missouri Archaeological Society, Memoir no. 2, Columbia.

MacNeish, Richard S.

1952a. The archeology of the northeastern United States. In “Arche-
ology of eastern United States,” p. 46-58. University of Chicago
Press, Chicago.

1952b. A possible early site in Thunder Bay District, Ontario. Annual
Report of the National Museum of Canada for 1950-51, Bulle-
tin 126, p. 23-47. Ottawa.

1954. The Pointed Mountain site near Fort Liard, Northwest Ter-
ritories, Canada. American Antiquity, v. 19, no. 3, p. 234-253.
Salt Lake City.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

243

1956. Archaeological reconnaissance of the Delta of the Mackenzie
River and Yukon Coast. Annual Report of the National Mu-
seum of Canada for 1954-55. p. 46-81. Ottawa.

Martin, P., G. Quimby and D. Collier

1947. Indians before Columbus. University of Chicago Press. Chicago.

Martin, P., and John B. Rinaldo

1951. The southwestern co-tradition, Southwestern Journal of An-
thropology, v. 7, p. 215-229. Albuquerque.

Maxwell, M. S.

1951. Woodland cultures of southern Illinois; Archaeological excava-
tions in the Carbonclale area. Logan Museum Publications in
Anthropology, Bulletin 7, Beloit College, Beloit, Wis.

Mayer-Oakes, William J.

1953. An archeological survey of the proposed Shenango River
Reservoir area in Ohio and Pennsylvania. Annals of the Car-
negie Museum, v. 33, p. 115-124. Pittsburgh.

1955a. Prehistory of the Upper Ohio Valley. Annals of Carnegie Mu-
seum, v. 34, 296 p. Pittsburgh.

1955b. Excavations at Globe Hill Shell Heap. West Virginia Archeo-
logical Society, Publication Series no. 3, p. 1-32. Moundsville.

McCary, Ben C.

1951. A workshop site of early man in Dinwiddie County, Virginia.
American Antiquity, v. 17, no. 1, p. 9-17. Menasha, Wis.

McMichael, Edward V.

1956. An analysis of McKees Rocks Mound, Allegheny County,
Pennsylvania. Pennsylvania Archaeologist, v. 26, no. 3-4, p.
128-151. Milton, Pa.

Miles, Suzanne W.

1951. A revaluation of the old copper industry. American Antiquity,
v. 16, no. 3, p. 240-247. Salt Lake City.

Miller, Rex K.

1941. McCain site, Dubois County, Indiana. Indiana Historical So-
ciety, Prehistory Research Series, v. 2, no. I. p. 1-60. Indian-
apolis.

Morgan, Lewis H.

1877. Ancient society. 570 p. Kerr, New York.

Morgan, Richard G.

1952. Outline of culture in the Ohio region. In “Archeology of

eastern United States,” p. 85-98. University of Chicago Press,
Chicago.

Neumann, Georg K.

1952. Archeology and race in the American Indian. In “Archeology

of eastern United States.” p. 13-34. University of Chicago

Press, Chicago.

Phillips, Philip, and Gordon Willey

1953. Method and theory in American archeology. American Anthro-
pologist, v. 55, no. 5, p. 615-633.

244

Annals of Carnegie Museum

vol. 35

Popham, Robert E., and J. N. Emerson

1954. Manifestations of the old copper industry in Ontario. Penn-
sylvania Archaeologist, v. 24, no. 1, p. 3-19.

Quimby, G. I.

1953. Indians had a lake level problem 60 centuries ago. Chicago
Natural History Museum, Bulletin, v. 24, no. 6, Chicago.

1954. Cultural and natural areas before Kroeber. American Antiquity,
v. 19, no. 4, p. 317-331. Salt Lake City.

Ridley, Frank

1954. The Frank Bay site, Lake Nipissing, Ontario. American An-
tiquity, v. 20, no. 1, p. 40-50. Salt Lake City.

Ritchie, William A.

1932a. The Lamoka Lake site. Researches and Transactions of the
New York State Archeological Association, v. 7, no. 4, Rochester.

1932b. The Algonkin sequence in New York. American Anthropolo-
gist, v. 34, p. 406-414. Menasha, Wis.

1940. Two prehistoric village sites at Brewston, New York. Researches
and Transactions of the New York State Archeological Associa-
tion, v. 9, no. 1, Rochester.

1944. The Pre-Iroquoian Occupation of New York State. Rochester
Museum of Arts and Sciences, Memoir 1. Rochester.

1945. An early site in Cayuga County, New York. Researches and
Transactions of the New York State Archeological Association,
v. 10, no. 2, p. 1-158. Rochester.

1949. The Pre-Iroquoian pottery of New York State. American An-
tiquity, v. 25, no. 2, p. 97-124. Menasha.

1951a. Ground slate; Eskimo or Indian? Pennsylvania Archaeologist,
v. 21, no. 3-4, p. 46-52. Milton, Pa.

1951b. Radiocarbon dates on samples from New York State. In “Radio-
carbon Dating.” Memoirs of the Society for American Archaeol-
ogy, no. 8, p. 31-32. Salt Lake City.

1953. A probable Paleo-Indian site in Vermont. American Antiquity,
v. 18, no. 3, p. 249-258. Salt Lake City.

1955. The northeastern Archaic— a review.

Mimeographed paper presented to Society for American Archaeol-
ogy conference, Bloomington, Ind., 1955.

Ritzenthaler, Robert E., and Paul Scholz

1946. The Osceola site, an “Old Copper” site near Potesi, Wisconsin.
Wisconsin Archeologist, v. 27, no. 3, p. 53-70. Milwaukee.

Ritzenthaler, Robert E., and W. L. Wittry

1952. The Ocento site— an Old Copper manifestation. Wisconsin
Archeologist, v. 33, no. 4, p. 199-223. Milwaukee.

Roberts, Frank H. H., Jr.

1940. Development in the problem of the North American Paleo-
Indian. Smithsonian Miscellaneous Collections, v. 100, p. 51-116.
Washington.

1958

Dragoo: Archaic Hunters of the Upper Ohio Valley

245

Rodgers, Edward H., and Murry H. Rodgers

1950. Archaeological investigations of the region about Lakes Mistas-
sini and Albanel, Province of Quebec, 1948. American An-
tiquity, v. 15, p. 322-377. Menasha, Wis.

Rouse, Irving

1953a. The strategy of cultural history. In “Anthropology today,”

p. 57-76. University of Chicago Press, Chicago.

1953b. On the correlation of phases. American Anthropologist, v. 57,
no. 4, p. 713-722.

Rowley, Graham

1940. The Dorset culture of the eastern Arctic. American Anthro-
pologist, v. 42, p. 490-499. Menasha, Wis.

Schmitt, Karl

1952. Archeological chronology of the Middle Atlantic States. In

“Archeology of eastern United States,” p. 59-70. University of
Chicago Press, Chicago.

Sears, William H.

1948. What is the Archaic? American Antiquity, v. 14, no. 2, p.
122-124. Menasha, Wis.

1954. A late Archaic horizon on the Atlantic Coastal Plain. Southern
Indian Studies, v. 6, p. 28-36. Chapel Hill.

Smith, Arthur C.

1954. Onondaga chert in Northern Ohio. Ohio Archaeologist v. 4,
no. 4, p. 9-11. Columbus.

Snow, Charles E.

1948. Indian Knoll skeletons. University of Kentucky Reports in
Archeology and Anthropology, v. 4, no. 3, pt. 2, p. 367-554.
Lexington.

Spaulding, Albert C.

1946. Northeastern archaeology and general trends in the northern
forest zone. In “Man in northeastern North America.” Papers
of the Robert S. Peabody Foundation for Archaeology, v. 30, p.
143-167. Andover.

1955. Prehistoric cultural development in the eastern United States.
In “New interpretation of aboriginal American culture his-
tory.” p. 12-27, Anthropological Society of Washington.

Stewart, Paul R., and Don W. Dragoo

1954. The Gay Shriver Archaic site, Greene County, Pennsylvania.

The Pennsylvania Archaeologist, v. 24, p. 106-114. Milton, Pa.

Thompson A. C.

1954. The Boyer’s Run Rock Shelter— a preliminary report. Pennsyl-
vania Archaeologist, v. 24, no. 3-4, p. 115-126. Milton, Pa.

Webb, William S.

1946. Indian Knoll. University of Kentucky Reports in Archaeology
and Anthropology, v. 4, no. 3, pt. 11, p. 115-365. Lexington.

1950. The Carlson Annis Mound. University of Kentucky Reports in
Archaeology and Anthropology, v. 7 no. 5, p. 267-354. Lex-
ington.

246

Annals of Carnegie Museum

vol. 35

1951a. Radiocarbon dating on samples from the Southeast. In “Radio-
carbon dating.” Memoirs of the Society for American Archaeol-
ogy, no. 8, p. 30. Salt Lake City.

1951b. The Parrish Village site. University of Kentucky Reports in
Archaeology and Anthropology, v. 7, no. 6, p. 407-451. Lex-
ington.

Webb, William S., and David L. Dejarnette

1942. An archaeological survey of the Pickwick Basin in adjacent
portions of the states of Alabama, Mississippi, and Tennessee.
Bureau of American Ethnology Bulletin 29, Washington.

Webb, William S., and William G. Haag

1940. The Cypress Creek Village, site 11 and 12, McLean County,
Kentucky. University of Kentucky Reports in Archaeology and
Anthropology, v. 4, no. 2, Lexington.

Wettlaufer, Boyd

1955. The Mortlach site in the Besant Valley of central Saskatche-
wan. Department of Natural Resources, Regina, Saskatchewan.
(Anthropological Series no. 1.)

Willey, Gordon

1945. Horizon styles and pottery traditions in Peruvian archaeology.
American Antiquity, v. 11 p. 49-56. Menasha, Wis.

1953. Archeological theories and interpretations; New World. In
“Anthropology today,” p. 361-385. University of Chicago Press,
Chicago.

Willey, Gordon, and Philip Phillips

1955. Method and theory in American archeology, II. American
Anthropologist, v. 57, no. 4, p. 723-819. Menasha, Wis.

Wintenberg, W. J. *

1939. Eskimo sites of the Dorset culture in Newfoundland. American
Antiquity, v. 5, p. 83-102. Menasha, *Wis.

Witthoft, John

1952. A Paleo-Indian site in eastern Pennsylvania; an early hunting
culture. Proceedings of the American Philosophical Society,
v. 96, no. 4, p. 464-495. Lancaster.

1953. Broad spearpoints and the transitional period cultures. Penn-
sylvania Archaeologist, v. 23, no. 1, p. 4-31. Milton, Pa.

Wittry, Warren L.

1951. A preliminary study of the Old Copper Complex. The Wiscon-
sin Archeologist, v. 32, no. 1, p. 1-18, Milwaukee.

Wittry, Warren L., and Robert K. Ritzenthaler

1956. The Old Copper Complex; an Archaic manifestation in Wis-
consin. American Antiquity, v. 21, no. 3, p. 244-254. Salt Lake
City.

(AMES L. SWAUGER AND ARTHUR M. HAYES

' t>

ANTHROPOLOGICAL SERIES, NO. 4

Publication of this article has been sponsored by the Point Park
Committee of the Allegheny Conference on Community Development

Cover design by Clifford J. Morrow, Staff Illustrator

P4-P (bE

V,.*?

ART. 11. HISTORIC ARCHEOLOGY AT FORT PITT, 1953
(anthropological series, no. 4)

By James L. Swauger* and Arthur M. HayesJ

INTRODUCTION

During the period from January 15, through December 31, 1953, the Sec-
tion of Man, Carnegie Museum, conducted archeological salvage work at
the site of the Point State Park in Pittsburgh. Until the latter part of Febru-
ary, the work was done as a part of the Section’s program under the Fund for
the Study of Man, a project sponsored by the Sarah Mellon Scaife
Foundation.

A Service Purchase Agreement was executed between Carnegie Museum
and the Department of Forests and Waters of the Commonwealth of Penn-
sylvania, effective February 26, 1953, and from that time forward the
Section worked as part of the Commonwealth’s general Point State Park
program. Under the terms of this agreement, the Section was able to broaden
and intensify its activities.

A mimeographed interim report, “Archeological Salvage at the Site of
Fort Pitt, 1953,” dated September 15, 1953, was prepared by the authors for
the Pennsylvania Department of Forests and Waters. Copies of it were dis-
tributed to interested persons. The interim report was written at that time
because actual earth-moving operations at Point State Park, the phase of
operation promising most likely reward for the archeologists, had ceased. We
considered it desirable to inform the Department of gross results obtained
as of the date of the report, since we did not believe site activity through the
rest of the year would yield important results. This belief was borne out
by events.

This present report is based in large degree on the 1953 interim report;
in fact, some of the interim report material will be found here verbatim.
In a sense this present article is an expanded version of the interim report
with the addition of information not present or not understood in Septem-
ber 1953.

PURPOSE OF THE ARCHEOLOGICAL WORK

The Historical Advisory Committee of the Point Park Committee of the
Allegheny Conference on Community Development recommended to the
Department of Forests and Waters that, in conjunction with grading opera-
tions to be conducted at the site of the Point State Park, effort be made
to rescue whatever items of historical significance might come to light and
be useful in development of general Point State Park plans. This recom-
mendation was favorably received by the Department and was made part

* Assistant Director, Carnegie Museum.

tAssistant Archeologist, Section of Man, Carnegie Museum.

247

Submitted for publication, November 28, 1958
Issued June 18, 1959

248

Annals of Carnegie Museum

vol. 35

of the General and Specific Conditions of the interim Point State Park
contract.*

Three specific duties were the responsibility of the Section of Man under
this contract:

1. Collecting of all pre-1800 man-made objects.

2. Salvaging of bricks from walls that would be inaccessible later be-
cause of planned highway developments.

3. Recording of gross fort features not previously discovered by archeo-
logical methods, and any previously discovered for comparison with known
maps and plans of the fort complex in the Point State Park area.

A fourth responsibility not stated in the contract but naturally assumed
by the Section was the maintenance of proper records: the taking, number-
ing, listing, and initial storage of specimens and photographs; the preparing
of necessary maps and drawings; the keeping of journals; the organizing of
data for study and report; the keeping of strict financial records; and other
normal field and laboratory routine procedures.

PERSONNEL

The following persons were the staff under direct supervision of the
Section of Man during the Point State Park work. The specific Park duty of
each is given in capital letters:

James L. Swauger, archeologist, Curator, Section of Man, Carnegie
Museum.

Lawrence S. Thurman, historian. Curator, Old Economy, Ambridge, Pa.

Arthur M. Hayes, assistant archeologist. Section of Man, Carnegie
Museum.

Mrs. Dorothy E. Dragoo, clerical and laboratory assistant. Section of
Man, Carnegie Museum.

Swauger was active throughout the whole period. Thurman worked from
January 15 through May 17, when, in a letter to Swauger, he announced
cessation of regular Point State Park activity because of pressure of his
work at Old Economy and his poor health. Hayes worked from January 26
through June 9 when he left for summer field work with the Section’s regular
field crew as agreed at the time of his hiring. Mrs. Dragoo was on part-
time status from January 16 through August 19.

At various times services were hired from Frank Bryan, Inc., McKees
Rocks, Pa., for excavation work; Surveys, Inc., and Braun and Fulton, Pitts-
burgh, Pa., for surveying work; Pittsburgh Photographic Library, University
of Pittsburgh, for photographic work; Frederick R. Matson, Pennylvania
State University, University Park, Pa., for ceramic study; and Carnegie
Museum office staff for routine processing and final preparation of financial
records and the like.

* Contract for Development of Point State Park (clearing, grading, miscellaneous
work), Pittsburgh, Allegheny County, Pennsylvania. Project No. 1440-1. Common-
wealth of Pennsylvania, John S. Fine, Governor, Harrisburg, Pennsylvania. Depart-
ment of Property and Supplies, Alan D. Reynolds, Secretary. Ralph E. Griswold &
Associates, Landscape Architects, 206 Gladstone Road, Pittsburgh, Pennsylvania.
Nov. 18, 1952.

1959 Swauger and Hayes: Archeology at Fort Pitt, 1953 249

Our indebtedness to numerous persons who voluntarily assisted us in
one \vay or another should be acknowledged here. There were so many it
would be folly to list them all, but in particular Edwin V. Pugh, Wells-
ville, Ohio; Miss Rose Demorest and her staff of the Pennsylvania Room,
Carnegie Library of Pittsburgh; Ralph A. McGiffin of the Department of
City Planning, Pittsburgh; David W. Rial, Wilkinsburg, Pa.; and Dr. O. E.
Jennings and Dr. E. R. Eller of Carnegie Museum furnished valuable in-
formation.

A special note of praise must be sounded for the contractor’s men and the
Commonwealth’s inspectors at the Point State Park. Clauses in the contract
provide for co-operation of the excavator’s people with the archeologists. The
employees of the Frank Bryan company adhered not only to the letter but
also to the spirit of these clauses and even exceeded them in their aid to
the archeologists. C. L. Smith, Superintendent, Howard Mauk and Anthony
Vice, foremen, were tireless in making certain that all curious objects and
structures encountered were reported, in advising as to potential use of
items of their equipment, and in suggesting ways and means to further the
investigation. Not once was there any indication of impatience or lack of
either interest or warm good will. They went so far as to furnish many
services without charge and on their own initiative which made the arche-
ologists’ lot not only more pleasant but also more useful.

The same spirited interest was shown by the Commonwealth’s inspectors,
John H. Reish, District Supervisor, and the resident inspectors, James K.
Chambers and James K. Warren. It was a pleasure to work with these men.

METHOD

Time for research prior to commencement of excavations at Point State
Park was limited. On January 12, three days prior to the beginning of
excavation and more than a month before formal hiring, the Section was
reliably informed it would be retained to perform its salvage task. Most of the
Section’s preparation for its work went on while the job was in progress, but
the conditions under which the Section operated were such that this simul-
taneous preparation and action procedure was not a serious handicap.

Most of the library research was done by Hayes. Particularly during the
first three months of work there were many, many days on which the soggy
condition of the site’s soil prevented digging by the contractor, days on
which he could only break sidewalks or tear down stone walls. Advantage
of this situation was taken by Hayes to do library work as a result of
which he assembled a mass of information concerned primarily with physical
features of the fort complex at the site. Both Swauger and Thurman did
some library work. Miss Demorest and her staff were most helpful to us
in this study.

John A. Renner of Ralph E. Griswold 8c Associates supplied us with a
set of maps prepared in 1943 for the Point Park Commission of the City
of Pittsburgh by the Department of City Planning, with copies of the
Grading Plan and Salvage and Construction Plan for the interim Point

250

Annals of Carnegie Museum

vol. 35

State Park project, and with a copy of the Bliss report.* This report contains
a vast amount of material concerning the fort complex. It was compiled by
Wesley A. Bliss, archeologist for the Point Park Commission’s 1942-1943
investigations at the site of the Point State Park, is based largely on the
archeological work done there during the three years 1941 through 1943,
but also contains much other information. Without these maps and the
Bliss report, we would have had no reliable guides for our own planning,
work, and interpretation. The thoroughness of the Bliss report is evidenced
by the fact that every check we made of his record against actual fort condi-
tions as we found them proved accurate within reason. We added but little
to his compilation. Rial later donated a copy of the Bliss report to the
Section.

Through Pugh we met McGiffin. Bliss had lodged with McGiffin while
doing his work at the site of the Point State Park. McGiffin had intimate
knowledge of Bliss’s work not only because of his close association with Bliss
but also because much of the research on which the 1943 Department of
City Planning maps were based was his. He talked to us of Bliss’s work,
showed us objects recovered by Bliss, and permitted us to borrow bricks
from Bliss’s pit “B” with which we could compare those we would find.

Armed with the Bliss report, the City Planning maps, the interim project
maps, and the McGiffin samples, we proceeded with assurance in this other-
wise unfamiliar historical-archeological project.

On the site we depended on observation and movement for result. Hayes
worked each day from his hiring until he left. Thurman and Swauger worked
alternate days until Thurman’s withdrawal from active participation, and
Swauger full time from June 9 until the middle of July, and at odd inter-
vals thereafter. Most of the time two men, and sometimes three, were at the
site following shovels, bulldozers, high-lifts, rollers, motor cranes, and laborers
as they dug into and moved the earth. Aided by known provenience of
fort features and levels as determined by Bliss and shown on the City Plan-
ning maps, we were able to check against the grading and salvage and con-
struction plans to predict likely areas for important finds. We used the
grading plan grid system as our own grid for location of finds.

Although we knew most of the area from which soil was being removed
(roughly the space bounded by Duquesne Way, Barbeau Street, Penn Ave-
nue, and a north-south line slightly east of the Block House) was composed
of fill laid down about 1900, we hoped residue of the forts and of build-
ings possibly made of fort material might be found as part of this fill. This
hope also kept us attentive to digging done in areas outside the known
fort perimeter.

To expose a portion of Fort Pitt’s brick and stone wall along Liberty
Avenue (a wall now permanently covered by a ramp for the new highway
development) we hired a bulldozer, pumps, and motor crane with clamshell
*“ Bliss report” is the commonly used term for the unpublished “Part One of the
Report of the Point Park Commission,” Pittsburgh, Pa., December 31, 1943. It was
copyrighted in 1944 by the Point Park Commission. So far as the writers can dis-
cover, only 12 copies of this report were prepared. The copy donated to the Section
by Rial is copy 9.

1959

SWAUGER AND HAYES: ARCHEOLOGY AT FORT PlTT, 1953

251

bucket from the Bryan people. To assist in removing and transporting
bricks and stones from the three pits dug here, indicated as C.M. 1, C.M. 2,
and C.M. 3, we hired power hammers, trucks, and, again, a motor crane.
Fig. 1 shows location of test pits (C.M. 1, C.M. 2, C.M. 3, C.M. 4 and C.M. 5)
dug by Carnegie Museum, and also identifies the location of the earlier
pits dug by Bliss. All of the sample removal in C.M. 1 and most of that
in C.M. 3 was done by hand by Hayes and Swauger. To expose a portion of a
palisade north of Penn Avenue in C.M. 4, we hired a bulldozer and high-
lift for rough work, and finished off by hand. The filling of these pits was
done by bulldozers.

Although empowered to stop the excavator from working in any given area
for fear of his damaging significant materials, we had to use this power only
twice. Once north of Penn Avenue we stopped the motor crane when un-
certain as to the importance of a brick structure revealed when a sewer pit
was being dug. The structure proved of no value to us. Once we stopped
the power shovel when we feared it might cut through the palisade in
C.M. 4, and this stoppage proved wise.

Fort Pitt had five bastions. Names applied to these are of relatively re-
cent origin, but are accepted in literature concerning the fort. That pointing
northeast is the Music Bastion; that pointing southeast, the Grenadier
Bastion; that pointing south, the Flag Bastion; that pointing southwest, the
Monongahela Bastion; and that pointing northwest, the Ohio Bastion.

The important gross features encountered— the wall between the Flag and
Grenadier Bastions exposed in C.M. 1, C.M. 2, and C.M. 3 on the south
side of the fort, and the wall between the Music and Ohio Bastions, C.M. 4,
on the north side, were mapped by professional surveyors. Levels on these
walls were taken for us by Mauk, Grade Foreman for the Bryan company,
as well as by the surveyors. We, of course, also made our own maps and
drawings of these features and elements thereof.

We devoted considerable attention to the making of a photographic
record. All activities at the Point State Park excavation, not only of the
strictly archeological work but also of the general excavation and grading
procedures, were recorded in both black and white and in colored photo-
graphs. The Section took 196 black and white photographs and 349 colored
slides, the Pittsburgh Photographic Library took 42 black and white photo-
graphs. Rial volunteered his services for the production of a group of colored
slides of bricks, stones, mortar, and logs from the Fort Pitt walls, and a
series of fort complex maps and diagrams in black and white transparencies.

In the laboratory, activity centered on numbering, describing, and cata-
loguing specimens and photographs, mounting colored slides, and pre-
paring jackets for black and white photographs. Mrs. Dragoo did most of
this work after initial recording of essential data by Hayes and Swauger. For
recording of exact colors of Fort Pitt bricks, Mrs. Dragoo used Robert
Ridgway’s “Color Standards and Color Nomenclature,’’ although for prac-
ical purposes we follow Matson’s advice* and refer to bricks in their general
color range as medium red, deep red, brown, etc. In the laboratory also,

* Memorandum to Swauger, June 29, 1953.

252

Annals of Carnegie Museum

vol. 35

Fig. 1. Location of Fort Pitt, and positions of archeological test pits

1959

SWAUGER AND HaYESI ARCHEOLOGY AT FORT PlTT, 1953

253

were finished journals, maps, and drawings necessitated by the project,
and were prepared routine and special reports and letters. So far as maps
are concerned, ours are mostly of detailed portions of the fort structures
since the Bliss report maps are excellent and quite sufficient for general
overall presentation.

Swauger attended several conferences concerned with the work at Point
State Park; Thurman attended some; Hayes, none.

Daily journals were maintained in triplicate by each of the persons under
direction of the Section. The original copy is held by the Section, the first
carbon copy by the person preparing it, and the second carbon copy ac-
companies the bulk of the specimens in storage. The original was held in
the contractor’s office at the Point State Park for ready reference by inter-
ested persons throughout the project. Forms for these journals were pre-
pared by the Museum office staff and, after her hiring, Mrs. Dragoo. Forms
were also prepared for entry of photographic data and expense accounts.

Matson visited the site on June 29, made an examination of the bricks
removed as well as those on the site, and forwarded a general memorandum
to Swauger as of June 29. A representative collection of bricks and mortar
was forwarded to Matson on December 22. These he will use as samples
for laboratory ceramic study and as part of a collection of bricks from
known sites to be assembled at the Pennsylvania State Univerity.

Dr. Jennings made identifications of timbers removed from the north
wall in C.M. 4, and Dr. Eller performed the same service for glass and
bottles collected from the whole site.

RESULTS
Objects recovered

Accomplishment of the Section’s first responsibility in this work in Point
State Park, the rescuing of 18th century objects, was a dismal failure. So
far as we can determine, we recovered not one single object except pieces
of the actual fort wall which can confidently be given a pre-1800 date. We
can not explain this lack of specimens. It is not due to our carelessness or
lack of ability, for certainly we were diligent and equally certainly the
three men who worked are capable of differentiating between late 18th
century or even early 19th century objects on which some doubt might be
held and late 19th or early 20th century material. We saw nothing we
would hesitate to place chronologically as anywhere but in the late 19th or
early 20th centuries.

This conclusion as to date is readily understandable so far as most of the
material moved by the Bryan company is concerned. Before we began work
we knew the area had suffered flood, fire, construction disturbance, and
finally a fill about 1900 covering much of it to a depth of 8 to 12 feet. But
we certainly hoped to get something from work along the fort walls.

The fill was disappointing. The Section had hoped to be able to accumu-
late a collection from a stratified section of the fill which would give a
chronological story of the city as told by its rubbish. Even this hope was
dashed, for the fill was a homogeneous mass of all sorts of litter— broken

254

Annals of Carnegie Museum

vol. 35

crockery, glass, sewer pipes, machine-stitched shoes, charred lumber, piles
of ashes, slag, bricks, cement, tar paper, metal conduit, gas lines, etc.

The size of the tools used and the immense amounts of earth displaced
at one time were detrimental to locating specimens. For instance, much of
the earth was torn from place by a power shovel with a bucket having a
capacity of two cubic yards. The shovel took three bites to fill a truck
which carried the dirt to the new fill area and dumped it. A bulldozer flat-
tened the dumped earth. A sheeps’-foot roller rolled and tamped it. The
process consumed about ten minutes per load and posed a pretty problem
in speedy observation for the archeologists.

However, the face of the shovel made a clean horizontal cut on the
surface of the ground. The bucket dribbled from its mouth as it swung to
the truck. The truck’s dumping made a pile with the surface easily scanned,
and the edges free from packing which might hide objects. The bulldozer’s
blade packed in the center of its sweep but feathered dirt out at its edges
to reveal objects. The rollers churned the dump as they went bringing some
things to the surface even as they tamped others down. We saw thousands
of objects, and it is unlikely chance alone prevented our seeing 18th-century
material. A small collection of this late fill material was made— crockery
fragments, bottles, and metal objects.

The work along the fort walls was begun initially with heavy machinery
and was thus subject to the same handicap as work in the fill areas proper
so far as observation was concerned; but all work along these walls was
finished by hand and the opportunity to see was optimum. Here again we
found dozens of objects, and here again we found nothing we are confident
dates from the 18th century.

Bliss worked many more pits than clicl we, and did them largely by hand,
yet his experience with artifact recovery was no happier than ours. He was
no more successful than were we in dating as 18th century any pieces re-
covered that were not definitely part of the fort structure. The following
statements are from the Bliss report:

“The few artifacts recovered could not be definitely placed stratigraphically.
Crockery, pottery, shoes, bones, fragments of iron utensils, bricks, bottles, glass,
logs, stockade posts, and boat fragments made up most of the material that came
out of the test pits.” p. 73.

“Fragments of corckery (sic), bottles, pottery, shoes, bones, utensils, wire, nails,
etc. came out of the test pits.” p. 74.

With Bliss we must assume that there probably are relics of the 18th
century at the site of Fort Pitt because of literary reference to the finding
of such objects in the past (Bliss report, p. 75). All we are certain of is that
we found none.

The following categories and numbers of objects were recovered and

catalogued:

Bricks and brick fragments 357

Mortar and boxes of crushed mortar 53

Stone from fort walls 19

Boxes of rubble 4

Bone 1

1959

SWAUGER AND HAYES: ARCHEOLOGY AT FORT PlTT, 1953

255

Oyster shell 1

Metal fragments 5

Pottery and crockery fragments 20

Bottles and glass fragments 54

Stone paving blocks 5

Logs and bark and bark fragments 14

Cement block 1

Marble slab 1

Metal ball from Mt. Washington 1

Motorcycle license 1

Unidentified 4

Total 541

Of the 541 objects catalogued, 466 numbered items were taken to the
building of the Historical Society of Western Pennsylvania at 4338 Bigelow
Boulevard, Pittsburgh, on December 16, 1953. They remain there until the
construction of a museum building at Point State Park permits their display
in their area of origin, or until other use is found for them by the Historical
Advisory Committee. The other 75 items were retained by Carnegie Museum
as Accession 15661, in its permanent collection as a representative series of
objects from the 1953 work at Fort Pitt. Of these 75 items 22 were forwarded
to Matson on December 22, for his study. Six items were given as tokens of
appreciation to persons instrumental in the Fort Pitt operations. Record of
the distribution of numbered items is in the catalogue book.

Brick salvage

We were more successful in the task of salvaging bricks. Pits C.M. 1,
C.M. 2, and C.M. 3 were dug specifically to locate the fort walls to be
covered by a highway ramp paralleling Liberty Avenue to the north and
to remove from them all bricks that could be saved. As will be explained in
more detail in discussion of fort features, brick wall was found in C.M. 1
and C.M. 3; none in C.M. 2. All bricks taken from C.M. 1 and a few of
those from C.M. 3 were retained as samples. Most of the bricks from C.M. 3
were salvaged for later usage at Point State Park.

Initially we removed bricks by hand with hammers, chisels, and bars.
The bricks are so soft however, the mortar so hard, relatively, that the
pounding necessary to separate bricks from mortar shattered so many bricks
we estimate a loss of 50% in C.M. 1 from this process. Finding the same
situation recurring in C.M. 3, we decided to use power tools, deliberately
losing some rows of bricks in order to get out great chunks of brick wall.
This process was far more successful than the first tried, and we estimate
a loss not exceeding 30%. Several hundred good bricks in great blocks and
many loose ones were removed by use of power hammers, lifted by motor
crane, taken by truck to the northwest comer of the intersection of Penn
Avenue and Barbeau Street, and there buried by motor crane.

We salvaged by no means as many bricks as were desired to carry out
initial restoration plans for the Monongahela and Ohio Bastions as out-
lined for us at the beginning of the project. This is due in part to the

256

Annals of Carnegie Museum

vol. 35

difficulty of pulling good whole bricks from their seats in the walls, but
mostly to the fact that there are not nearly so many bricks left in the walls
as had been hoped. Salvage operations by 18th century contractors were
evidently very thorough,* and they didn’t leave many bricks for us to re-
cover. Those bricks we did salvage, however, will aid in the manufacture
of new bricks on the old model for the restoration work.

Some mention must be made of the bane of the archeologists’ existence-
souvenir hunters. From beginning to end of the excavation we were badgered
by spectators for pieces of brick and stone, mortar and wood, from the fort
walls. We did give away some valueless fragments of authentic material
and some whole bricks, chiefly to members of the Bryan company and other
persons who were working with us on the site. We also gave away, and
people took, modern bricks from the fill, fragments of cement from the
railroad retaining walls, chunks of slag and cinders, bags of earth from the
fill, and the like. Some of the more persistent souvenir seekers were ap-
parently quite satisfied with bricks from the sidewalk on Liberty Avenue,
bricks stamped plainly “Toronto” and ‘‘Pittsburgh Buffalo Co.”, which they
evidently mistook for Fort Pitt bricks. For future reference, only those
objects carrying the catalogue designation “PPP/number” on a square of
yellow paint, or those authenticated by a certificate signed by Hayes or
Swauger should be considered actual Fort Pitt material.

Fort features

General. Gross features of the fort were located in the four pits dug at
the request of the Section and numbered C.M. 1, C.M. 2, C.M. 3, and
C.M. 4. It is likely part of the wall between the Flag and Grenadier Bastions
was found in a hole dug to cap a water line under Liberty Avenue. This pit
we numbered C.M. 5. We did no work with C.M. 5 except to collect bricks
as samples and to photograph the wall since it was inexpedient to close
Liberty Avenue to traffic. Fig. 1 gives the positions of these features.

Test pit C.M. 1. In this pit, 38.35 feet of the fort wall were uncovered.
This portion of the fort’s eastern wall is the angle at the southwest corner
of the Grenadier Bastion and is composed of part of the southwest arm of
that bastion (We call this the eastern arm of the angle) and part of the
arm running more southerly into Liberty Avenue where it angles more or
less east into the Flag Bastion (We call this the western arm of the angle).
These arms are illustrated in Fig. 2. Both arms of the angle are interrupted
by a cement wall 14.43 feet from the corner along the eastern arm, 23.92 feet
along the western arm.

Only part of the original wall exists. In the corner of the angle were
perhaps two hundred whole bricks and many fragments. Most of the whole
bricks were removed for samples, those directly at the corner being left for
the benefit of future diggers. A section was taken through the stone body
of the wall to determine its form (Fig. 6) and a profile cut was made at
the foundation’s southern face to determine structure and depth. Stones
from the wall were removed for samples. Mortar samples were taken from
both the brick facing and the stone wall.

* Pennsylvania Archives, 1854, v. 10, p. 483.

1959

SWAUGER AND HAYES! ARCHEOLOGY AT FORT PlTT, 1953

257

CORNER OF ANGLE TQ

mark on curb- 44. o'

EL. FT., MIDWAY EASTERN ARM-8'

EL. PT., MID WAY WESTERN A RM - 12*

MARK TO MON., NE CORNER, INTERSECTION OF PENN
AND BARBEAU-495^ jjG-

1 ABOUT !"■ 20' SKETCH SCALE 1

GENERAL PLAN

PROFILE AREA

DIRECT COPY, FIELD SKETCH OF 15 APR., *53

Fig. 2. Location of test pit C.M. 1

258

Annals of Carnegie Museum

vol. 35

The wall averages 8 to 10 feet in width. It is made up of four “steps”. The
lowest of these we call “foundation”; the second, “footer”; the third and
fourth, “wall”. At the profile section the height of structure remaining is

4.92 feet, of which 2.92 feet is foundation; 0.5 foot, footer; 1.55 feet, wall.
Levels at the footer of the western arm average 718.67 feet above the sea-level
datum plane, comparing with Bliss’s levels averaging 718.4 feet for this part
of the fort wall.* The footer of the eastern arm is at an average elevation
of 717.52 feet, it being about one foot lower than the footer of the western
arm, and this compares favorably with the Bliss level given above. The high-
est elevation for the portion of wall remaining in C.M. 1 is 720.17 feet.

The foundation is made of eight layers of flat, roughly face-dressed stones
averaging 1.5 feet in width, and a little more than 0.25 foot in thickness,
with the exception of the top course which is carefully face-dressed stone,
with an average thickness of 0.33 foot. Some of these foundation stones are
mortared to each other; others are not. All had been laid with care and are
quite plumb. Excavations at the rear of the profile cut were not carried to
the base but revealed a vertical wall corresponding to that in the front but
much more carelessly laid, and not mortared.

A ditch, commonly called the moat, ran along the eastern side of the fort.
Bliss located it in his first test pit, “A,” and established its elevation.

In the moat area, measured vertically from the top course of the founda-
tion, we have 1.83 feet of clay and fine gravel to a layer of heavy gravel
and black muck which continues 1.42 feet to the base of the foundation.
This black muck and gravel layer has a superior elevation of 716.67 feet
and compares favorably with Bliss’s elevation of 716.6 feet for the floor of
the moat. The general relationships of this and other features of the wall
can be best understood by reference to Fig. 2, where a copy of the field
sketch of the profile is shown.

The footer is set back an average of 0.67 foot from the face of the
foundation, consists of two courses, one a shimming layer averaging 0.17
foot thick, the other the footer stones proper averaging 0.33 foot thick.
Each of these courses is carefully face dressed, and the footer stones were
laid with such precision that it is impossible to pry one loose without
levering out its neighbors to right and left at the same time. The footer
facing stones average 1.5 feet in length, and maintain their 0.33-foot thick-
ness for 0.75 foot back from their faces to where, from a definite transverse
ridge, they taper from the top back to an average thickness of 0.22 foot,
the bottoms remaining level. The footer stones proper average one foot in
width. Since the wall proper is set back almost two feet from the face
of the footer, thin stones laid along the sloping rear of the footer stones
carry their line back into the wall where it is lost in the stones of the wall.

The wall at the profile cut is made up of two steps. The first is set back

1.92 feet from the footer face, and averages 0.33 foot in thickness. It is a single
layer of carefully face-dressed stones. The second is set back 2.75 feet from
the footer face, and averages 1.25 feet in thickness. It is made up of four layers
of roughly face-dressed stones. Most of the wall stones average a little

* Bliss report, p. 76.

1959

SWAUGER AND HAYES: ARCHEOLOGY AT FORT PlTT, 1953

259

more than one foot in width, and are roughly rectangular in most instances,
but there is great variation in shape of those which are not rectangular.
These eccentric stones were undoubtedly used to fill gaps in the more
regular courses where errors of judgment as to size on the part of the
masons had to be rectified.

Corresponding to the superior elevation, of the foundation and 6.42 feet
back from its face is a peculiar set of two “steps” completely incased in
other wall stones. The lower of these is 0.83 foot thick and is made up of
four courses of stone. The higher is 0.46 foot thick and is composed of two
layers of stones. All stones used average 0.75 foot in width. All are care-
fully face dressed and are tightly mortared together. The presence of this
subsidiary and “hidden” wail suggests that after the foundation had been
laid a mortared wall was built up corresponding to the front vertical wall
face and more than half way to the rear face. We can conjecture no specific
purpose behind this method of construction.

Most of the wall stones are carefully mortared for two rows back of the
interior wall discussed above. The final two or three courses, however, are
carelessly set in mud packing. This is as true of the highest remaining
layers of the wall as of the rear of the footer and foundation.

We were impressed by the exceeding hardness of the mortar used and
by the great quantity used. In many instances it was easier to chisel stone
away from mortar than mortar away from stone while the profile cut was
being made. In many areas of the profile cut, we found the mortar was
thicker than the stones it held together. Almost it looks as if the masons had
poured great batches of mortar over a layer of stones and set the next
course on the mortar to sink to position by weight alone.

The bricks were used as facing for the stone wall. In test pit C.M. 1 the
bricks had been worked at so much in times past that it is now difficult to
determine the number of bricks used in each course either horizontally or
vertically. In test pit C.M. 3 the number was more readily apparent and
a full discussion of the bricks will be given in the section devoted to
that pit.

A total of 111 bricks and brick fragments, 50 mortar samples, 18 wall
stone samples, and some miscellaneous boxes of rubble, bone, shell, glass,
and metal pieces were removed from C.M. 1. We can not say that anything
but the bricks, mortar, and stones are 18th century.

Test pit CM. 2. Less than 10 feet of the arm running south into Liberty
Avenue were uncovered. The position is shown in Fig. 3. It is only foundation,
footer, and very little wall, all corresponding to those elements in C.M. 1.
No bricks are present. Since at the time C.M. 2 was opened we were search-
ing primarily for bricks, the portion here had no bricks and could tell us
nothing we had not learned in C.M. 1, and by contract we were expected
to get on with our work as rapidly as possible, we did no work at this
portion of the wall once its general character had been ascertained.

This decision was productive of good results, for Smith, with Renner’s
approval, diverted into C.M. 2 a sewer line planned to run directly through

260

Annals of Carnegie Museum

vol. 35

Fig. 3. Location of test pits C.M. 1, C.M. 2 and C.M. 3

1959

SWAUGER AND HAYES: ARCHEOLOGY AT FORT PlTT, 1953

261

Fig. 4. Location of test pit C.M. 4

262

Annals of Carnegie Museum

vol. 35

C.M. 3 at almost footer depth, and we were thus free to work in C.M. 3
without interference.

Test pit C.M. 3. In this pit, 33.92 feet of the arm coming south from
C.M. 1 and C.M. 2 were uncovered. The position is shown in Fig. 3. The ex-
cavation for this pit ran into Bliss’s pit “B” and included about ten feet
of the wall he found.

Fig. 5. Portion of wall of Fort Pitt between Grenadier and Flag Bastions.

Test pit C.M. 1

Here is a section of wall consisting of 1.83 feet of foundation with a top
elevation of 718.37 feet, comparing favorably with Bliss’s elevation for
this element,* and a base elevation of 716.54 feet where it is bedded in
gravel below a layer of muck. Above the footer, which is of the same character
as in C.M. 1, is a brick wall laid in English bond, 2.17 feet high with a
backward slope of 1.02 inches in this height. There is then a stone wall
2.67 feet high above the brick wall. From the face of the foundation to the
rear of the stone wall the average distance is 9.67 feet. Fig. 7 gives a good
general view of test pit C.M. 3.

The stone wall is built in a series of four steps above the footer. Bricks
are in place on the footer, as shown in Fig. 8, and for a height of a little

* Bliss report, p. 76.

1959 Swauger and Hayes: Archeology at Fort Pitt, 1953 263

over two feet above it, but it is probable they once faced the whole wall.
The bricks are only four layers wide on the footer but due to their “reverse
step” method of laying, they run eight to ten layers wide two feet above it.
They tie into the stone wall not only by means of the “reverse steps,” but
also by heavy mortar bedding. In one or two places, stones for the wall above
a layer of bricks extend over the layer and are mortared to them. This is
a random occurence, however.

Fig. 6. Section through wall of Fort Pitt between Grenadier and
Flag Bastions. Test pit C.M. 1

Bricks average 8.5 inches long, 3.5 inches wide, and 2.25 inches thick.
Exceptions to this standard occur, particularly among those bricks resting
directly on the footer, and usually the eccentricity is in length, some bricks
being as much as 1 1 inches long. A few bricks, too, are not truly rectangular
on all faces but are chopped and hewed to different angles and curves to fit
special niches in the walls where bricks of standard size and shape were not
efficient. A representative two-square-foot section of the wall shows in its
face 12 stretchers and 24 headers. Fig. 9 gives a good view of the brick wall.

The bricks, and here we derive most of our information from Matson’s
memorandum of June 29, are of a red burning clay with a light intermixture
of river sand or a clay rich in fine-grained sand. Some of the brick are
dense, heavy for their size, others almot punky in texture, light for their

264

Annals of Carnegie Museum

vol. 35

size. Probably these discrepancies are due to positions of bricks in the kiln.
In color they fall generally into the deep and medium red classes (We here
follow Matson in describing in a few general visual categories rather than
in the specific categories used by Mrs. Dragoo in preparing her catalogue
descriptions) with some pale orange, and some brown bricks. Some are
burned black around the edges. Some have vitreous glazing over portions
of them.

Fig. 7. Brick facing on wall of Fort Pitt between Grenadier and
Flag Bastions. Test pit C.M. 3

The bricks were made in a mold, probably one of wood. This mold
formed three sides and both ends of the bricks, the upper broad surface
being formed by tamping with fingers or palms of hand. There usually
is a little overhang on the upper surfaces, and in broken bricks, the irregular
upper surfaces are clearly seen.

The mortar is light brown, a clayey, sometimes muddy looking mixture,
with considerable lime washing and lumps included. Some brick lumps are
found in the mortar but Matson does not believe they were included as
grog. There is a great deal of variation in the amount of mortar used in any
two groups of bricks, and an even greater variation among groups of stones
in the wall.

1959

SWAUGER AND HAYES I ARCHEOLOGY AT FORT PlTT, 1953

265

Construction of the wall, details of the foundation and footer, and stratig-
raphy in the moat area are the same as in C.M. 1.

The situation in regard to specimens is also the same as that in C.M. 1;
that is, only objects pulled from the wall can be considered 18th century
material. Altogether, 216 bricks and brick fragments, two mortar samples,

Fig. 8. Detail of brick facing at footer of wall of Fort Pitt between
Grenadier and Flag Bastions. Test pit C.M. 3

266

Annals of Carnegie Museum

vol. 35

and one stone sample were removed from C.M. 3 and kept as specimens.
Many other bricks and some stone, as noted earlier, were buried.

Test pit C.M. 4. As illustrated in Fig. 5, 175 feet of a stockade line paral-
leling the wall between the Music and Ohio Bastions were exposed. The po-
sition of this line as surveyed by Braun & Fulton does not correspond exactly
to features given in Sheet A2 of 7, “Forts and Fort Features . . .”, of the
Department of City Planning maps. It is our opinion that it represents the

Fig. 9. Detail of brick facing of wall of Fort Pitt between Grenadier and
Flag Bastions. Test pit C.M. 3

interior line of the earth wall structure between the Music and Ohio
Bastions.

The portion of the fort revealed by excavation at C.M. 4 consists chiefly
of a stockade line 147 feet long running almost due east and west, its
eastern end lying 356.26 feet west of the monument at the northeast corner
of the intersection of Penn Avenue with Barbeau Street, and 101.94 north
of the 7-foot running line on the north Penn Avenue sidewalk. Its western
end is 494.8 feet west of the monument, 146.06 feet north of the running
line. It is interrupted in several places by gas lines, by intrusive buildings,
and by blank areas which have suffered disturbance but of exactly what
type we did not determine. In general it is whole, and, as illustrated in
Fig. 10, it is quite plain.

1959

SWAUGER AND HAYES! ARCHEOLOGY AT FORT PlTT, 1953

267

At the eastern end of the main part of the stockade, 15.6 feet of stockade
incline almost due north. It is a clearly defined section but is interrupted
by a fill at its northern end. We excavated across the line of its direction
north of its last post, but although we went to a cement railroad retaining
wall, we had no success.

At the western end of the main part of the stockade, about 12.5 feet
of post line is barely discernible inclining to the south. We traced it by

Fig. 10. Top of line of stockade posts, north wall of Fort Pitt.

Test pit C.M. 4

means of organic smudges, remnants of posts, as illustrated in Fig. 11, and
one fairly well preserved post top, this latter the most southerly element
found. Excavation out its line of departure met with no success.

Elevations taken for us by Mauk average 720 for the tops of the stockade
posts. We doubt that these levels are significant so far as the wall itself is
concerned, but they may be in establishing the height of the parade ground,
for obviously the posts had been chopped off almost at ground level at
some time in the past probably so buildings could be constructed with
either foundations or even first floors at this 720-foot level.

The stockade, as shown in Fig. 12, is composed of two types of posts,
large ones averaging 0.83 to 1 foot in diameter, with six feet of their length
well preserved and in place, and smaller ones averaging 0.33 to 0.5 foot in

268

Annals of Carnegie Museum

vol. 35

Fig. 11. Smudges and remnants of west portion of line of stockade posts,
north wall of Fort Pitt. Test pit C.M. 4

diameter, with an average length of three feet well preserved and in place.
The large posts had not been sharpened, had not been driven into position
in the wall by mauls, but had been stood in a ditch dug to receive them
as indicated in Fig. 4. Earth had then been tamped in around them. At the
one vertical section cut, a short horizontal log was still in place lying snug
against the northern face of the post line and covered with tamped earth.

1959

SWAUGER AND 1 1 AYES: ARCHEOLOGY AT FORT Pi I T, 1953

269

The shorter posts had been sharpened and driven into the filled ditch to
strenghen the post line and to fill interstices between the large posts. While
it is difficult to to be certain of the number of posts remaining, since the
entire stockade line was not completely cleaned and profiled, 159 large, and
small post tops were counted in the wall, and it is likely the exact total is
not significantly different. Logs brought to the museum as samples were
identified by Dr. Jennings as elm and black oak.

Fig. 12. Profile cut revealing stockade posts in place, north wall of Fort Pitt.

Test pit C.M. 4

The superior level of a representative post at the vertical section is
723.47 feet. Subtracting the post’s length, six feet, from this figure, gives a
level for the base of this particular post, and essentially a practical level
for the posts in the section, of 717.47 feet, very nearly the elevation for the
footing of the stone wall.# This is, of course, slim evidence, but it may in-
dicate an effort to have the bases of all walls at nearly the same level.

East of the stockade line and to all practical purposes on line with it at a
distance of 10 feet is a burned stump chopped off at a level corresponding
to the tops of the stockade posts; beyond that 3.83 feet a square-cut post, that
we do not believe is contemporary with the fort, chopped off at the same
level; and beyond that 8.17 feet another burned stump chopped off as was
* Bliss report, p. 76.

270

Annals of Carnegie Museum

vol. 35

the first. Ten feet beyond this stump, 32 feet from the stockade line, and
still on line with it, were found three sharpened stakes which were recovered.
These average a little over three feet in length, and, as identified by Dr.
Jennings, are hickory, red oak, and beech. It has been conjectured they were
cheveaux de frise, but since they are not robust, they may have been impaling
stakes for sinking in the moat with their points up, or they may have been
only stakes used in stretching lines.

From test pit C.M. 4, six bricks and brick fragments, three pieces of
crockery, and 13 pieces of wood and bark were removed as samples. The
wood and bark we are certain are 18th century. The bricks and crockery
we feel sure are 19th or early 20th.

Test pit C.M. 5 was not dug at the request of the Section of Man but is
numbered and mapped. Its potentials were discussed, earlier, and its position
is shown in Fig. 1.

The pump. The map Sheet A2 of 7 mentioned on page 266 shows a
pump about sixty feet south of the west end of the stockade line. At approxi-
mately this position normal grading operations uncovered a tub-like struc-
ture with its exterior made of vertical staves like those of a barrel, and its
interior made of alternating circles of bricks and curved pieces of wood.
Digging in this structure revealed only a mass of fine, dark dust shot with
silvery specks of metal. The bricks were inspected by Matson during his
visit on June 29, and he agreed with our conclusion that they do not resemble
known fort bricks. The contractor was told the element was not of con-
suming interest to us, and it was covered as part of the regular grading work.
Since that time, however, we have seen an illustration* of a 17th century
well barrel at Jamestown, Va., which to some extent resembles the structure
described above except for the interior lining. It is not impossible that
this may have been part of the catch basin for the pump indicated on the
“Forts and Fort Features . . .” plan although it is not probable.

RECORDS

Note-books, journals, correspondence, etc.

Considered as an integral part of the collection of material made during
work at the Point State Park are the records maintained. These are briefly
discussed here with notice as to their locations, copies, and the like, for the
benefit of future workers at the Fort Pitt site.

Journals were kept by Hayes, Thurman, and Swauger. The books them-
selves remain in the possession of those who made them, but typed copies
are held by the Section, as well as a copy of the journal prepared by Mrs.
Dragoo. Duplicate copies of these journals accompany the specimens at
present in the building of the Historical Society of Western Pennsylvania.
They will move with the objects to the Point State Park museum. In such
records, of course, the Bliss report must be given a place. It is history of
the Point area, of forts such as was Fort Pitt, specific discussion of specific
items, and discussion of specific archeological activities.

*Fig. 195, illustration for Jean C. Harrington, "Historic Site Archeology in the
United States.” In "Archeology of Eastern United States,” James B. Griffin, ed.,
University of Chicago Press, Chicago, 1952.

1959

SWAUGER AND HAYES: ARCHEOLOGY AT FORT PlTT, 1953

271

Swauger’s field note-book, a mining transit book, remains in the Section’s
possession. It contains only a few items of textual information not con-
tained in the journal but does contain sketches which may be of future use
including, for instance, one showing the location of the burial pit for
bricks and stones from C.M. 3.

Copies of all reports made to the director of the museum or to various
officers of the Plistorical Committee are kept by the Section. These include
progress reports, estimates of the situation as the excavations developed, and
the like. The Matson memorandum of June 29 is held with these reports.

A detailed descriptive catalogue of the objects recovered, giving assigned
number, origin within the excavated area, full description, and disposition
of each object was drawn up by Hayes, Mrs. Dragoo, and Swauger. The
original is in two volumes. Typed copies were also made. One of these is
held by the Section; one accompanies the specimens.

Copies of correspondence are kept in the Point State Park record books
by the Section. Used in concert with the journals, these copies provide an
excellent listing of persons concerned with the project either officially or
with only a general interest.

The Section has kept a fairly full publicity record of its activities at the
Point State Park. At present the clippings which compose this record are
kept in the Point State Park record books but in time they will be transferred
to the Section clipping file where they will always be available for study.

Maps and drawings

The following groups of maps and drawings are held by the Section as
part of the Point State Park records. The first seven are those prepared by
the Department of City Planning of the City of Pittsburgh for the city’s
Point Park Commission in 1943. Their scale is 1"=50'.

1. Archaeological Excavations of 1942-1943, Location of Test Pits and
Fort Features. Sheet A-l of 7.

2. Forts and Fort Features of 1754 to 1792 Superimposed upon a 1942
Map of Point Area. Sheet A-2 of 7.

3. Undisturbed Sub-surface Relevant to Surface and Sub-surface Fort
Features Showing Basement Areas. Sheet A-3 of 7.

4. Physiographic Changes from 1754 to 1942 Relevant to Surface and
Sub-surface Fort Features Superimposed upon a 1942 Map of Point Area.
Sheet A-4 of 7.

5. Sub-surface Utilities Relevant to Surface and Sub-surface Fort Features
Superimposed upon a 1942 Map of Point Area. Sheet A-5 of 7.

6. Owners of Record and Character of Buildings as of 1940. Sheet A-6 of 7.

7. Land and Buildings Assessed Values as of 1940. Sheet A-7 of 7.

The next four maps were prepared by Ralph E. Griswold & Associates
in 1952.

8. Plot Plan, Point State Park. Scale l"=50r.

9. Survey Diagram, Point Park. Scale 1"=100'.

10. Grading Plan, Point State Park. Scale V' — bO'.

11. Salvage and Construction Plan, Point State Park. Scale 1"=50'.

272

Annals of Carnegie Museum

vol. 35

The surveyors hired by the Section produced these maps:

12. Carnegie Museum Pit 1, Point Park project. Rudolph Agresti. March.
1953. Scale l"-5'.

13. Uncovered Portion of Stockade of Original Fort Pitt in Point Park.
Braun and Fulton. July, 1953. Scale 1" = 50'.

Swauger prepared two drawings labeled “Sketch Maps” on tracing paper
for overlay use during the excavation work. Their scale is 1"=50'. They are:

14. Sketch Map no. 1, Excavations and Results, Bliss Test Pits. February

21. 1953.

15. Sketch Map no. 2, Probable Excavation Area, P.P.P. February 22, 1953.
Swauger and Flayes prepared a series of drawings labeled “Charts” for use

in both planning and record procedures. Their scales vary from chart to
chart and in instances within charts for either horizontal or vertical read-
ings. In such instances, the first scale given is horizontal, the second, vertical.
In the listing given here, Swauger’s charts are identified by the initials
“J.L.S.”; Hayes’s by “A.M.H.”.

16. Chart 1. Point Park Elevation Chart, No. 1: “O” line. Feb. 4., 1953.
J.L.S. Scale 1"=50'; 0.2"= 1'.

17. Chart 2. Point Park Elevation Chart, No. 2: “N-l” line. Feb. 4, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

18. Chart 3. Point Park Elevation Chart, No. 3: “N-2” line. Feb. 7, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

19. Chart 4. Point Park Elevation Chart, No. 4: “N-3” line. Feb. 7, 1953.
J.L.S. Scale 1"=50'; 0.2"= 1'.

20. Chart 5. Point Park Elevation Chart, No. 5: “N-4” line. Feb. 9, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

21. Chart 6. Point Park Elevation Chart, No. 6: “S-l” line. Feb. 9, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

22. Chart 7. Point Park Elevation Chart, No. 7: “S-2” line. Feb. 9, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

23. Chart 8. Point Park Elevation Chart, No. 8: “S-3” line. Feb. 9, 1953.
J.L.S. Scale 1" = 50'; 0.2"= T.

24. Chart 9. Point Park Elevation Chart, No. 9: “S-4” line. Feb. 9, 1953.
J.L.S. Scale 1" = 50'; 0.2"= V.

25. Chart 10. Stratigraphic Charts, Wesley Bliss test pits. Feb. 21, 1953.
J.L.S. No horizontal scale; vertical scale, 0.2"= T.

26. Chart 11. Stratification, Bliss test pits, Land Wall. Feb. 22, 1953.
J.L.S. Scale 1"=50'; 0.2"= 1'.

27. Chart 12. Stratigraphic Charts, Contours and Pits, S6-S7 Contours,
Pits “A” and “B.” Feb. 24, 1953. J.L.S. Scale 1"=50'; 0.2"= T.

27. Chart 13. Stratification and Horizontal Plan, Land Wall. Feb. 24,
1953. J.L.S. Scale 1"=50'; 0.2"= T.

28. Chart 14. Stratification in Pits, Southeast Wall. March 14, 1953.
J.L.S. Scale 1"=50'; 0.2"= T.

29. Chart 15. Pit Stratification, C.M. 1, Bliss pits “A” and “B.” March

22, 1953. J.L.S. No horizontal scale; vertical scale, 1"=1'.

1959

SWAUGER AND HAYES! ARCHEOLOGY AT FORT PlTT, 1953

273

30. Chart 16. Archeological Knowledge, Southeast and East Walls of
Fort Pitt. (Prepared for presentation with report of April 7, 1953, to Gris-
wold from Swauger.) April 7, 1953. J.L.S. Scale 1"=50'; 0.2"= T.

31. Chart 17. Point Park Project, Miscellaneous Figures, Carnegie Mu-
seum pit 1. March 22, 1953. J.L.S. Varying scales.

32. Chart 18. Point Park Project, Miscellaneous Figures, Carnegie Mu-
seum pit 1. March 22, 1953. J.L.S. Varying scales.

33. Chart 19. First, second and third sections, Retaining wall, C.M. 1.
March 25, 1953. A.M.H. 0.1"= 1".

34. Chart 20. Various views, Wall in C.M. 3. A.M.H. April 10, 1953.
1"=2'.

35. Chart 21. Fourth and fifth sections, Retaining wall, C.M. 1, April,
1953. A.M.H. 1"= V.

36. Chart 22. Various views, Wall in C.M. 3. A.M.H. April 22 to May 1,
1953. A.M.H. Varying scales.

Photographic records

There are three categories in the photographic record:

1. Kodachromes taken chiefly by Swauger and Rial.

2. Black and white photographs taken chiefly by Swauger.

3. Black and white photographs taken by the Pittsburgh Photographic
Library.

Kodachromes, negatives and prints of Swauger’s black and white photo-
graphs are in the possession of the Section, held as part of the Point State
Park record, and are entered and numbered in routine fashion in the
Section’s photographic file. Prints of the Pittsburgh Photographic Library’s
black and white photographs are held by the Section; the negatives are re-
tained by the Pittsburgh Photographic Library, where prints can be obtained.
The prints from the Pittsburgh Photographic Library bear that Library’s
numbers.

Not all the photographs taken by Swauger (and his pictures form the bulk
of the photographic record) are good photographs. However many that are
blurred in places through being out of focus or because of movement are
retained because portions of the slide or print retain sharp impressions of
part of the work. They are part of the listing below which gives the categories
of subjects, the kind of picture, the taking agency, and the proper reference
numbers in the files of the taking agencies. A complete descriptive list,
slide by slide, print by print, that includes an estimate of the condition of
the slide as a photograph is held by the Section.

274

Annals of Carnegie Museum

vol. 35

List of photographs

Kodachromes Black and white Black and white

Swauger and Rial Swauger Pittsburgh Photographic

Library

General views of the site

901-905, 908, 911, 678, 679, 686-690,

912, 918, 919, 923- 746, 747, 818-822,

936. 939-958, 1561, 829-834.

1567-1573, 1576-

1578, 1591-1602,

1605, 1612, 1619,

1656, 1661, 1663-
1681.

Historic markers, buildings, etc.

913-917,920, 921,

937, 938, 1581-1590.

General use of heavy equipment

906, 907, 909, 910,

922, 951-954, 1532-
1537, 1562-1566,

1574, 1575, 1579,

1580, 1603, 1604,
1606-1611, 1613-
1618, 1652, 1655,
1657-1660, 1662,

1678.

Test

959-999, 1500-1506,
1508-1513, 1524-1531,
1540.

680-685, 691, 692,
770, 789, 807-817,
823-828, 835-837,
868-874.

pit C.M. 1.

700-745, 748-764,
769, 771-774, 778-
781.

11608-11611.

11600-11607, 11621-
11623.

1518.

1507, 1514-1517,
1519-1523, 1538,
1539, 1541-1560,
1643.

1620-1642, 1644-
1651, 1653.

Test pit C.M. 2.

Test pit C.M. 3.

765-768, 776, 777,
782-788, 790-806,
854.

Test pit C.M. 4.

838-853, 855-867.

1683-1750.

Test pit C.M. 5.
693-697.

11593-11599, 11612-
11620.

11566, 11568-11580.

Miscellaneous

1654, 1682.

698, 699.

11577.

ART. 12. SYSTEMATIC NOTES ON NORTH AMERICAN BIRDS

3. The Northeastern Races of the Long-billed Marsh Wren
( Telmatodytes palustris)

By Kenneth C. Parkes
Associate Curator of Birds, Carnegie Museum

This is the third of a series of papers on the systematics and nomenclature
of certain North American birds. A general introduction and acknowledg-
ments for the series as a whole will be found in the first paper (Parkes, 1955).
In connection with the present paper I wish to add my thanks to Mr. Robie
W. Tufts of Wolfville, Nova Scotia, through whose good offices I was able
to examine two of the three known Nova Scotia specimens of Long-billed
Marsh Wren (the third, in the National Museum of Canada, I have also
seen). Most of the work on which the present paper is based was done in 1949.

The Long-billed Marsh Wren ( Telmatodytes palustris) has a wide distribu-
tion in North America. Like the Sharp-tailed Sparrow (Ammospiza caud-
acuta), it has both inland, fresh-water populations, and populations con-
fined to narrow coastal strips. Unlike the Sharp-tailed Sparrow, the Marsh
Wren’s range reaches the Pacific coast as well as the Atlantic. This study is
restricted to the tangled taxonomic and nomenclatorial history of the Long-
billed Marsh Wrens in that portion of the species’ range lying east of the
Rocky Mountains and north of the mouth of Chesapeake Bay.

As of the fourth edition of the A.O.U. Check-list (1931), the birds of the
area in question were divided as follows:

Telmatodytes palustris palustris (Wilson), Long-billed Marsh Wren.
Telmatodytes palustris laingi Harper, Alberta Marsh Wren.

Telmatodytes palustris dissaeptus (Bangs), Prairie Marsh Wren.

These vernacular names will be frequently used for convenience in the dis-
cussion beyond, since the usage of the subspecific names has been subjected
to much shifting back and forth.

Todd (1937) claimed that Bangs’ dissaeptus (type from Wayland, Massa-
chusetts) was not available for the Prairie Marsh Wren, as he (Todd) was
unable to separate Bangs’ type series from a series Todd considered to repre-
sent typical palustris. The A.O.U. Check-list Committee apparently agreed
with Todd, and in the Twentieth Check-list Supplement (Wetmore et al.,
1945, p. 446) the name for the Prairie Marsh Wren reverted to the next oldest
available name, iliacus Ridgway (type from Wheatland, Indiana). Todd’s
1937 paper was cited as authority for this change.

Todd had mentioned certain differences between the type specimen of
iliacus and other specimens from neighboring localities, but attributed these
differences to possible post-mortem color changes. Aldrich (1946, p. 131)
claimed that the type specimen of iliacus represented a migrant of the
Alberta subspecies known to that time as laingi Harper. Since the A.O.U.
Check-list Committee agreed with Todd in 1945 that dissaeptus was a
synonym of palustris , and (assuming Aldrich to have been correct about the
identity of the type of iliacus) since the name iliacus applied properly to the
Alberta Marsh Wren, it would appear that the Prairie Marsh Wrtn was
without a name. Oddly enough, the A.O.U. Check-list Committee reversed

275

INSTITUTION J-0:

1 3 fSfft

Submitted for publication, March 24, 1959
Issued, July 10, 1959

276

Annals of Carnegie Museum

vol. 35

itself in the Twenty-third Supplement (Wetmore et aL , 1948, p. 441), citing
Aldrich as authority for a change from iliacus back to dissaeptus for the
Prairie Marsh Wren. Actually, Aldrich left the question open, confining
himself merely to a reidentification of the type of iliacus as a side remark
in the description of a new race from the far West. He made no mention of
either the name dissaeptus or the population to which it had been applied.
The A.O.U. committee gave no indication as to why they no longer agreed
with Todd that dissaeptus should be considered a synonym of palustris.

As of the fifth edition of the A.O.U. Check-list (1957), therefore, the
nomenclatorial status quo is as follows:

Telmatodytes palustris palustris (Wilson), Long-billed Marsh Wren.
Telmatodytes palustris iliacus Ridgway, Alberta Marsh Wren.
Telmatodytes palustris dissaeptus (Bangs), Prairie Marsh Wren.

Over two decades ago Welter (1935, p. 4) pointed out that the so-called
Prairie Marsh Wren was actually a composite of two recognizable races, a
western and an eastern, for which he used the names iliacus and dissaeptus
respectively. His comments, which appear only in a brief foot-note in a
general life-history paper, have been overlooked or ignored in subsequent
years. Welter's diagnosis of the characters of the two alleged races appeared
in more detail in his Ph.D. thesis (Welter, 1932), the taxonomic portion of
which was never published. Only Oberholser (1938, p. 448), after Welter’s
paper appeared, recognized both iliacus and dissaeptus as distinct subspecies,
and he may have come to this conclusion independently, since he does not
quote Welter. Oberholser used the vernacular name Prairie Marsh Wren
for the western iliacus , and called the eastern dissaeptus Massachusetts
Marsh Wren.

There are three taxonomic points which obviously must be decided before
the nomenclature of the northeastern Marsh Wrens can be straightened out.
These questions may be listed as follows:

1. Was Todd correct in considering the type series of dissaeptus insepar-
able from palustris ?

2. Was Aldrich correct in identifying the type specimen of iliacus as a
migrant of laingi , the Alberta Marsh Wren?

3. Were Welter and Oberholser correct in believing that four rather
than three races should be recognized in the area covered by this study?

The treatment in the fifth edition of the Check-list, as given above, re-
quires that questions 1 and 3 be answered “no”, and question 2 answered
“yes”.

Believing that, as with so many other seemingly complex taxonomic prob-
lems, the solution lay in assembling, for once, a really adequate series of
specimens, I was able to bring together approximately 670 skins of Tel-
matodytes palustris representing only those northeastern populations under
discussion here. Museums and individuals from whom specimens were bor-
rowed are listed in the first paper of this series (Parkes, 1955).

The three questions listed above will be taken up one by one, and then a
summary of the characters and distribution of the subspecies I recognize
will follow.

1959

Parkes: Systematic Notes on North American Birds

277

1. The status of the name dissaeptus Bangs

Among the specimens examined during this study were two topotypes and
a number of near-topotypes of dissaeptus. Comparison of this series with a
near-topotypical series of palustris (vicinity of Philadelphia, Pa.) indicated
that the Massachusetts birds were nearer the inland race as exemplified by
specimens from central and western New York. There is some intermediacy
toward palustris, but I feel that the relationships of the Massachusetts
population to which Bangs applied the name dissaeptus are definitely with
the inland form. Much of the earlier confusion, including Todd’s rejection
of dissaeptus as a synonym of palustris, can be traced to the use of speci-
mens from the vicinity of the District of Columbia as typical of palustris
when making comparisons. Unfortunately there are far fewer specimens of
good palustris in museums than there are of these District of Columbia
birds. The latter show much intergradation with dissaeptus and also with
waynei, a race of the coastal marshes of southeastern Virginia and North
Carolina which is outside the scope of the present paper. When topotypes
of both palustris and dissaeptus are compared, as mentioned above, it be-
comes evident that the latter name is not a synonym of palustris, but is
available for an inland race.

2. The status of the name iliacus Ridgway

I have examined the type specimen of Telmatodytes palustris iliacus
Ridgway (U.S.N.M. 90199, from Wheatland, Indiana). It is most unfor-
tunate that this particular specimen should have been chosen as a type, as
it may or may not have been a migrant when collected (April SO, 1883).
After careful comparison with excellent series of both populations to which
this type has been assigned (the Prairie and Alberta Marsh Wrens), I have
come to the conclusion that this specimen does not represent a migrant of
the Alberta race. It is admittedly paler than the great majority of Prairie
Marsh Wrens, but it is by no means as pale as laingi, nor are the black por-
tions of the plumage suffused with light brown as in typical laingi. The
specimen represents either a light-colored extreme or (as Mr. Todd believed)
a foxed specimen of the Prairie Marsh Wren. Geographic evidence is even
more pertinent; I have seen no migrant specimens of the Alberta Marsh Wren
taken east of Oklahoma, and it is apparently uncommon even that far east.
The chances of the Indiana-taken type of iliacus being an actual derivative
of the Alberta population are, from the geographic standpoint, infinitesimal.
The Alberta Marsh Wren thus reverts to its proper name of Telmatodytes
palustris laingi Harper, as listed in the fourth edition (1931) of the A.O.U.
Check-list.

With the earlier name dissaeptus available for the inland population that
has been called Prairie Marsh Wren, the fate of the name iliacus depends on
the answer to question 3.

3. The number of recognizable northeastern races

Examination of the excellent series of specimens assembled for this study
convinced me that Welter and Oberholser were perfectly correct in their
assertion that two subspecies are involved within the range described for
dissaeptus in the fifth edition (1957) of the A.O.U. Check-list, making a

278

Annals of Carnegie Museum

vol. 35

total of four within the area encompassed by this study. Names are avail-
able for all races (palustris, dissaeptus, iliacus , laingi), and no major change
of concept of the names as used in the fourth edition (1931) of the A.O.U.
Check-list is necessary other than the revival of iliacus Ridgway for the
western component of the “Prairie Marsh Wren”.

A summary of the characters and distribution of the four recognized races
follows.

Telmatodytes palustris palustris (Wilson)

Certhia palustris Wilson, American Ornithology, 2, 1807, p. 58 (borders of
the Schuylkill and Delaware rivers at Philadelphia, Pennsylvania).
Characters: All browns, especially on sides of neck and flanks, more grayish
or olivaceous, less reddish or buffy than in other races; browns of back and
rump more variable, but show the same tendency. Crown the blackest of the
four races, with the brown central stripe often reduced. When present, this
crown stripe is almost always well defined, not diffuse. Lores white. Black
and white interscapular region averaging somewhat more extensive, with
broader white streaks. White superciliary stripe well marked, breaking up
at posterior end into white dashes which tend to be continuous with white
dashes of interscapular region. A brown collar seldom present between black
of crown and of interscapular region. Throat and breast usually immaculate
white. Slightly smaller and bill more slender than inland races (see measure-
ments).

Range : Strictly a coastal form; all inland specimens I have examined from
eastern United States north of the Florida peninsula are referable to dis-
saeptus or iliacus no matter how they were originally identified. This in-
cludes, for instance, such specimens as that from Berea, Kentucky (which I
have examined), identified as palustris by Wetmore (1940, p. 549). Nominate
palustris breeds, in general, in coastal and estuarine marshes from Rhode
Island to Virginia. Intergrades with dissaeptus in Connecticut (presumably),
the Hudson valley, western New Jersey, and southeastern Pennsylvania.
Birds from Chesapeake Bay average nearest palustris; those from Washing-
ton, D.C., and adjacent parts of Virginia and Maryland are variously inter-
mediate between palustris and dissaeptus , with some individiuals showing
an approach to waynei, the next race to the south along the coast. Inter-
gradation with waynei manifests itself most conspicuously in the presence
of barred upper tail coverts. Winters in much of its breeding range, and
south through most of peninsular Florida. Accidental in Nova Scotia.

Telmatodytes palustris dissaeptus (Bangs)

Cistothorus ( Telmatodytes ) palustris dissaeptus Bangs, Auk, 19, 1902, p. 352
(Wayland, Massachusetts).

Characters : All browns more reddish than those of palustris. Brown always
present on crown, almost always diffuse, often covering almost all of crown,
frequently extending backwards and laterally as a collar between black of
crown and of interscapular region. Lores usually white, sometimes pale
buffy. Black interscapular region averaging somewhat less extensive, white
markings narrower than in palustris. Superciliary line often less pure white,
less distinct. Postocular spot or line usually brown (black in palustris). Brown
of sides averaging more extensive, especially anteriorly. Breast frequently
washed with pale brown or orange brown.

1959

Parkes: Systematic Notes on North American Birds

279

Range : Breeds from southern New Brunswick, southern Maine, Vermont,
New Hampshire, southwestern Quebec and southern Ontario south to the
mountains of western Virginia, eastern West Virginia, south-central Penn-
sylvania, and Ohio. Iritergrades with palustris just inland from the coast
from Rhode Island to Virginia, and with iliacus in central Michigan; un-
doubtedly also intergrades with iliacus in western Ohio, eastern Indiana,
and south-central Ontario. Winters at least casually through most of the
breeding range, south to the Gulf states from Mississippi eastward. Accidental
in Nova Scotia.

Telmatodytes palustris iliacus Ridgway

Telmatodytes palustris iliacus Ridgway, Proceedings of the Biological Society
of Washington, v. 16, 1903, p. 110 (Wheatland, Knox Co., Indiana).
Characters'. All browns lighter and more reddish than in dissaeptus. Crown
almost always more suffused with brown than in that race. Lores and an-
terior portion of superciliary line usually buffy, sometimes white. Ear covert
region more rusty. Brown of crown tends to be more extensive posteriorly.
Light markings of outer webs of secondaries and tertials more extensive,
tending to be confluent in some specimens. Always at least slightly stained
with rusty below, often most of the underparts so washed.

Range : Breeds from southwestern Michigan, western Indiana and western
Ontario south to southern Illinois, Missouri and eastern Kansas, west through
the Dakotas and Manitoba. Most Manitoba specimens are of this race; some
from the western part of the province show the influence of laingi. As is
also the case with dissaeptus , the type locality of iliacus is near the eastern
periphery of the range of the subspecies, rather than near the center of
differentiation. The characters of the race to which the name iliacus is being
applied are shown to best advantage in specimens from the upper Mississippi
drainage, particularly Minnesota and Wisconsin. This subspecies winters
through much of southeastern United States, west at least to southeastern
Texas. Like many other midwestern birds, iliacus has a southeast-northwest
migration route, and may be found on migration within the breeding range
of dissaeptus , and casually as far east as Long Island, New York, in the range
of palustris (A.M.N.H. 66451, Montauk Point lighthouse, October 13, 1888).

Telmatodytes palustris laingi Harper

Telmatodytes palustris laingi Harper, Occasional Papers of the Boston So-
ciety of Natural History, v. 5, 1926, p. 221 (Athabaska Delta, Main Branch
[9 miles above mouth], Alberta, Canada).

Characters : A very distinctive race, with all browns paler and buffier than
in iliacus. All black areas of other races, including interscapular region, are
in laingi more or less invaded by browns. Under tail coverts immaculate or
barred with pale rusty, never with black barring ( iliacus almost always has
at least a trace of black in the barring, which is often heavy). Differs from
plesius and other western races in lacking bars on the upper tail coverts.
Within the range here assigned to laingi , there is some geographic variation
in size. Topotypes average slightly longer in wing and tail than any of the
other races discussed here. Saskatchewan birds average somewhat smaller in
these measurements. In the accompanying table, after the column giving

280

Annals of Carnegie Museum

vol. 35

measurements of a composite series of male laingi comparative wing and tail
measurements are given for a small series each of male topotypes from Alberta
and males from Stalwart and Last Mountain Lake, Saskatchewan. The only
female laingi measured were from the latter series.

Range: Breeds east of the Rocky Mountains in Alberta. The breeding Marsh
Wren of Montana east of the Rockies is probably this race (no specimens
seen), and is so assigned by the A.O.U. Check-list (1957). Birds of all but
easternmost Saskatchewan are closest to laingi in color, but may show some
signs of approach to iliacus. Presumably intergrades with plesius along the
east face of the Rocky Mountains; no breeding birds from this area seen,
but winter and migrant specimens which appear to be laingi-plesius inter-
grades have been examined from New Mexico, Texas, and Mexico. Prin-
cipal winter range apparently in Mexico; migrants examined from Colorado,
Oklahoma, New Mexico, and Texas.

REFERENCES

Aldrich, John W.

1946. New subspecies of birds from western North America, Pro-
ceedings of the Biological Society of Washington, v. 59, p.
129-136.

American Ornithologists’ Union

1931. Check-list of North American birds, ed. 4. 526 p.

1957. The same , ed. 5. 691 p.

Oberholser, Harry C.

1938. The bird life of Louisiana. Louisiana State Department of
Conservation Bulletin 28. 834 p.

Parkes, Kenneth C.

1955. Systematic notes on North American birds. 1. The herons and
ibises (Ciconiiformes). Annals of the Carnegie Museum, v. 33,
p. 287-293.

Todd, W. E. Clyde

1937. Critical remarks on the Long-billed Marsh Wren. Proceedings
of the Biological Society of Washington, v. 50, p. 23-24.
Welter, Wilfred A.

1932. The natural history and taxonomy of the Marsh Wren Tel-
matodytes palustris (Wilson). Cornell University Ph.D. thesis,
unpublished. 167 p.

1935. The natural history of the Long-billed Marsh Wren. Wilson
Bulletin, v. 47, p. 3-34.

Wetmore, Alexander

1940. Notes on the birds of Kentucky. Proceedings of the United
States National Museum, v. 88, p. 529-574.

Wetmore, Alexander (Chairman) and others

1945. Twentieth supplement to the American Ornithologists' Union
Check-list of North American birds. Auk, v. 61, p. 436-449.

1948. Twenty-third supplement to the American Ornithologists’
Union Check-list of North American birds. Auk, v. 65, p.
438-443.

TABLE 1. MEASUREMENTS OF TELMATODYTES PALUSTRIS

1959

Parkes: Systematic Notes on North American Birds

281

j© Ch'

.5 eg S 3 5

q J XO — 'CM > —

00

xO

M 5 ^rr-

*^3 © nt< ©

-©

PS

V3 i— H
CO I>*

■Ofi oo

CO

°° ^oo

CO

00

•CO

GO • v^wi ■ w cm

d *> oo £

00

' ' iO
O CM

CM CM^o
CM
CO
©

•«> pj °9 (S'

^ ^ fX 10 ^ ^

■S < >9 S ^

!3

©
^ CM
xO

'Sb 8 g- g s-

•5 S«S«$2eSS3«

>— i -W-CM '«-'© v — . ©

xO ^ '

©

CM

00 '©■GO

a ” © cm ph ^ cm

.53 H^^trJH^HCOooX^
^2 xo '^'co -w©

- © ^

xO

©
©
CO

©

©

CM

^9 • oo

CM «? s T
i— h vi^CM •©'CD

’’f

xO

2©

© S «9 © &

© CO o

00

*5 CM -©*© ©“• CM '©'t" '©'CM f©'
2 °° -- • 0O • xfj CQ 50

^ nfs ^ ^ © i >-h co go c5

iu 0O oo •^'I>» '-'XT5 n_- I ©

£ S o

t3 ^ co

« ©

xo ^ijo ;©'xr> j©eo ^©'©

CO wi4» s-'l>

©

oo

CM

00

S”®

,«2

2 oo
^ <*
«

3^2©

^ c* ?? cm ©

'© ^XJO
®° ©

CM

CM

,-•00 -©- I— I

^ <® CM ^ '"T

»? ^ ^ ^
i— t GO ' — ^©

xfj ©

GO

-©-xo ©no
© • © CM

© CO

CO cO o

^ "-"©
00
CM

| ,

1/3

PS

■5

CO U

*5h

•*-»

CO

O

p0

x->

PS

bo

fl

<D

CO

w ^

M f-i

a j§
< &

G

a

o

3

\

£

Hh

OJ

T3

< 2
s§.

W

f*n b0

S bo

43

C

rS

r— 1

Hh

_

fl

n3

£

cS

h

3

<U

Q

s

g

H

PQ

(0.305) (0.321) (0.316) (0.317)

'.7 3
pt&S 4z

ART. 13. NOTES ON A HERPETOLOGICAL COLLECTING
TRIP THROUGH THE SOUTHEASTERN
LOWLANDS OF BOLIVIA

By Carl Gans

Carnegie Museum and The University of Buffalo, Buffalo, N.Y.

INTRODUCTION

This paper is a report on a brief collecting trip across the department of
Santa Cruz which includes the southeastern lowland region of Bolivia (chaco
boliviano). This region has been made much more accessible by the con-
struction of the Brazil-Bolivia Railway, and it was the courteous help of its
officials and personnel that made the trip both possible and successful.

I was accompanied on the trip by Padre Francisco Silverio Pereira, C.M.F.,
of the Departamento de Zoologia (Sao Paulo, Brazil). Because of our personal
interests, collections consisted mainly of insects, amphibians and reptiles,
though small groups of other forms were also collected (c/.. Lane, 1956;
Bechynee, 1956, 1958; and Parodiz, 1958). This paper presents the general
itinerary, a brief description of the region traversed and an annotated list of
the approximately 900 specimens of reptiles and amphibians resulting from
the fourteen-day trip (February 21 to March 5, 1954). The 800 frogs represent
30 forms, thus tripling the fauna reported by the first German Gran Chaco
Expedition, which visited this region from August to October 1926 during
the dry season which is much less propitious for amphibian collecting. In
view of the scanty information regarding the life histories of most of these
species, field observations are also included.

One of the aims of publication is to call the attention of specialists
to the existence of this collection. The status of the herpetofauna of central
South America is still such that even preliminary identification of a collection
often requires some revisionary work on the specific and generic level, as
well as an acquaintance with the species in regions far removed geographically.
In the absence of recent revisions, identifications of reptiles were carried to
the specific level only. Various specimens were made available to specialists
who had revisions in progress.

Specimens were deposited in the following institutions. They are iden-
tified by code letters throughout:

A.M.N.H. American Museum of Natural History, New York.

C. M. Carnegie Museum, Pittsburgh, Pa.

D. Z. Departamento de Zoologia, Seer, de Agricultura, Sao Paulo, S. P.,

Brazil. (A representative series of frogs assigned M.C.Z. numbers
will also be transferred to this institution).

M.C.Z. Museum of Comparative Zoology, Cambridge, Mass.

ACKNOWLEDGMENTS

It is a pleasure to acknowledge the aid and assistance of the many friends
who made the trip possible. At the risk of omitting others, I should like to
express particular thanks to Dr. Lauro Travassos of the Instituto Oswaldo
Cruz and Dr. Carlos de Paula Couto of the Museu Nacional both of Rio de
Janeiro; Dra. Maria Apparecida Vulcano de Andretta, Mr. Werner C. A.
Bokermann, and Dr. Paulo E. Vanzolini, of the Departmento de Zoologia,

283

Submitted for publication, April 12, 1958.

Issued June 3, 1960.

284

Annals of Carnegie Museum

vol. 35

Sao Paulo; Dr. Alphonse R. Hoge and Dr. Helio E. Belluomini of the In-
stituto Butantan; Capt. Luis Nicolau Velasco of the Instituto Geographico
Militar, La Paz; and Mr. Benjamin R. Moser of the United States Depart-
ment of State.

I am very grateful to the Comissao Mixta Ferroviaria Brasileiro-Boliviana,
in particular to Dr. Luis Alberto Whately, Eng. Chefe and Dr. Francisco
Gonsalves de Aguiar, Eng. Chefe Substituto, who arranged for room, board
and transportation during our travels. The hospitality and assistance of Dr.
Carlos Moreno T., Mr. Alberto Silles Rogas, Dr. Armillo Monteiro, Dr.
Alfonso Valderama, Dr. Rolando Chasal, all of the Comissao Mixta, are also
acknowledged with pleasure.

Specimens were contributed by several individuals, in particular by Madre
Esther Bottega of the Colegio “Santa Clara” of San Jose de Chiquitos, and
by Sr. Orivaldo Oliveira Nogueira of El Porton.

The lizards were identified by Dr. P. E. Vanzolini. Since circumstances
prevented Mr. W. C. A. Bokermann from completing the identification of the
frogs, the collection was reviewed by Mr. Benjamin Shreve of the Museum
of Comparative Zoology. The latter also reviewed certain doubtful iden-
tifications of snakes made by me. I certainly appreciate having had the benefit
of his wide knowledge of South American reptiles and amphibians.

Joseph R. Bailey, Robert L. Dressier, Coleman J. Goin, M. Graham Net-
ting, Neil D. Richmond, William J. Riemer and Ernest E. Williams read
and commented on the manuscript. Messrs. Lee Bolton and Robert D.
Weigel printed the photographs and Howard W. Campbell confirmed some
of the scale counts.

The trip discussed here was made while I was the recipient of a John
Simon Guggenheim Memorial Fellowship, and 1 should like to thank the
Foundation and especially Messrs. Henry Allan Moe and James F. Mathias
for their continuing and appreciated interest. The report was completed with
the aid of a postdoctoral Research Fellowship from the University of Florida.

I should finally like to thank my colleague Padre Francisco Silverio Pereira,
C.M.F., for his pleasant companionship during our journey.

GENERAL

Since the time was unfortunately limited it was known from the beginning
that the trip could at best amount to a preliminary survey of this poorly
known region. In order to achieve the maximum results in terms of collec-
tions it was planned to restrict the stops to not more than three days each.
Side trips involving auxiliary transport were also avoided in order to permit
most efficient use of the brief period available. It was reasoned that any
station within a recently opened area would have approximately equal prob-
ability of yielding interesting material and that the greatest need was for
adequate series of specimens. However, this restriction to the transect formed
by the railroad right of way did make for a limited view of the country and
the comments should be evaluated with this in mind.

Frog collecting was most productive from the start and most of the work
was therefore done at night to take maximum advantage of this.

The collecting stations (Fig. 1) follow the right of way of the railway be-
tween Corumba, Mato Grosso, Brazil and Santa Cruz de la Sierra, Bolivia,

O'

Map of collecting stations and profile of the railroad right of way from Corumba, Brazil to Santa Cruz de la Sierra, Bolivia.

Modified from map furnished by C.M.F.B.B.

286

Annals of Carnegie Museum

vol. 35

capital of the department of the same name. With the exception of Corumba
all stations were within the Bolivian department of Santa Cruz, which is the
largest in that country, though far from being the one best known, herpeto-
logically or otherwise. Through it passes the divide between the Rio de la
Plata and Amazonas drainages, but with few exceptions the divide is very
low, and the drainage of large districts may be either way depending upon
the relative rainfall. Haseman (1912) who made extensive fish collections in
the vicinity, comments upon the fact that the barriers, if any, to the distribu-
tion of fluviatile species between South American watersheds are generally
not formed by the divides.

The main branches of the two river systems flow in parallel but opposite
directions and effectively delimit the region in question. East of the Paraguay
stretch the lowland Pantanals, which apparently extend up from the South.
Corumba lies on a bluff on the west (at this point actually south) shore of
the river, which here describes an arc, though its general course is southward.
Near the other side of the department lies the northward flowing Rio Grande
(Guapay), with the city of Santa Cruz situated some 40 km. west of this. The
extreme flatness of the region between the two river systems is broken only
by the several Serranias and may be noted from the profile of the right of
way shown in Fig. 1. For the 230 km. between San Jose and Pto. Pailas the
variation of height amounts to less than 25 m., most of which occurs at the
eastern end of the stretch. East of the crossing of the Serrania de Santiago,
at El Porton, the profile undergoes greater changes because the railway line
has been located away from the annually flooded lowlands.

Rainfall and temperature data (Fig. 2 and 3) are available for only the
two terminal points of the traverse.

Readers desiring more detailed climatological data for Brazil are referred
to Margarinos Torres and Mortera (1948) and Anonymous (1941), as well as
to the various publications of the Instituto de Geographia e Estatistica in
Rio de Janeiro. Bolivian data are poorer and much harder to obtain. Some
information is given by Pena and Escobar (1947), while Osborne’s (1954)
little booklet “Bolivia” gives a good bibliography of other material. Geo-
logical information on Bolivia may be found in Ahlfeld (1946).

The data for Santa Cruz in Fig. 2 and 3 are theoretically applicable only
to the station at El Porton, since Santa Cruz lies on the outlying foothills
of the Andean chain and at a higher elevation than all but this station. The
apparent effect of this on the vegetation is mentioned by Herzog (1913, p. 2)
who comments on the presence of Brazilian, Andean and Argentinian ele-
ments in the Santa Cruz flora.

The monthly data however, do give a very excellent indication of the
extremely seasonal rainfall pattern. Thus the rainy season lasts from October
through March, and the country is partially inundated during this time.
Prior to the advent of railroad and plane service, transportation used to be
close to a standstill during this season. Upon the cessation of the rains there
is said to be a rapid drop in the water-table and the winter is dry.

The vegetation appears to reflect the rainfall cycle, and the peculiar inter-
digitation of pantanal and cerrado. As shown by the views in Fig. 4-7, most
regions are covered by a scrub forest ranging in many places to a height of
approximately 10 m., with a high frequency of thorny and xerophytic species.

Monthly Temperatures - -Degrees Centigrade

1960

Gans: A Herpetological Collecting Trip in Bolivia

287

Fig. 2. Monthly temperature data for Corumba and Santa Cruz de la Sierra.
The overall and minimum mean monthly temperatures for Corumba are from
the Normais Climatologicas (Anonymous, 1941) and represent averages for the
years 1922-35. The Santa Cruz data are from Pena and Escobar (1947) and

apply to 1945 only

288

Annals of Carnegie Museum

vol. 35

Fig. 3. Monthly rainfall data for Corumba and Santa Cruz de la Sierra from
the same sources as Fig. 2. The Corumba points represent an average for the
years 1912-13 to 1935. Those for Santa Cruz give the rainfall for 1945 only

1960

Gans: A Herpetological Collecting Trip in Bolivia

289

Particularly noticeable were a tall cactus, five meters or more in height
(Fig. 6) and a spiny bottle trunk tree ( Chorisia ?) of the family Bombacaceae
(Fig. 7), scattered through the forest. There appear to have been no studies
made of this flora except for purely taxonomic ones, and Mann’s popular
volume (1951) offers only some general comments.

The first recent account of the area resulted from an Argentine expedition
(Lizer, 1919) investigating the distribution of Schistocerca paranensis. A more
extensively illustrated general report is given in the travelogue of the 1925-26
Instituto Oswalda Cruz medical expedition (Fonseca, 1929). The herpeto-
logical results never appeared in print even though it is stated that the snakes
were turned over to Dr. Oswaldo Mello. Another very much popularized
and somewhat lurid description will be found in Duguid’s book “Green Hell”
(1931).

It may be of interest to compare the present collection with that compiled
in the dry season by the first German Gran Chaco Expedition, which spent
more than two months in the Province. Such comparison (shown in table
on page 290) has been carried out for the amphibians by using Muller and
Hellmich’s report (1936). Of the species reported by them from this region,
all but one (Hyla phrynoderma from Santa Cruz de la Sierra) were obtained
during the present trip. Nine additional forms recorded by them from north-
ern Paraguay and Argentina, and 10 forms not mentioned at all are added to
the known fauna of the province by this report.

Reference should also be made to the extensive collections made in Santa
Cruz de la Sierra over many years by the elder J[ose] Steinbach. His speci-
mens are now scattered in many museums, with many in the Carnegie Mu-
seum and in the Museum of Zoology of the University of Michigan. The
present collection complements rather than overlaps his material as he does
not appear to have visited the eastern lowland regions, and restricted his
activities to the slopes of the Andes between the cities of Cochabamba and
Santa Cruz de la Sierra.

Recently there appeared the report of the second Oswaldo Cruz Expedition
(Travasso, et al., 1957). This expedition obtained 26 reptiles and 68 amphib-
ians, representing a total of 21 species collected and sacrificed for a study
of their parasites. All of the Bolivian species recorded by them are also
listed in the present report.

The construction of the railway line has already affected the nature of the
country in two ways. First is the excavation of drainage ditches and borrow-
pits, from which soil had apparently been removed for fill. (Because
of the lack of suitable rock outcrops it has been necessary to transport the
track ballast from the vicinity of Corumba.) Many of the borrow-pits were
filled with water and provided excellent frog breeding pools. The second and
far less superficial change was produced by the increasingly intensive cattle
grazing and the cutting of the forest to provide “lena” (fire-wood) for the
wood-burning locomotives. Experience in Brazil has shown the remarkable
extent to which the grazing of cattle will keep down the shrubs and smaller
plants of the cerrado. When the larger trees are selectively logged at the
same time there is little opportunity for replacement. Seedlings will be
cropped before they can grow to size and the character of the landscape
changes.

290

Annals of Carnegie Museum

vol. 35

Dendrobates flavopictus
Bufo granulosus major
paracnemis
typhonius

Hyla geographica punctatissima
leucophyllata
megapodla
minuta
rubra nasica
raniceps
senicula

Phrynohyas zonata
Sphoenohyla nana
Phyllomedusa hypochondrialis
pa i Iona
sauvagi

Pseudis paradoxa bolbodactyla
Leptodactylus bufonius
mystaceus
my st acinus
ocellatus

pentadactylus labyrinthicus
p. podicipinus
podicipinus petersi
sibilatrix
Physalaemus cuvieri
fuscomaculatus
Eupemphix natteri
Elachistocleis ovalis

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

- X
X

- X

- X

X

X

X

X

X

X

X

X

X

X

Hypopachus mulleri

Travassos & Travassos (1957)

1960

Gans: A Herpetological Collecting Trip in Bolivia

291

LIST OF HERPETOLOGICAL LOCALITIES IN ORDER OF
ITINERARY

CORUMBA, Mato Grosso, Brazil (19°S.-57°38'W., elev. 215 m., km. 0).*

We arrived during the night of February 20, and on the evening of Feb-
ruary 21 collected frogs in the flooded fields near and beyond the C.M.F.B.B.
station. On February 22, Dr. Francisco Gonsalves de Aguiar arranged for a
trip to the vicinity of Mt. Urucum, type locality of many Peracca species.
In the evening we continued the frog collecting both at the old localities
south of the town and along the shore of the Rio Paraguay below the bluff

Fig. 4. View of the Corumba bluff from the shore of the Rio Paraguay. The
material in the foreground is fill extending out into the fringing marshes
that harbor large populations of Leptodactylus ocellatus. The city of Cor-
umba lies atop the bluff on the far side of the palm-flanked avenue

on which the town is situated (Fig. 4). Further frog collections were made on
February 23 and 24, much of the last day being spent in obtaining a series
of Teius teyou in scrub areas west of town.

EL CARMEN, Estacion Rivero Torres, Santa Cruz de la Sierra, Bolivia
(18°48/30"S.-58°33'50"W., elev. 212 m., km. 113).

The station lies slightly above the northern extensions of the flooded pan-
tanals, often characterized by open stands of palms, which intervene between

* These co-ordinates are estimated from the map; all others are from Pena and
Escobar (1947). The elevations are from a map kindly furnished by the C.M.F.B.B.
and refer to those of the railway stations, while the last figures give the distance (in
kilometers) west along the right of way from the Corumba station.

292

Annals of Carnegie Museum

vol. 35

Corumba and Carmen, and approach to within one kilometer of its eastern
side. For a view of the cut-over, flooded low country see Figure 2, plate 29
in Duellman (1958), which was taken near Palmito 25 km. east of El Carmen.
We arrived by train on the evening of February 25 and enjoyed the memor-
able hospitality of Dr. C. Moreno T., who did everything possible to facilitate
our work. The weather was clear and hot during our stay and the last rains
were said to have fallen on February 23.

On the evening of February 25 we collected frogs and insects in a series
of pools and cattle wallows along the right of way, one to three kilometers
east of the station. During the day of February 26 we visited the station

Fig. 5. View of retaining wall in Rivero Torres Station (El Carmen) show-
ing two of the drainage holes in the foreground. All of the small frogs were
taken in the short grass below the wall

reservoir and returned at night to the region east of the town. On the 27th
we worked over several log piles in the railway yard and also collected
through a farmed region to the south.

At the suggestion of Dr. Moreno T. we spent the early evening in an in-
spection of a low stone wall within the station (Fig. 5). This locality proved
to be immensely fruitful. Large numbers of juvenile Physalaemus cuvieri and
fuscomaculatus , a series of Elachistocleis and several species of Leptodactylus
were taken in the short grass before the retaining wall or in its drainage holes.
It was far from clear what attracted this unusual aggregation of small frogs
to this relatively barren site. A series of Phyllopezus was later collected on
walls of the station building.

We left on the morning of February 28, travelling by train to NARANJOS

I960

Gans: A Herpetological Collecting Trip in Bolivia

293

Fig. 6. Cactus of undetermined species approximately seven meters tall at
the edge of dense scrub forest near El Carmen

294

Annals of Carnegie Museum

vol. 35

Fig. 7. Spiny bottle trunk tree, possibly Chorisia sp., family Bombacaceae in
scrub near El Carmen. The picture had to be taken diagonally to avoid
obstructions in the foreground

1960

Gans: A Herpetological Collecting Trip in Bolivia

295

(elev. 260 m., km. 180), where we had time to shoot two of several specimens
of Tropidurus torquatus sunning themselves on a pile of ties. From here
we continued to Robore by motorized hand-car.

ROBORE, Santa Cruz de la Sierra, Bolivia (18°20T2"S.-50°59T3"W., elev.
530 m., km. 251).

This town lies to the south and just below the first hills of the Serrania
de Santiago, and a small creek from the mountains passes through the town.
We stayed in the station guest house and spent the first evening collecting
frogs which were calling from the many small pools and puddles left by the
day’s rain. The following day (March 1) we walked up the ravine shooting
Teius and collecting insects. The forest of the hills seemed denser and higher
than that of the lowlands through which we had passed from Naranjos. That
evening we obtained a ride to a spot some four km. south of the station. Here
we collected a number of species and observed a Caiman in a system of

Fig. 8. Carnival celebration at Limoncito, 33 km. east of El Porton. Com-
pare with illustrations in Fonseca (1929)

interconnected sloughs and pools fringed by the open scrub forest on very
sandy soil. We collected a number of other forms while walking back to the
station and flushed many night-hawks resting on the road. We left by motor-
ized hand-car after noon on March 2.

EL PORTON, Santa Cruz de la Sierra, Bolivia (18°10'S.-60°8'W., elev. 550
m. at highest point of track, km. 298).

This station (Fig. 10) is located on the southeastern end of the railroad cut
which forms the highest point between Corumba and Santa Cruz. The line

296

Annals of Carnegie Museum

vol. 35

runs along the base of the escarpment (Fig. 9), passes through the cut and
descends after passing across the divide into the Amazon drainage system.

We arrived at dusk on March 2 and during the night collected through
the flat station area and the railroad cut. The temperature was noticeably
lower than in the lowlands, and the last of the frogs had stopped calling
by 0130 hours. Mr. Orivaldo O. Nogueira presented us with five preserved
snakes collected in the vicinity.

Prior to our departure the next morning we walked again through the cut
where we found large numbers of Dendrobates flavopictus and shot various
lizards on the rock walls.

SAN JOSE DE CHIQU1TOS, Santa Cruz de la Sierra, Bolivia (17°50'S.-60°45'
W., elev. 350 m., km. 385).

This very old settlement lies in a region of low scrub some kilometers north
of the escarpment of the Serrania de San Jose which is visible from the out-

Fig. 9. View of the southern escarpment of the Serrania de Santiago from
railway track 5 km. east of El Porton

skirts of town. We arrived in a drizzle on the afternoon of March 3, and after
being established in a guest house spent most of the night collecting around
the station. On the morning of March 4 we visited the church, which had
been founded in the early eighteenth century and now housed a missionary
group. They showed us about the compound and assisted us in collecting
bats, lizards and frogs in and about the old buildings. Madre Esther Bottega
and several of her pupils accompanied us on an insect collecting trip to the
east of town during the afternoon.

We left by motorized hand-car on the morning of March 5 and were

1960

Gans: A Herpetological Collecting Trip in Bolivia

297

dropped at TUNAS (elev. 335 m., km. 492) where we collected some frogs,
while waiting for the rail-operated tank-truck that was to take us to the
Rio Grande.

EL PAILON, (5 km. from E. shore of the Rio Grande), Santa Cruz de la

Sierra, Bolivia (17°42/S.-63°44/W, elev. 350 m., km. 601).

After several hours of travel in the rain we arrived at this station at dusk
on March 5, to find that it would be impossible to cross the river that night.
The station was located in slightly higher forest close to the edge of the river
zone which was characterized by small streams and pools. We slung our
hammocks in a shack belonging to the railroad and immediately started to
collect. Enormous and deafening choruses sounded from all directions and
it was difficult to distinguish the voices of individual species in the general
din. Far greater numbers of individuals could have been taken if the aim
had not been directed toward obtaining series of a maximum number of
species. Though most of the pools appeared to be identical to the observer
it was repeatedly found that certain species had congregated in or around a
single pool, with no specimens being taken anywhere else. Sixteen species
were collected during the night and it appears certain that others were also
present, particularly since congregations of two of the forms were not located
until just before exhaustion forced me to retire.

Fig. 10. View south across station at El Porton. The main line passes from
the center toward the left, and only a spur passes toward the right. Hillside
from which the picture was taken was locality where Hyla senicula was heard
and collected. Hyla megapodia , Eupemphix natteri, Physalaemus cuvieri,
Leptodactylus ocellatus and L. p. podicipmus were collected in the wet grassy
region in the center of the picture

298

Annals of Carnegie Museum

yol. 35

Since the track was in too poor condition to permit the use of powered
vehicles we walked the five kilometers to the edge of the Rio Grande, while
our baggage followed on a hand-car pushed by two workers. The forest was
much higher near the river, though the term gallery forest would still seem
inappropriate. Small but swollen creeks meandered through it and had under-
cut the roadbed in a number of spots. The Rio Grande was wide, shallow
and interrupted by large sand-bars. We crossed it on a barge powered by
paddles and tow ropes. The station at the west shore was PUNTO DE
PAILAS (elev. 350 m., km. 608) where we collected only a single toad be-
fore the night train took us to SANTA CRUZ DE LA SIERRA (17°46T2S.-
63° 1 1'00"W., elev. 536 m., km. 651). The collections were packed here and
returned to Brazil with Padre Pereira, while I continued by plane to LaPaz.

ACCOUNT OF SPECIES
SALIENTIA

(Below the name of the species , digits at the beginning of each item indicate the

number of specimens)

Dendrobates flavopictus (A. Lutz), 1925.

32 El Porton, March 3. C.M. 36155 (15), M.C.Z. 29813-25 (+4).

These small brightly colored frogs were extremely common within the rail-
road cut through the mountain and around the reservoir drainage leading
down from the ridge. Their call is a piping cry hoowit-hoowit and could be
heard from ditches along the sides of the track.

Fig. 11. Southeast view of the northern escarpment of the Serrania de
Santiago, taken from the track some 3 km. west of El Porton. The forest
appeared much less disturbed here than in the lower areas

1960

Cans: A Herpetological Collecting Trip in Bolivia

299

The species appears to be definitely diurnal. A diligent search during the
night disclosed no specimens, though they were found to be abundant in
the early morning. In the cut they were restricted to the shady sections
where the rock, ballasted road-bed, flanked by shallow ditches, passed be-
tween two sheer rock walls. Here they would sit along the bottom of the
rock face or on the crushed stone of the road-bed. When disturbed they
would take a number of jumps up the very steep rock face and then enter
cracks or crevices.

The black body bears markings of bright yellow and red. Cott (1940, p.
204) remarked on the effective warning coloration of this genus and his com-
ments were brought home in force while watching these frogs frequenting
exposed localities in broad daylight.

Numbers of the specimens were carrying 8-10 tadpoles on their backs
The larvae had a body length of approximately 4 mm. and were arranged
crossways in two rows with the tails toward the outside. When shaken or
handled roughly the frogs would loose the tadpoles fairly readily. Several
frogs were observed to drop off larvae while crossing the ditches flanking the
track. The frogs were not discovered until we were almost ready to leave
the locality, and in my haste I neglected to sex the carriers.

Bufo granulosus major Midler and Hellmich, 1936.

36 San Jose, March 3-4. C.M. 36217 (18), M.C.Z. 29826-43.

These small toads were extremely common in open areas throughout the
town. They were found calling at the margins of small muddy puddles in
the middle of the roads and in closely cropped fields. Several pairs were
collected in amplexus. The call is a high nasal whine.

San Jose is the type locality of the race tentatively named B. g. major by
Muller and Hellmich (1936). Gallardo (1957) who has revised the group
considers this race valid.

Bufo paracnemis A. Lutz, 1925.

16 Corumba, Feb. 21-22. C.M. 35908-16, M.C.Z. 29302-08.

7 Robore, Feb. 28. C.M. 35892-6, M.C.Z. 29290-1.

3 El Porton, March 2. C.M. 35907, 36153, M.C.Z. 29301.

19 San Jose, March 3-4. C.M. 35897-906, M.C.Z. 29292-300.

2 El Pailon, March 5. C.M. 36168-9.

1 Pto. Pailas, March 6. C.M. 36152.

These large toads are apparently abundant throughout this area. Speci-
mens were collected hopping about the roads after dark and males were
observed calling in Corumba, Robore, San Jose, Pailon and Pto. Pailas,
though none was found at El Carmen.

Ticks from Corumba and Robore toads were identified as Amblyomrna
dissimile Koch by Dr. Joseph Bequaert of the Museum of Comparative Zool-
ogy.

Bufo typhonius (Linnaeus), 1758.

1 El Porton, March 2. C.M. 36233.

10 El Pailon, March 5. C.M. 36244-8, M.C.Z. 29844-8.

The Porton specimen was found sitting on the floating leaf of a plant in
the center of a small pool. The Pailon series was calling from a flooded
ditch near the railroad embankment about two kilometers east of the station.

300

Annals of Carnegie Museum

vol. 35

The call is a clack-clack-clack loudly and rapidly repeated several times. Two
pairs were seen in amplexus.

Hyla geographica punctatissima (Reinhardt and Liitken), 1862.

6 Four km. S. of Robore, March 1. C.M. 36108-10, M.C.Z. 29849-51.

The frogs were calling from a flooded thicket on the edge of a deep and
exceedingly ill-smelling pool. They were calling while sitting on the hori-
zontal branches of shrubs within 30-60 cm. of the water surface. Their call
is a peculiar chuckling cry, shreek-eck-eck-eck , repeated irregularly.

When alive the specimens were a dirty yellow brown dorsally and had a
whitish venter. The hands and feet were red. The belly was speckled and
the sides of the legs and the body were barred.

Shreve advises that although these specimens would be included under
punctatissima according to Parker’s ideas, more recent studies indicate that
the Bolivian population may be a distinct race.

Hyla leucophyllata (Beireis), 1783.

11 El Pailon, March 5. C.M. 36111-5, 36212, M.C.Z. 29852-6.

These small hylids always called from within 25 cm. of the water-level of
the ditches, generally while sitting on reeds, stems of bushes, grass blades,
etc. The call is a multisyllabic chuckle or chatter. One pair was collected in
amplexus.

In life the markings of these small frogs are a bright saffron yellow.

Hyla megapodia Miranda-Ribeiro, 1937.

20 Corumba, Feb. 21-23. C.M. 36141 (10), M.C.Z. 29857-66.

5 El Porton, March 2. C.M. 36147-9, M.C.Z. 29867-8.

24 El Pailon, March 5. C.M. 36151 (12), 36175, M.C.Z. 29869-79.

These frogs were extremely abundant at Corumba and Pailon, at which
latter site they were one of the chief contributors to the deafening choruses.
A large chorus was also heard in Robore though the site was unfortunately
inaccessible.

In this species the restriction of calling site appears to be far less severe
than in some of the smaller hylids. The Corumba frogs were calling while
clinging to the vertical stems of bushes and weeds growing in shallowly flooded
fields at the edge of a pool. The Robore site appeared to be similar to this.
At Porton the frogs were congregated around a flooded depression in a
meadow and were calling from the ground. In spite of the fact that there
were no bushes around the pool this might still indicate a more advanced
portion of the breeding cycle, were it not for the fact that several of the
Corumba pairs were in amplexus, while none of the Porton individuals were
observed in mating position. In Pailon again they were by no means re-
stricted to a single site, but specimens were observed and taken while calling
along the entire length of the embankment. Some of these frogs were sitting
on the sand around the tracks, on the grassy sides of the embankment and
at various elevations in the bushes flanking both sides of the embankment.

The Corumba chorus was composed of equal numbers of Hyla megapodia
and Hyla r. nasica. The latter form called from the same plants and pairs of
it were also found in amplexus. No mixed pairs were observed. The calls
of the two species were rather easy to distinguish, differing in frequency and
depth, and the average size difference of the two forms may also assist in
species recognition.

1960

Cans: A Herpetological Collecting Trip in Bolivia

301

Shreve informs me that he has seen the cotype of H. megapodia (M.C.Z.
15668) and does not believe it to be a synonym of 77. raniceps (Cope), 1862
as stated by Cochran (1955, p. 96).

Hyla minuta Peters, 1872.

7 El Porton, March 2. C.M. 36134(4), M.C.Z. 29880-2.

All of these specimens were calling from the aerial parts of plants rooted
in the water in the center of a pool one meter deep on the northwestern side
of the pass. Their call is more piercing than that of Sphoenohyla nana and
is given singly instead of in repetitive series. The latter form was calling from
various spots on the periphery of the same pool, generally from grasses
or reeds.

The color pattern of the specimens appears to fall into pattern (a) of
Cochran (1954, p. 122) though it differs in that the posteriormost marking
is a fairly regular V.

Hyla rubra nasica Cope, 1862.

26 Corumba, Feb. 22-23. C.M. 36161 (13), M.C.Z. 29883-95.

1 El Carmen, Feb. 25. M.C.Z. 29896.

2 1.5 km. east of El Carmen, Feb. 26. M.C.Z. 29898-9.

1 El Carmen, Feb. 27. M.C.Z. 29897.

1 San Jose, 1954, (M. Bottega). C.M. 34847.

3 El Pailon, March 5. C.M. 36100-1, M.C.Z. 29900.

All calling specimens and pairs of these frogs were collected in bushes
at or near water. In Corumba they were very common and were calling in a
mixed chorus with Hyla megapodia (See notes on that form) clinging to the
vertical stems of bushes and weeds rooted in the shallowly flooded fields at
the edge of a pool. The specimens from 1.5 km. east of Carmen were call-
ing from bushes 1-2 m. above and 3 m. from the edge of a pond. The Pailon
specimens were in bushes next to a flooded ditch. Both specimens collected
in Carmen were on building walls, one within the veranda of the residence
and the other on the inside wall of the railroad shed.

A number of the Corumba specimens were in amplexus as was the single
pair from Pailon. My notes mention that the calls of the two Pailon males
were more irregular than those of the Corumba specimens.

Shreve advises me that if this name has been correctly applied in this case,
the form seems to be but slightly distinct from northern II. rubra. What is
called II. megapodia in this report has in the past been referred to as H.
nasica.

Hyla raniceps (Cope), 1862.

1 San Jose, 1954. (M. Bottega). C.M. 34848. (See note under 77. megapodia .)

Hyla senicula Cope, 1868.

1 Corumba, Feb. 21. C.M. 36121.

3 El Porton, March 2. M.C.Z. 30137-9.

21 El Pailon, March 5. C.M. 36207, 36150(10), M.C.Z. 29901-10.

The Corumba specimen was calling from the same bush at the edge of a
small pool from which the individual specimens of Phyllomedusa sauvagi
and Sphoenohyla nana were collected. A number of specimens were calling
in the bushes of the bluff overlooking the station at Porton, during the early
part of the evening when we were collecting other species in the flooded

302

Annals of Carnegie Museum

vol. 35

station meadows. It started to cool off as the evening progressed and the
specimen collected was the only one still calling at 2200 hours. It called ap-
proximately 10-15 meters away from and above the water. All of the Pailon
specimens were calling from bushes 1 to 1.5 meters above the ground. No
females were noted or collected. The call is an extended wailing snore,
repeated several times.

Shreve advises that senicula and melanargyrea both seem to occur in
Bolivia. It is therefore doubtful whether they are both subspecies of mar-
morata (See also Cochran, 1955, p. 174).

Phrynohyas zonata (Spix), 1824.

2 juv. 1 km. east of El Carmen, Feb. 25. C.M. 26213, M.C.Z. 29911.

15 (7 $ $, 1 9, V juv.) El Pailon, March 5. C.M. 36087-90, 36143-6, M.C.Z.
29912-8.

These large hylids were calling from a floating position in the center of
various small pools. They were very shy, sinking when disturbed and re-
appearing at the edge of the pool amid the partially submerged grass. When
they rise, they hold the body almost vertically, and only the nostrils and eyes
break the surface. The single female was collected in a bush approximately
one meter above and in the center of the pool from which a number of
males were calling. The characteristic call is similar to the roar of a bull.

The Carmen juveniles were collected at night when they were sitting on
the sandy railroad embankment, some 300 meters from the nearest pool. The
Pailon juveniles came from similar localities.

The juveniles have the same whitish secretion as the adults. This is a
heavy slime of the consistency of well masticated bubblegum. It soon hardens
to a brownish solid that is extremely hard to remove from the hands. There
also appeared to be some toxic effect connected with it as my hands burned
and were slightly inflamed on March 6. It is possible, however, that part of
this was due to the continued exposure to formalin.

Sphoenohyla nana (Boulenger), 1889.

1 Corumba, Feb. 21. M.C.Z. 29919.

37 1.5 km. east of El Carmen, Feb. 25-26. C.M. 36171-3, 36183-5, 36201 (10),
26202(10), M.C.Z. 29920-30.

16 Robore, Feb. 28. C.M. 36203 (8), M.C.Z. 29931-8.

4 Three km. south of Robore, March 1. C.M. 36271-4.

7 El Porton, March 2. M.C.Z. 29939-45.

5 San Jose, March 3-4. C.M. 36174, 36178-9, M.C.Z. 29946-7.

4 El Pailon, March 5. C.M. 36180-1, M.C.Z. 29948-9.

All but one specimen of this abundant frog were taken or observed while
sitting on low bushes, reeds or other plants rooted in pools or growing in
temporarily flooded areas. The single exception was a specimen from
Corumba (M.C.Z. 29919) calling from a bush near a pool in an open field.
This observation is in agreement with Goin’s remarks on habits (1957, p.
29) in his discussion of this genus.

All of the series included calling males, and many more individuals were
heard than taken in every instance. The second Corumba specimen was col-
lected from reeds along the marshy shore of the Rio Paraguay, the Carmen
specimens from water plants in a pond, the Robore series from dead brush

1960

Gans: A Herpetological Collecting Trip in Bolivia

303

in a flooded slough at the edge of a lake. The Porton specimens were found
on reeds in the shallower portions of a stream and the San Jose and Pailon
series were also calling from water plants.

This small frog has a sibilant repeated call somewhat like the chirp of a
cricket.

Phyllomedusa hypochondrialis (Daudin), 1803.

11 1.5 km. east of El Carmen, Feb. 25-27. C.M. 36102-6, 36176, M.C.Z. 29950-4.

Males and females were sitting or walking on the aerial stems and leaves of
various plants growing in the shallow end of a pool. Most of the specimens
were within 10 cm. of the surface. These frogs moved very slowly when dis-
turbed and the males were often found calling with the legs extended asym-
metrically and hands and feet grasping stems and small branches. The
call was an irregular chatter that is difficult to describe.

In life these medium-sized frogs had a bright green dorsum. Their venter
was white and the sides of the legs were orange barred with black.

Shreve advises that these frogs may belong to the race azurea Cope, 1862,
which was not recognized by A. Funkhouser (1957).

Phyllomedusa sauvagi Boulenger, 1882.

1 Corumba, Feb. 21. C.M. 36092.

7 1.5 km. east of El Carmen, Feb. 26. C.M. 36137-40, M.C.Z. 29955-7.

All of the specimens were collected in the immediate vicinity of ponds or
pools, with the males calling. The Corumba specimen was two meters high
in the single bush of an open grassy area next to a small pool. The El Car-
men specimens were in trees around two small ponds, most of the specimens
being two to three meters up in the tree, with the sole exception of a male
calling from the top of a bush, 50 cm. high, in the center of one of the ponds.
Most of these frogs could be obtained by climbing for them, but one had to
be shot out of the top of a tall but slender bush.

The frogs moved very slowly with alternate movements of the limbs. They
were rarely seen to jump. When picked up they produced a white, slimy
secretion. The call is an irregular chuckling Corr-rr-rr-ack, and when several
specimens were calling at a single site they seemed to alternate.

These relatively large frogs are a bright metallic green in life, set off
sharply with white lateral lines and some isolated white spots.

Phyllomedusa pailona Shreve, 1959.

6 El Pailon, March 5. C.M. 36278-80, M.C.Z. 29677-9.

1 juv. El Pailon, March 5. M.C.Z. 29680.

The frogs of the first series were calling from points more than a meter
above the ground in trees and bushes along the flooded ditches. The single
partly digested juvenile was taken from the stomach of the hatchling
Chironius pyrrhopogon. The call of this frog is an irregular chuckle.

In life these specimens were green dorsally with yellow markings on the
sides of the body and along the legs.

This is the type series of the only new frog described from this collection.

Pseudis paradoxa bolbodactyla A. Lutz, 1925.

3 1.5 km. east of El Carmen, Feb. 26. C.M. 36222-3, M.C.Z. 29958.

7 El Pailon, March 5. C.M. 26218-21, M.C.Z. 29959-61.

The Carmen specimens were heard to call and were found floating amid

304

Annals of Carnegie Museum

vol. 35

the grass fringing the shore of a small pond. The Pailon specimens may
have been calling, but the din produced by the many other species was suf-
ficient to mask the call of all but the most vociferous forms. When disturbed
the specimens seem to submerge backwards and disappear without a ripple
showing on the surface (See Gans, 1956, for a popular discussion of habits.
Gallardo has suggested in conversation that the egg-mass illustrated in the
1956 note probably belongs to a species of Leptodactylus rather than to
Pseudis.) The streamlining of the muscular body and a very slippery mucus
made them extremely difficult to grasp. One of the Pailon males was in
amplexus when caught. The female escaped.

Leptodactylus bufonius Boulenger, 1894.

5 One km. east of El Carmen, Feb. 25-26. C.M. 36187, M.C.Z. 29962-5.

7 El Carmen, Feb. 27. C.M. 36156-60, 36187, M.C.Z. 29966.

14 San Jose, March 3-4. C.M. 36229 (7), M.C.Z. 29967-73.

22 El Pailon, March 5. C.M. 36091 (10), 36107, M.C.Z. 29974-84.

The San Jose and Pailon specimens were calling from open places near
grassy spots close to pools and puddles. They have a whistling call.

Others were collected on the road-bed at night, and the six Carmen speci-
mens occupied the openings of as many drainage holes in the retaining wall
of Carmen station (See Fig. 5). Every hole contained a specimen of this
species, a tarantula, or both. These frogs are very wary and excellent jumpers.
If surprised in the open they will move off with a series of three to six meter-
long jumps. The third of these jumps is always at an angle of 60 to 90 de-
grees to the preceding one.

Besides the more usual secondary sex characters this species exhibits strik-
ing differences in the shape of the rostrum. The males show a sharp spatulate
edge extended forward to present a profile view of a straight line from eye
to rostral tip. The rostral edge shows strong projection beyond the tip of
the nose. The females have the top of the head descending sharply and the
rostrum bluntly rounded. The occurrence of various secondary sex characters
in this and related genera might prove to be a profitable subject for in-
vestigation.

Leptodactylus mystaceus (Spix), 1824.

1 Corumba, Feb. 23. C.M. 36162.

2 1.5 km. east of El Carmen, Feb. 25-26. C.M. 36097, M.C.Z. 29985.

4 San Jose, March 4. C.M. 36117-8, M.C.Z. 29987-8.

The Corumba specimen was under a log, the east of Carmen specimens at
the edge of a pond, while the San Jose specimens came from buildings of a
church compound into which they had strayed.

Leptodactylus mystacinus Burmeister, 1861.

2 El Carmen, Feb. 27. C.M. 36247, M.C.Z. 29986.

The specimens were found sitting in the short grass before the retaining
wall at the station.

Leptodactylus ocellatus (Linnaeus), 1758.

3 Corumba, Feb. 22-23. C.M. 36154, 36182, M.C.Z. 29989.

24 1.5 km. east of El Carmen, Feb. 25-26. C.M. 36215 (11), M.C.Z. 29990-
30002.

2 El Carmen, Feb. 27. C.M. 36214, 36275.

1960

Gans: A Herpetological Collecting Trip in Bolivia

305

5 Robore, March 1. C.M. 36134-6, M.C.Z. 30003-4.

1 El Porton, March 2. C.M. 36099.

51 San lose, March 3-4. C.M. 36204-6, 36242-4, 36251-70, M.C.Z. 30005-13
(+16).

1 Tunas, March 5. M.C.Z. 30013.

13 El Pailon, March 5. C.M. 36167 (7), M.C.Z. 30014-9.

None of these frogs was observed while it was calling nor were pairs
found in amplexus, which seems curious in view of the fact that many speci-
mens were observed and that most other species in the same region appeared
to be breeding. The specimens were collected in a variety of habitats, in
roads, at the margin of the Paraguay River, along the margins of ponds and
pools, etc.

Leptodactylus pentadactylus labyrinthicus (Spix), 1824.

1 Reservoir, 1 km. north of El Carmen, Feb. 26. C.M. 36166.

I El Porton, March 2. M.C.Z. 30135.

1 juv. ? El Porton, March 2. C.M. 36188.

1 El Pailon, March 5. C.M. 36170.

Two adult specimens were seen at El Porton, both in well protected situ-
ations. The first one was noted close to the entrance of a deep crevice in
the rock wall out of which water was running. It escaped back into the wall
before it could be caught or shot. The second one was sitting in the entrance
of a drain-pipe, which discharged into a concrete drainage canal. The frog
was caught after the pipe behind it had been blocked. The Pailon specimen
was floating at the edge of a deep pool. Several other specimens were also
seen here, but these always dived before I could get within several meters.
These large frogs appear to have good eyesight. They are very wary and
select well protected calling sites.

This species is supposed to have a toxic secretion. One of the persons with
whom I collected in Brazil would start to sneeze as soon as he handled a
specimen. The active substance must have different effects on different per-
sons as I neither suffered from dermatitis nor had any other symptoms at
any time.

Leptodactylus p. podicipinus (Cope), 1862.

1 Corumba, Feb. 22. C.M. 36165.

5 1.5 km. east of El Carmen, Feb. 26. C.M. 36234-5, M.C.Z. 30020-2.

2 El Carmen, Feb. 27. C.M. 36236-7.

8 Three km. south of Robore, March 1. C.M. 36238-41, M.C.Z. 30023-6.

2 El Porton, March 2. C.M. 36248, M.C.Z. 30027.

5 San Jose, March 3. C.M. 36119-20, 36164, M.C.Z. 30028-9.

1 Tunas, March 5. M.C.Z. 30136.

All but the specimens from Carmen and vicinity were collected near water.
The Corumba specimen was sitting in the open at the edge of a roadside
pool, while the Robore specimens were partially sheltered by small logs and
the Porton specimens were found under rocks. These specimens from east of
Carmen were collected in a drying cattle wallow, those from Carmen proper
in the grass before the station wall. The San Jose specimens were found on
the floor of buildings of the church compound into which they seemed to have
strayed inadvertently. Many desiccated corpses in building corners indicated
that the few steps to the outside must form a formidable obstacle to their

306

Annals of Carnegie Museum

vol. 35

escape. The Robore and Porton specimens seemed to be calling from hiding.

The single poorly preserved juvenile (M.C.Z. 30136) from Tunas was tenta-
tively assigned to L. podicipinus petersi (Steindachner) by Shreve.

Muller and Hellmich (1936) refer to this form as L. caliginosus Girard.

Leptodactylus sibilatrix (Wied), 1824.

4 Corumba, Feb. 22-23. C.M. 36231-2, M.C.Z. 30030-1.

15 San Jose, March 3-4. C.M. 36230(8), M.C.Z. 30032-8.

The Corumba specimens were sitting near roadside puddles at the edge of
the Rio Paraguay. Others were found in open grassy spots. The San Jose
specimens also called from grassy areas, but the specimens would always be
partly hidden under tufts of grass so that they were difficult to see even where
the grass was cropped short. The species gave its whistling call from both
wet and dry places.

Physalaemus cuvieri Fitzinger, 1826.

9 Corumba, Feb. 21-23. C.M. 36124-31, 36163.

2 Reservoir, 1 km. north of El Carmen, Feb. 26. C.M. 36116, M.C.Z. 30039.
70 El Carmen, Feb. 27. C.M. 36142 (35), M.C.Z. 30040-74.

2 Robore, Feb. 28. C.M. 32122, M.C.Z. 30075.

2 El Porton, March 2. C.M. 36208, M.C.Z. 30076.

3 El Pailon, March 5. C.M. 36249-50, M.C.Z. 30077.

These frogs were abundant and could be heard calling every night, and in
certain regions during the day as well. (The Carmen specimens were col-
lected while they were calling during the afternoon.) In spite of this it was
difficult to obtain adequate series because of their calling habits.

The males call from shallow water (1-3 cm. deep) at or near the edge of
larger bodies of water. They sit almost always in small open spots, but
slightly larger than their inflated body, and shaded by tufts of grass. When
calling the skin is stretched to translucence so that it becomes difficult to
recognize them even with direct vertical illumination. Their call is a whine
resembling the protracted meow of a cat.

A large series of juveniles was found sitting in the short grass below the
stone wall at El Carmen station. Others were noted at the entrances to the
drainage holes which they shared with specimens of a large tarantula spider
and adult Leptodactylus bufonius.

Physalaemus fuscomaculatus (Steindachner), 1863.

6 1.5 km. east of El Carmen, Feb. 25-26. C.M. 36209-11, M.C.Z. 30078-80.

56 El Carmen, Feb. 27. C.M. 36133 (28), M.C.Z. 30081-108.

8 Robore, Feb. 28. C.M. 36093-6. M.C.Z. 30109-12.

1 San Jose, March 4. C.M. 36123.

The form was found calling at Robore. The males were calling from the
floating position in the center of a small stream pool. They were very strongly
inflated, puffed into a ball with the major portion of the body riding above
the water.

The call is a loud whaam-ump with the accent on the second note.

Some of the specimens from east of Carmen were collected while they were
sitting on the road-bed at night, others in a drying cattle wallow. The juve-
niles of this species and of P. cuvieri formed the greatest part of the large
number of small frogs collected in the grass before the station wall at El
Carmen.

1960

Gans: A Herpetological Collecting Trip in Bolivia

307

Eupemphix natteri Steindachner, 1863.

2 El Porton, March 2. C.M. 36098, M.C.Z. 30113.

These specimens were collected while calling from under logs lying on
moist ground in the station clearing at the southern end of the pass. Their
call is a booming sound repeated several times.

Elachistocleis ovalis (Schneider), 1799.

1 Corumba, Feb. 22. C.M. 36177.

3 One km. east of El Carmen, Feb. 25. C.M. 36198-200.

19 El Carmen, Feb. 27. C.M. 36190-7, M.C.Z. 30114-24.

4 El Pailon, March 5. C.M. 36245-6, M.C.Z. 30125-6.

Only a single specimen was calling and only a single pair was noted in
amplexus. Both were found at Pailon. The frog that appeared to be call-
ing was positioned at the base of a dense reed thicket at the edge of a pool.
It had almost completely left the water, only the toes of the hind feet touch-
ing the surface. The hind legs and body were extended vertically upward
into a straight line and the fore limbs touched the reed stem. The call was
a very high thin piping whistle, almost completely masked by the clamour
of the dozen other species calling within 10 meters.

The pair in amplexus was resting on the floating leaf of a water plant
in the center of the same pool.

All other specimens were collected at night in a number of semi-covered
sites. The Corumba specimen was crouching under a piece of rock in the
center of an open road, while the east of Carmen specimens were crouching
near lumps of rock or pieces of manure on the track, possibly waiting for
passing insects.

The largest series was taken at the base of the rock wall at Carmen station.
They were sitting between the tufts of closely cropped grass, and when dis-
turbed would flee not by hopping off but rather by burrowing, winding their
way between the interlaced network of stems and exposed roots as if lubri-
cated. Certain specimens appeared to protrude from tunnels leading to the
grass roots.

The color in life is greyish, with prominent but irregularly shaped bright
red spots on the sides of the body and legs. Only a single specimen (M.C.Z.
30124) lacked these red spots and demonstrated instead a light gray mid-
dorsal line that continued onto the back of each hind-limb.

Hypopachus miilleri (Boettger), 1885.

2 Three km. south of Robore, March 1. C.M. 36189, M.C.Z. 30127.

15 El Pailon, March 5. C.M. 36216(8), M.C.Z. 30128-34.

The Pailon males were calling from the floating position in track-side
pools. Their bodies were enormously inflated, giving them a circular outline
in plan view, broken only by the head which retains its original form. When
disturbed they would deflate and sink below the surface of the muddy pool.
They were very easy to collect as they did not change their position when
diving, but just sunk vertically to the muddy bottom of the pool. Three pairs
were collected in amplexus.

The call is a loud protracted snore reminiscent of a foghorn.

The two juvenile Robore specimens were collected at night under the
bark of fallen logs. They inflated themselves when picked up.

308

Annals of Carnegie Museum

vol. 35

CROCODILIA

fCaiman sp.

Two small (0.75-1.00 m.) specimens were seen from the train, one near
Naranjos and the second before San Jose. Another slightly larger specimen
was 'seen at night, while I was wading in the slough three km. south of
Robore. Though apparently stunned after being shot, it escaped into deep
water before it could be reached.

TESTUDINATA

Testudo sp.

Fragments of a burned shell were found at the Tunas siding.

LACERTILIA

Phyllopezus pollicaris przewalski Koslowsky, 1895.

(Lagarto domestico, Escorpion)

5 El Carmen, Feb. 27. C.M. 35856-7, D.Z. (3).

1 juv. 3 km. south of Robore, Feb. 28. D.Z. (1).

2 El Porton, March 2. C.M. 35858-9.

4 San Jose, March 3. C.M. 35860-1, D.Z. (2)

3 San Jose, 1954, (M. Bottega). A.M.N.H. 74730, C.M. 34839-40.

These harmless lizards are held in considerable awe by some of the local
people who believe that their bite is highly venomous. With the exception
of the juvenile Robore specimen picked up at night under the bark of a
stump and of the two Porton adults collected from under the bark of a fallen
trunk during the day, all specimens were collected in or near buildings. They
were fairly numerous on the walls of the railroad shed in Carmen and many
more were seen than collected on the walls of the old church in San Jose.
The lizards seem to be strictly nocturnal.

Since the racial characters differentiating the two subspecies (Vanzolini,
1953) are those of mature specimens, the juvenile can not be definitely as-
signed. A check of other specimens of this species in U. S. museums dis-
closed a record for P. p. przewalski from Villa Montes, Bolivia, (M.C.Z. 28632)
and one for the nominate race from Rio Reto, Brazil (M.C.Z. 3132). Both of
these records also agree with the scheme proposed by Vanzolini.

Teius teyou (Daudin), 1802.

16 Corumba, Feb. 24. C.M. 35879-86, D.Z. (8).

3 El Carmen, Feb. 26-27. C.M. 35887-8, D.Z. (1).

2 Robore, March 1. C.M. 35889, D.Z. (1).

These lizards were fairly common in many open bushy areas. The fact that
none was collected on the western leg of the trip possibly reflects the rainy
weather rather than absence of the species.

Tropidurus spinulosus (Cope), 1862.

2 Reservoir, 1 km. north of El Carmen, Feb. 26. C.M. 35874, D.Z. (1).

1 El Carmen, Feb. 27. C.M. 35878.

5 San Jose, March 4. C.M. 35875-7, D.Z. (2).

3 San Jose, 1954, (M. Bottega). C.M. 34836-8.

The Carmen reservoir specimens were shot during the day while they were
sitting approximately six meters above the ground on the trunks of trees;
the Carmen juvenile was taken under the bark of a log in the station at night.

The form was abundant on the buildings of the church compound in San
Jose and was observed while it was hunting on the roof during the day.

1960

Gans: A Hf.rpetological Collecting Trip in Bolivia

309

Tropidurus torquatus (Wied), 1820.

9 El Carmen, Feb. 27. C.M. 35863-8, D.Z. (3).

2 Naranjos, Feb. 28. C.M. 35869, D.Z. (1).

7 El Porton, Feb. 3. C.M. 35870-3, D.Z. (3).

Part of the Porton series was collected on rock walls during the day; all
others were on piles of lumber and railroad ties. The lizards were always
found close to the ground and were strictly diurnal.

Mabuya frenata frenata (Cope), 1862.

1 El Carmen, Feb. 27. C.M. 35862.

Shot on woodpile in Carmen station. A second specimen was seen in a
similar location, 0.5 km. north of the village, but escaped.

OPHIDIA

Boa enydris (Linnaeus), 1754.

1 $ El Porton, 1953, (O. O. Nogueira). C.M. 34845.

Snout-vent plus tail length of the specimen is 615+175 mm.

Chironius pyrrhopogon (Wied), 1825.

2 El Pailon, March 5. C.M. 34823-4.

These hatchling specimens were coiled at the ends of small branches in
a bush overlooking a track-side pool at night. One disgorged a juvenile
Phyllomedusa (M.C.Z. 29680).

They were identified by Dr. J. R. Bailey who states that “they are very
close to pyrrhopogon of my latest paper (1955). They differ essentially only
in having 29 and 30 maxillary teeth, which is high but not unexpected as
can be seen from the truncated frequency distribution. I have not worked
over Bolivian material which perhaps averages higher.”

Dryophylax pallidus (Linnaeus), 1758.

1 $ San Jose, 1954, (M. Bottega), C.M. 34820.

This specimen has 148 ventrals, 56/56 caudals, 19 dorsals, 7 upper labials
and a divided anal. The scales are faintly keeled. Snout-vent plus tail length
is approximately 420+105 mm.

Elapomorphus tricolor Dumeril and Bibron, 1854.

2 (1 juv., 19) San Jose, 1954, (M. Bottega). C.M. 34846-7.

The two specimens have 208 and 199 ventrals respectively; both have di-
vided anals and 15 dorsals at midbody and their tails are incomplete. Snout-
vent lengths are 198 and 545 mm.

Erythrolamprus venustissimus (Wied), 1821.

1 9 El Porton, March 3. C.M. 34835.

The snake was picked up by a track walker while it was crossing the road-
bed in the early morning. The specimen has 198 ventrals, 33/33 caudals, a
divided anal and 15 midbody dorsals. Snout-vent plus tail length is 628 + 65
mm.

The head is black. A white band crosses behind the eye, and produces a
mottling of the parietals. The body is red with black bands, each divided
into two by a central white band. All dorsal scales show a tendency toward
black pigmentation of their tips and posterior edges. The nuchal collar is
single and black. It extends over the first six vertebrals.

310

Annals of Carnegie Museum

vol. 35

hnantodes cenchoa (Linnaeus), 1758.

1 $ El Porton, 1953, (O. O. Nogueira). C.M. 34819.

The specimen has 243 ventrals, 141/141 caudals, a divided anal and 17
midbody dorsals. Snout-vent plus tail length is 280+110 mm.

Leimadophis iyphlus (Linnaeus), 1758.

1 9 D.O.R., road Corumba to Mt. Urucum, Feb. 23. C.M. 34841.

3 (19,2 juv.) San Jose, 1954, (M. Bottega). C.M. 34842, 34821-2.

The juveniles have a clear upper lip, while there is some invasion of the
dorsal color in adults. The smallest specimen still has the black nuchal collar.
This is followed posteriorly by a lighter band, and some black speckling on
the back. The other juvenile has the dorsal dark bordered. The D.O.R.
adult was a striking blue-green clorsally before preservation. The venter is
clear in all specimens. The adult coloration is very similar to that of a male
specimen from “Santa Cruz de la Sierra” (M.C.Z. 11860, J. Steinbach coll.)
identified as L. t. brachyura (Cope) by Shreve.

Scale counts and measurements are:

Sex

Ventrals plus
caudals

Dorsals

Anal

Snout-vent plus
tail length

C.M. 34841

9

167 + 53/53

19

D

544+125

C.M. 34842

9

169 + x

19

D

580+x

C.M. 34821

juv

149 + 48/48

19

D

240 + 47

C.M. 34822

juv

x + x

19

D

125 + 22

M.C.Z. 11860

a

165 + 54/54

19

D

412 + 91

Leptodeira annulata pulchriceps Duellman, 1958.

2 San Jose, 1954, (M. Bottega). C.M. 34103-4.

These two specimens were sent to Dr. W. Duellman who was revising the
genus at the time. They form part of the series upon which this new race was
established (1958).

Lystrophis semicinctus (Dumeril and Bibron), 1854.

1 juv. El Carmen, Feb. 27. C.M. 34829.

2 juv. El Porton, 1953, (O. O. Nogueira). C.M. 34830-1.

1 juv. San Jose, 1954, (M. Bottega). C.M. 34832.

The Carmen specimen was collected under a rock near the village.
These four juveniles show a coloration similar to that of several adults
(M.C.Z. 11856, 29038, 29255-8) collected by the elder Steinbach in Santa Cruz
and Buena Vista respectively. The entire series shows decided individual vari-
ation in the degree of ventral melanism. There also seems to be an in-
creasing north-south cline in ventral counts through this region. The speci-
mens furthermore appear to show sexual dimorphism in tail length and
caudal counts. The extent of this may be estimated from the table.

Sex

Ventrals plus
caudals

Dorsals

Anal

Snout-vent plus
tail length

C.M. 34829

$

151 + 38/38

21

D

182 + 28

C.M. 34830

9

143 + 30/30

21

D

163 + 23

C.M. 34831

9

151 + 30/30

21

D

172 + 22

C.M. 34832

S

154+40/40

21

D

177 + 28

M.C.Z. 11856

S

151 + 37/37

21

D

412+28

M.C.Z. 29038

9

154 + 26/26

21-19

D

370+45

Upper labials always 8.

1960

Gans: A Herpetological Collecting Trip in Bolivia

311

Oxyrhopus rhombifer Dumeril and Bibron, 1854.

1 El Carmen, Feb. 26. C.M. 34826.

1 El Porton, March 3. C.M. 34827.

1 San Jose, 1954, (M. Bottega). C.M. 34825.

The Carmen and Porton specimens were found under logs near the track.
The Carmen specimen contained remnants of two lizards, apparently Tropi-
durus.

The identification is that of Dr. J. R. Bailey, who states (in litt.) that the
specimens appear to be intermediate between O. r. inequifasciatus Werner
and O. r. septentrionalis Vellard.

? Philodryas psammophideus Gunther, 1872.

1 $ San Jose, 1954, (M. Bottega). C.M. 34851.

This small and very poorly preserved specimen agrees fairly well (except
as noted) with Boulenger’s description (1896, p. 132) of this rare species,
previously recorded from Uruguay and northern Argentina only. The verte-
bral stripe is indistinct on the body and is not edged with black, and the
lateral stripe is not apparent though traces may be seen on the cheek. There
is a considerable amount of black speckling under the chin and in the gular
region.

The specimen has 198 (± 4) ventrals, 90/90 caudals, a divided anal, 19
scale rows at midbody and 8 (45) / 9 (56) upper labials. Snout-vent plus tail
length is 305 + 90 mm.

Pseudoboa guerini Dumeril and Bibron, 1854.

1 juv. San Jose, 1954, (M. Bottega). C.M. 34844.

This (alcoholic) specimen still has the juvenile color pattern. The snout
is dark brown to a point just posterior to the eyes. A whitish-yellow band
crosses the nuchal region, including the parietals and the first six to seven
vertebrals. Beyond these the dorsum is dark brown, starting to lighten
caudad of the first quarter of the body. The sides of the neck are dark, but
the brown color begins to break up with the third, fourth, fifth and sixth
rows of dorsals being irregularly invaded by lighter color. The chin and
venter are a dirty yellow-white.

The specimen has 200 ventrals, 94 caudals, 19 dorsals, a single anal and
8 upper labials. Snout-vent plus tail length is 256 + 77 mm.

Rhadinea genimaculata (Boettger), 1885.

2 San Jose, 1954, (M. Bottega). C.M. 43852-3.

These two specimens proved most difficult to identify because of the pres-
ence of an incomplete light band across the nuchal region. Such a band
was not present on the nape of the only other specimen (M.C.Z. 3000,
“Brazil” Thayer Expedition) initially available for comparison, nor was it
mentioned in descriptions of this species.

Examination of five additional specimens in the Museum of Zoology of
the University of Michigan, indicated what appears to be a geographical
gradient in this character. The band was absent in U.M.M.Z. 108766 from
Barreiras, Bahia and U.M.M.Z. 108765 from Pirapora, Minas Gerais, both in
Brazil. A light spot was present on each side of the nuchal region between
the vertebral and lateral bands of U.M.M.Z. 108764 from Porto Esperan^a,
Mato Grosso, Brazil, while two further Steinbach specimens (U.M.M.Z. 64549)

312

Annals of Carnegie Museum

vol. 35

from Santa Cruz, Bolivia, agreed with the San Jose specimens in showing the
incomplete nuchal bands.

The San Jose specimens, a male and a female respectively, have 196 and
194 ventrals, 45/45 and 58/58 caudals, divided anals, 17 dorsal scale rows,
and 8(45) upper labials. Snout-vent plus tail length are 341 + 69 and 328 + 83
mm.

The tails of both specimens appear superficially blunted in that they are
fairly stout and then suffer a sharp reduction near their ends. This is also
the case in the M.C.Z. and U.M.M.Z. specimens and appears to be a char-
acteristic of this (and possibly several other related) species.

Sibynomorphus turgidus (Cope), 1868.

1 $ 3 km. south of Robore, March 1. C.M. 34833.

1 $ San Jose, 1954. (M. Bottega). C.M. 34834.

The Robore specimen was caught at night while it was crossing a sandy
road in low scrub forest near a flooded slough.

The specimens were identified by Dr. James A. Peters.

The male specimen has 160 ventrals, 54/54 caudals, a single anal, and 15
dorsals. Snout-vent plus tail length is 269 + 65 mm. The ventral count of
the female can not be determined. The specimen has 45/45 caudals, a single
anal, 15 dorsals and snout-vent plus tail length of 341 + 69 mm.

Xenodon merremi (Wagler), 1824.

2$ San Jose, 1954, (M. Bottega). C.M. 34817-8.

Counts of these two specimens are ventrals 148, 147, anals divided, caudals
37/37 both, and dorsals 19-19-17 and 19-19-15. Snout-vent and tail lengths are
778+116 and 705+114 mm.

Xenodon severus (Linnaeus), 1754.

1 juv. El Porton, 1953, (O. O. Nogueira). C.M. 34843.

This juvenile specimen shows the typical coloration described by Boulenger
(1894, p. 149). There are nine light dorsal bands from nape to anus. The
scutellation is: ventrals 136, divided anal, 36/36 caudals, 21 dorsals, and 8
upper labials. Snout-vent plus tail length is 208 + 29 mm.

Bothrops neuwiedi Wagler, 1824.

1 $ San Jose, 1954, (M. Bottega). C.M. 34828.

The anterior portion of this snake is severely damaged.

REFERENCES

Ahlfeld, Federico

1946. Geologia de Bolivia. Revista del Museo de La Plata (Nueva
Serie), Seccion Geologia, v. 3, p. 5-370.

Anonymous

1941. Normais climatologicas. Servigo de Meterologia, Servi^o de
Informa^ao Agricola, Ministerio da Agricultura, Rio de Ja-
neiro, 176 p.

Bailey, Joseph R.

1955. The snakes of the genus Chironius in southeastern South
America. Occasional Papers of the Museum of Zoology, Uni-
versity of Michigan, no. 571, 21 p.

1960

Gans: A Herpetological Collecting Trip in Bolivia

313

Bechyne, Jan

1956. Reise des Herrn G. Frey in Siidamerika: Galerucidae (Col.
Phytophaga). Entomologische Arbeiten des Museums G. Frey,
v. 7, no. 1, p. 241-358.

1958. Notizen zu den neotropischen Chrysomeloidea (Col. Phy-
tophaga). Entomologische Arbeiten des Museums G. Frey, v. 9,
no. 2, p. 478-706.

Boulenger, George Albert

1894. Catalogue of the snakes in the British Museum (Natural His-
tory). London, v. 2, 382 p.

1896. Catalogue of the snakes in the British Museum (Natural His-
tory). London, v. 3, 727 p.

Cochran, Doris Mable

1954. Frogs of southeastern Brazil. Bulletin of the United States
National Museum, no. 206, 423 p.

Cott, Hugh

1940. Adaptive coloration in animals. London, 508 p.

(1957 reissue) .

Duellman, William E.

1958. A monographic study of the colubrid snake genus Leptodeira.

Bulletin of the American Museum of Natural History, v. 114,
art. 1, p. 1-152.

Duguid, Julian

1931. Green hell, adventures in the mysterious jungles of eastern Bo-
livia. Century Company, New York, 339 p.

Fonseca, Olympio Oliveira Ribeiro da

1929. Diario de uma viagem ao oriente da Bolivia, de Porto Esper-
an^a a San Jose de Chiquitos. Memorias do Instituto Oswaldo
Cruz, Rio de Janeiro, v. 22, p. 175-222.

Funkhauser, Anne

1957. A review of the neotropical tree-frogs of the genus Phyllome-
dusa. Occasional Papers of the Natural History Museum of
Stanford University, no. 5, p. 1-90.

Gallardo, Jose

1957. Las subspecies Argentinas de Bufo granulosus Spix. Anales del
Museo Argentino de Ciencias Naturales “Bernardino Riva-
davia”, v. 3, no. 6, p. 337-74.

Gans, Carl

1956. Frogs and paradoxes. In Animaland, Staten Island Zoological
Society, New York, 4 p.

Goin, Coleman Jett

1957. Status of the frog genus Sphoenohyla with a synopsis of the
species. Caldasia, v. 8, no. 36, p. 11-31.

Haseman, John D.

1912. Some factors of geographical distribution in South America.

Annals of the New York Academy of Science, v. 22, p. 9-112.

314

Annals of Carnegie Museum

vol. 35

Herzog, Th.

1913. Die von Dr. Th. Herzog auf seiner Reise durch Bolivien in den
Jahren 1910 und 1911 gesammelten Pflanzen. Teil 1. Mede-
deelingen van ’s Rijks Herbarium Leiden, no. 19, p. 1-84.

Lane, Federico

1956. Cerambycoidea Neotropica nova IV (Coleoptera). Dusenia,
v. 7, no. 1, p. 1-32.

Lizer, Carlos A.

1919. Informe sobre la expedition al Chaco boliviano. Boletin del
Ministerio de Agricultura de la Nacion, Buenos Aires, v. 24,
no. 1, p. 1-47.

Magarinos Torres, Francisco Eugenio and Armando Mortera

1948. Atlas pluviometrico do Brasil (1914-38). Secciao de Hidrologia,
Divisao de Aguas, Ministerio da Agricultura, Rio de Janeiro,
bull. 5.

Mann F., Guillermo

1951. Esquema ecologico de selva, sabana y cordillera en Bolivia.

Instituto de Geografia, Universidad de Chile, pub. no. 3, 236 p.

Muller, Lorenz and Walter Hellmich

1936. Wissenschaftliche Ergebnisse der Deutschen Gran Chaco-Ex-
pedition. Amphibien und Reptilien. Teil 1. Amphibia, Che-
lonia, Loricata. Stuttgart, 120 p.

Osborne, Harold

1954. Bolivia, a land divided. Royal Institute of International Affairs,
London, 144 p.

Parodiz, Juan

1958. A new Bolivian land snail of the genus Drymaeus. Breviora
no. 95, p. 1-3.

Pena, Abel and Lillo Escobar

1947. Sintesis geografica de Bolivia. La Paz, 416 p.

Shreve, Benjamin

1959. A new Phyllomedusa from Bolivia. Breviora no. 113, p. 1-6.

Travassos, Lauro, and others

1957. Excursao cientifica realizada nas zonas das Estradas de Ferro
Noroeste do Brasil e Brasil-Bolivia em janeiro e fevereiro de
1955. Publica^oes avulsas do Museu Nacional, Rio de Janeiro,
no. 20, 44 p.

Vanzolini, Paulo Emilio

1953. Sobre o genero Phyllopezus Peters (Sauria, Gekkonidae) . Pa-
peis Avulsos do Departamento de Zoologia de Secretaria da
Agricultura, Sao Paulo, v. 11, no. 22, p. 353-69.

POPULAR SUMMARY

Contrary to popular belief, many parts of the
world are still incompletely known zoologically.

This publication is a report upon a fourteen-day
collecting trip in 1954 through a little known region
of eastern, or Amazonian, Bolivia, along the right-of-
way of the newly constructed Brazi J-Bo 1 ] via Railway,

In addition to information on the physiography,
climate and localities in the area traversed, detailed
reports are given upon the 53 kinds of amphibians and
reptiles represented in the collection of about 900
specimens obtained by the two-man party. Other mater-
ials brought back include a large collection of in-
sects and a number of batches of shells. The collections
were divided between several i nst i tut ions wi th most of
the specimens of amphibians and reptiles going to the
Carnegie Museum of Pittsburgh, Pennsylvania,

Additional copies of this publication may be obtained
from Carnegie Museum, Pittsburgh 13, Pa., for $.95
each plus ,04 tax for Pennsylvania residents.

r

Z> 7. 7 3
7f PCn-2*

?

ART. 14. LATE PLEISTOCENE RECORDS OF THE YELLOW-
CHEEKED VOLE, MICROTUS XANTHOGNATHUS (Leach)

By John E. Guilday* and Martin S. BenderJ

The yellow-cheeked vole, Microtus xanthognathus (Leach), has been re-
ported from numerous, widely scattered localities in the subarctic of western
North America, from about 52° N. latitude in Alberta, north to the Mac-
kenzie River delta, and from the Kuskokwim River in western Alaska, east to
Fort Churchill and the Nelson River in Manitoba. It is apparently very
local in distribution and is not often taken by collectors.

This is one of the largest species in the genus Microtus closely approach-
ing the size of the water vole, Microtus richardsoni (DeKay). It is a colonial
animal, an extensive burro wer, and one of the rarest North American micro-
tines.

Skeletal remains referred to this species have been recovered from two
Late Pleistocene cave deposits in the Appalachian Mountain region, one in
south-central Pennsylvania and the other in west-central Virginia. These
constitute the first record of the animal from the Pleistocene of eastern North
America. (Fig. 1.)

An additional Pleistocene record may be a specimen recovered from perma-
frost deposits at Chicken, Alaska. The specimen, a mummy with hair, is in
the American Museum of Natural History. In the opinion of the collector,
O. W. Geist, who noted several deposits of vivienite on it, the specimen is
a true fossil. We wish to thank Dr. Richard G. Van Gelder for calling our
attention to this specimen.

The first site to yield remains of this mouse was Natural Chimneys, Mt.
Solon, Augusta County, Virginia, 38° 20' N. lat., 79° 5' W. long., alt. 1500
feet, Stanton Quadrangle, Va.-W.Va. U. S. Geological Survey topographical
map, 1894 edition. The Natural Chimneys, located on the western side of
the Shenandoah Valley, are a group of natural limestone pillars with a shal-
low cave at their base. The pillars seem to be erosion remnants of a former
fissure system. Bones and teeth were recovered from several spots in the floor
and walls of the cave by Ted B. Ruhoff and later by Carnegie Museum field
parties under the direction of J. LeRoy Kay. Within historic times a stream
flowed about the base of the “chimneys,” and the bone-bearing deposit may
have been reworked by water since its initial deposition. The fauna is prob-
ably derived from several time periods but none is believed older than
late Pleistocene. The collection is unstudied but contains at least the fol-
lowing forms: Condylura, Scalopus, Sorex, Microsorex , Blarina, Cryptotis,
Myotis, Eptesicus, Pipistrellus , Mephitis , Mustela , Lynx, Ursus, Marmota,
Tamias, Glaucomys , Sciurus, Tamiasciurus, Neotoma, Peromyscus, Synaptomys,
(. Mictomys ), Phenacomys, Clethrionomys, Microtus, Pitymys, Ondatra, Castor,
Castoroides, Zapus, Napaeozapus, Lepus, Odocoileus ?, Mylohyus ?, plus
fragments of birds, snakes (including a large Crotalus near adamanteus ver-
tebra), amphibians and fish. All material is fragmentary. The Castoroides,
one upper molar, was discovered by Mr. Ruhoff. Six specimens collected by
the Carnegie Museum field parties were identified as Microtus cf. xantho-
gnathus (Leach). These include two broken right mandibles containing the

* Assistant Curator of Comparative Anatomy, Carnegie Museum.

•{■Instructor in Geology, University of Pittsburgh.

315

Submitted for publication, September 21, 1959
Issued June 10, 1960.

316

Annals of Carnegie Museum

vol. 35

first two molars, 1 broken left mandible with ml-m2, 1 broken left mandible
containing the stumps of ml-m2, 1 left maxilla with complete dentition and
1 palate with M1-M2 on both sides, all catalogued under C.M. 5861.

Fig. 1. Outline map of North America. A. Present range of Microtus
xanthognathus (Leach). (Modified from Hall 8c Kelson, 1959.) B. Lloyd’s
Rock Hole, New Paris, Bedford County, Pennsylvania. C. Natural Chimneys,
Mt. Solon, Augusta County, Virginia

1960

Guilday and Bender: Yellow-cheeked Vole

317

The second locality is “Lloyd’s Rock Hole” or Sink-hole no. 4, li/2 miles
northeast of New Paris, Bedford County, Pennsylvania, 40° 7' N. lat., 78° 37'
W. long., alt. 1500 feet. Bedford Quadrangle, U. S. Geological Survey, 1908
edition. This highly fossiliferous fissure deposit was discovered in the spring
of 1958 by Carnegie Museum and National Speleological Society field parties
composed of Rita Battistoli, Ralph C. Bossart, Gerald W. Frederick, Roscoe
Hall, Harold Hamilton, Beverly Hofecker, J. LeRoy Kay, John A. Leppla,
Allen D. McCrady, and other members of the Pittsburgh Grotto of the N. S. S.

Excavation is not complete at this writing but the fauna contains at least
the following mammalian forms: Condylura, Parascalops, Sorex, Microsorex,
Blarina, Myotis, Eptesicus, Pipistrellus, Martes, Mustela, Marmota, Tamias,
T amiasciurus , Glaucomys, Neotoma, Peromyscus, Synaptomys, (Mictomys),
Phenacomys, Microtus, Pitymys, Clethrionomys, Zapus, Napaeozapus, Erethi-
zon, Mylohyus, and Odocoileus ?. The material is both abundant and well
preserved. The entire fauna appears to have been contemporaneous with
possible intrusions of a few modern burrowing forms. Pollen analysis of the
clay matrix by Paul S. Martin, Geochronology Laboratories, University of
Arizona, indicates a boreal, pine-dominant forest cover.* This fauna is cur-
rently under study and will form the topic of a future paper.

Microtus xanthognathus (Leach) was the second most common microtine
in the deposit, exceeded only by Clethrionomys cf. gapperi (Vigors). Remains
referred to the yellow-cheeked vole include, to date, 21 partial skulls, 24
right mandibles, 23 left mandibles, and 42 additional bones representing most
of the major skeletal elements. Remains of this mouse occurred throughout
the excavation from four feet beneath the surface to the present excavation
level some 20 feet beneath the mouth of the sink-hole. It was apparently in
temporal association with Microtus pennsylvanicus (Ord) and Microtus chro-
torrhinus (Miller) both of which are represented by well preserved skulls
and mandibles.

A comparative study of the New Paris fossil M. xanthognathus material and
modern crania of various ages and localities was made. Results are pre-
sented below.

The authors wish to thank Dr. J. Kenneth Doutt, Dr. Claude L. Hibbard,
and Dr. J. LeRoy Kay for much valuable advice and assistance; the members
of the Pittsburgh Grotto, National Speleological Society for their unstinting
work on the New Paris excavations; John A. Leppla for the photography,
Harry K. Clench and Dr. Wendell P. Rand for mathematical assistance and
advice; and Caroline A. Heppenstall, Alice M. Guilday and Patricia M.
Guilday for editing, clerical assistance and art work. We would also like to
thank the following people who loaned us specimens under their care: Dr.
Barbara Lawrence, Museum of Comparative Zoology at Harvard; Dr. Richard
G. Van Gelder, American Museum of Natural History; Dr. David H. John-
son, United States National Museum; Dr. Austin W. Cameron, Canadian
National Museum; Dr. L. J. Rowinski, University of Alaska Museum. We
also wish to thank the two men who, by their co-operation and generosity,
were most responsible for this project’s existence, Mr. Gordon E. Brown,
owner of Natural Chimneys, Mt. Solon, Virginia, and Mr. Oscar Miller of

* A radiocarbon date of 11,300 ±1000 years run on carbon particles (forest fires?) in
the matrix has just been received from Yale University Geochronometric Laboratory.

318

Annals of Carnegie Museum

vol. 35

New Paris, Pennsylvania, owner of Lloyd’s Rock Hole. Excavations at the
latter site are being financed by a grant from the A. W. Mellon Educational
and Charitable Trust, and conducted by field parties from Pittsburgh Grotto,
National Speleological Society and Carnegie Museum.

All measurements of 10 mm. or less were taken with an ocular micrometer
at a magnification of 10 X- Larger measurements were taken to the nearest
0.5 mm. with an E. G. Bogush measuring scale. All camera lucida renderings
by one of us (Bender) and dental studies were done at 10 x under a dissect-
ing microscope.

Dental Variation

Variation in gross size, due to age, is especially great in those microtines
whose molars continue to grow throughout the life span of the animal. In
a sample of 30 modern skulls of M. xanthognathus from various localities (12
immature, but with the molars erupted and the occlusal pattern established,
and 18 adults), the alveolar length of the upper tooth-row varied from 6.0
mm. to 9.1 mm., an increase of 3.1 mm. or 34% from youngest to oldest. The
occlusal length of Ml varied from 2.0 mm. to 3.0 mm., a gain of 33%%.

Once a functional triturating surface has been established early in the life
of the animal, there is no change in the occlusal patterns of the individual
molars. Wear does take place, but since the molars are compensating for
this by elongation, and the occluding surface of each molar is a plane, tooth-
wear does not express itself in any way that can be followed. It is, therefore,
often impossible to tell, from isolated molars alone, whether a given speci-
men is from an adult of a smaller species such as M. pennsylvanicus or an
immature individual of a larger species such as M. xanthognathus. In many
cases this is true of isolated mandibles with complete dentitions (although
the occlusal patterns of ml and m3 are usually diagnostic). This is no great
problem when age changes in the skull can be examined or when dealing
with modern material; but as much fossil material is fragmentary and scanty,
gross measurements of individual molars may be of little significance when
considered alone.

Each upper incisor of M. xanthognathus bears a shallow, longitudinal groove
on its anterior face. This is very apparent on adult specimens but is not
noticeable on smaller, immature animals. Grooving was noted on the upper
incisors of the fossil New Paris material exactly as in modern specimens. In
profile, the incisors of the New Paris fossil population (Fig. 2) appear more
pro-odont than their modern equivalents. Their tips appear less recurved,
but the degree of curvature of the incisor appears to be the same in both
samples. This difference is mechanical in origin rather than genetic, due to
“settling” of the fossil incisors in their sockets. The maxilla of a fresh
Microtus pennsylvanicus (Ord) was dissected and the developing root of the
incisor examined. In a fresh specimen the developing bulb of the incisor
normally sits on the floor of the maxilla immediately anterior to the alveolus
of the first upper molar. But as the skull dries and tissues shrink, a hollow area,
about as deep as the width of the incisor, is left between the back of the
incisor and the wall of the maxilla. This same area was examined in a fossil
M. xanthognathus skull (C.M. 5571). In this case the delicate posterior
rim of the incisor was crushed down again the maxilla. The space that
originally housed the soft tissues of the developing tooth was eliminated and
it was quite obvious that the incisor had been pushed towards the rear. As

1960

Guilday and Bender: Yellow-cheeked Vole

319

Fig, 2. Pseudo-procumbent incisors of fossil Microtus xanthognathus (Leach)
due to post-mortem alveolar settling
Modern Fossil

M.C.Z. 5584 C.M. 5556

M.C.Z 5585 C.M. 5559

U.A. 1871 C.M. 5553

each incisor is a segment of a circle any displacement backwards would make
the tips appear to migrate forwards, and hence be more pro-odont in profile.
This same condition is often noted in modem skulls that have been macer-
ated and the incisors subjected to any pressures. This displacement was
also noticed in many of the lower jaws as well as in rodent skulls of other
species from Lloyd’s Rock Hole.

There is a considerable amount of individual variation in the molar teeth
of M. xanthognathus. The thickness of the enamel band, as seen from the
occlusal surface, increases, apparently with age. Salient angles and re-entrant
valleys may vary from broad and U-shaped (C.M. 5585, Fig. 4), to sharp and
acute (C.M. 5572, Fig. 4). Individual alternating triangles of the molars may
be broad or narrow and their exact configurations are highly variable. The
pattern of the occlusal surface of the molars remains fairly constant; that is,
it varies within definite limits. Lower molars are more variable than the
uppers.

Upper first molar. Normal pattern is an anterior crescent followed by two
alternating triangles and a posterior crescent. Of the 28 modem and 13
fossil skulls examined but one variant was noted. A modem specimen (U.A.
10, Fig. 3) had the anterior crescent of the left Ml nearly divided in half

Fig. 3. Occlusal patterns of upper

A. C.M. 5554 Right M1-M3 fossil

B. C.M. 5555 Right M1-M3 fossil

C. C.M. 5556 Right M1-M3 fossil

D. C.M. 5562 Right M1-M3 fossil

E. U.S.N.M. 7702 Right M1-M3 recent

F. U.A. 10 Left M1-M3 recent

G. C.M. 5578 Left M1-M3 fossil

H. C.M. 5557 Left M1-M3 fossil

I. C.M. 5558 Left M1-M3 fossil
f. C.M. 5577 Left M1-M3 fossil

K. A.M.N.H. 16006 Left M1-M3

recent

L. C.M. 5557 Left M1-M3 fossil

1960

Guilday and Bender: Yellow-cheeked Vole

321

by an anomalous medium anterior re-entrant valley. This valley was filled
with cement and lined with unbroken enamel. The right Ml was normal.

Upper second molar. Normal pattern is an anterior crescent followed by
two alternating triangles and posterior crescent. No variation was noted in

AB eDEF g H I

mm. 2 4 6 8

I -J -J 1 _l

Fig. 4. Occlusal patterns of lower molars of Microtus xanthognathus (Leach)

A. C.M. 5540 Left ml-m3 fossil

B. C.M. 5541 Left ml-m3 fossil

C. C.M. 5572 Left ml-m3 fossil

D. C.M. 5582 Left ml-m3 fossil

E. U.S.N.M. 128330 ml -m3 recent

F. U.S.N.M. 271710 ml-m3 recent

G. C.M. 5583 Right ml -m3 fossil
FI. C.M. 5589 Right m 1 -m3 fossil

I. A.M.N.H. 16006 Right ml-m3

recent

J. C.M. 5543 Left ml fossil

K. C.M. 5562 Left ml fossil

L. C.M. 5564 Left ml fossil

M. C.M. 5566 Left ml fossil

N. C.M. 5566a Left ml fossil

O. C.M. 5581 Left ml fossil

P. C.M. 5584 Left ml fossil

Q. C.M. 5585 Left ml fossil

R. C.M. 5586 Left ml fossil

S. C.M. 5590 Left ml fossil

322

Annals of Carnegie Museum

vol. 35

28 modern and 13 fossil dentitions.

Upper third molar. Highly variable. Normal pattern is an anterior crescent
followed by three alternating triangles and a posterior, lingually directed loop
composed of two lingually directed “arms” which are confluent buccally.
The third alternating triangle may or may not be confluent with the posterior
loop, but it usually is. An incipient fourth salient angle may be present
on the buccal side of the tooth (C.M. 5561), or it may be completely absent.
It was present in four out of 28 modern and 1 out of 12 fossil examples. An
incipient lobe, rarely well developed, may be present at the back of the
posterior lobe (C.M. 5554, Fig. 3). An incipient fourth lingual re-entrant is
rarely present, but in one modern (U.A. 10, Fig. 3) and one fossil dentition
(C.M. 5578, Fig. 3) the fourth lingual re-entrant had split the posterior lobe
into two sub-lobes. The incipient posterior lobe was noted in 8 out of 28
modem dentitions, and 4 out of 10 of the New Paris fossils. One fossil speci-
men (C.M. 5557, Fig. 3) was unique in that the third lingual re-entrant valley
extended clear across the tooth, isolating its posterior lobe.

First lower molar. Normal pattern is a posterior triangle, five isolated alter-
nating triangles and an anterior trefoil. This tooth, especially its anterior
end, is highly variable, and no two, modern or fossil, were quite alike. The
anterior lobe of the trefoil may be blunt and bulb-shaped (C.M. 5584, Fig.
4) or attenuated and directed antero-laterally (C.M. 5583, Fig. 4). The
posterior lobes of the trefoil (the 5th lingual and the 4th buccal salients)
may be blunt or sharp, aborted or robust. The anterior re-entrants of the
trefoil may be present (C.M. 5586, Fig. 4) or absent (C.M. 5582, Fig. 4), and
are usually quite shallow. Rarely the fifth alternating triangle is broadly con-
fluent with the anterior trefoil. One fossil specimen was unique (C.M. 5540,
Fig. 4); the first and second alternating triangles had united into a crescent,
and all triangles and crescents were narrowly confluent with one another.
Rarely (C.M. 5566a, Fig. 4) the fourth lingual re-entrant valley meets the
third buccal re-entrant valley, a condition more typical of M. pennsylvanicus.
Normally (C.M. 5572, Fig. 4) the fourth lingual re-entrant turns forward and
penetrates the base of the trefoil.

There were so many intermediate variations present in the anterior part
of ml that they could not be classified into particular varieties. No variations
were observed in the modern dentitions, however, that could not be matched,
or nearly so, with fossil specimens from New Paris or Mt. Solon.

Second lower molar. Normal pattern is a posterior crescent, two alternating
triangles and an anterior buccally-directed lobe. The basic pattern does not
vary, but the degree of isolation of the triangles does to a limited extent.
They may be completely isolated from one another and from the posterior
crescent and the anterior lobe (U.S.N.M. 271710, Fig. 4), or confluent (C.M.
5540, Fig. 4). The anterior lobe varies in shape from triangular (C.M. 5572,
Fig. 4) to vermiform (C.M. 5589, Fig. 4), this was noted in both modern and
fossil specimens.

Third lower molar. The third lower molar is a highly variable tooth and
can hardly be said to have a “normal” pattern. Three types of occlusal pat-
terns were noted. (Fig. 5). 1. A posterior crescent, two widely confluent
alternating triangles and an anterior cresent (C.M. 5583, Fig. 4). 2. A

posterior crescent, two isolated alternating triangles and an anterior crescent.

1960

Guilday and Bender: Yellow-cheeked Vole

323

3. The two alternating triangles are completely confluent resulting in an
occlusal pattern of three crescents (C.M. 5589, Fig. 4).

New Paris fossil. Sample size, 8 Modern. Sample size, 25
number percentage number percentage

Type 1 3 38% 10 40%

Type 2 2 25% 5 20%

Type 3 3 38% 10 40%

Fig. 5. Chart showing frequency of dental variations in the lower third molar
of Microtus xanthognathus (Leach)

Cranial Variation

The fossil skulls from New Paris average slightly smaller in most dimen-
sions, particularly in the rostral area, than the modern material we have ex-
amined. Fortunately the fossils are so well preserved that their individual
ages may be estimated. Age criteria used were: degree of development of the
interorbital crest, presence or absence of frontal pit, and general rugosity
of the skull. Evidence of Haversian canal activity, indicating active bone
deposition was noted on the dorsal surface of the zygomatic arms of the
maxilla.

The fossil series was judged to be almost entirely composed of young,
actively growing animals, probably sexually mature but not yet fully grown.
Only one specimen (C.M. 5576) is believed to be from a fully adult animal.
As a result it is hard to obtain modern comparative material of a comparable
age level (and growth changes in M. xanthognathus are so extreme that this
is a necessity). The apparent scarcity of the young adult age class in the
modern material, and its almost universal presence in the fossil collection
is probably because a sink-hole trap does not operate at random throughout
a population. There is a greater probability that young animals, rather than
older established ones, will be trapped. The amount of random wandering,
hence the chances of blundering into a sink-hole, would be greatest among
the younger, unestablished animals. Modern specimens are collected pri-
marily by trapping. A mousetrap is somewhat less selective than a sink-hole,
producing a greater agd^pectrum in any given collection. And since modern
collections of this animal 5re rather small and growth changes so pronounced,
any conclusions arrived at without making due allowance for age are apt
to be spurious. We have interpreted the slightly smaller average measurements
of the fossil crania as being primarily due to the biased age composition of
the sink-hole trapped collection, rather than to any genetic difference be-
tween the two populations^fossil and modern. The extent of these genetic
differences, if they existed, was apparently not expressed in the skeleton or
the dentition, as preserved in the cave material. We are unable to find any
consistent morphological differences between the fossil and the modern popu-
lations of Microtus xanthognathus (Leach) in the material examined by us.

This close similarity between the modern form and its late Pleistocene
Pennsylvania counterpart is, we believe, due to the lack of any significant
degree of geographical isolation or of ecological change that may have stim-
ulated evolutionary selection. Admittedly we have considered only the skull
and the dentition of the animals involved, but, based upon the evidence at
hand, plus the boreal forest affinity of the associated fauna at New Paris,
the Pennsylvania record, and by inference the Virginia one as well, is in-

324

Annals of Carnegie Museum

vol. 35

dicative of a former cooler time in the area. The present range of the yellow-
cheeked vole, therefore, represents but a remnant of an earlier maximum
glacial extension south to at least 37° N. latitude in the Appalachian region.
It has, apparently migrated to the north, during post-Wisconsin times, keep-
ing pace with climatic changes.

As both Microtus xanthognathus and Microtus chrotorrhinus occurred to-
gether in the Lloyd’s Rock Hole local fauna, with no trace of intergradation
in the dentition or the crania, this would seem to argue for the validity of
the two species (See Hall and Kelson, 1959, p. vi of v. 1 and p. 741 for discus-
sion of this point). The interesting question then presents itself of why M.
xanthognathus should be restricted today to the western half of the Hud-
sonian life zone, while M. chrotorrhinus appears to be restricted to the eastern
half, plus a relict distribution down the Appalachians. The two have been
distinct forms since at least the Late Pleistocene. They occurred together
in the Appalachians at one time. Yet they do not occupy common ground
today. More research must be done on the present habitat requirements of
these voles and on Late Pleistocene fauna from the area before this question
can be answered.

SUMMARY

Skeletal remains identified as Microtus xanthognathus (Leach) have been
recovered from Late Pleistocene cave deposits at Natural Chimneys, Mt.
Solon, Augusta County, Virginia and at Lloyd’s Rock Hole, New Paris, Bed-
ford County, Pennsylvania. This constitutes a range extension of about 1400
miles in aerial distance and 14 degrees of latitude. The present range of
this vole is from Alaska to Manitoba, south to about latitude 52 degrees north.

Dental variation in Microtus xanthognathus (Leach) is discussed and il-
lustrated.

Microtus xanthognathus (Leach), and Microtus chrotorrhinus (Miller) oc-
curred in apparent temporal association in the Lloyd’s Rock Hole local fauna,
Bedford County, Pennsylvania, during Late Pleistocene times.

REFERENCE

Hall, E. Raymond, and Keith R. Kelson

1959. Mammals of North America. Ronald Press Company, New
York. v. 1, p. 1-546; v. 2, p. 547-1083.

LIST OF SPECIMENS EXAMINED

Skulls and skeletal material as follows

Alaska

Yukon River: A.M.N.H. 16004.

Yukon River, mouth of Porcupine River: A.M.N.H. 16006, A.M.N.H.
16010, M.C.Z. 1593, M.C.Z. 5403, M.C.Z. 5307, M.C.Z. 5308.

Yukon River, 200 miles southwest of Porcupine River: M.C.Z. 1601,
M.C.Z. 5302, A.M.N.H. 16009, M.C.Z. 5584, M.C.Z. 5585.

Yukon River, Charlie Creek: U.S.N.M. 128330.

Takotna: U.S.N.M. 271710.

500 miles up the Kuskokwin [szc] River: M.C.Z. 17001.

Castle Rocks, 63° 40' N., 152° 07' W.: U.A. 9, U.A. 10, U.A. 11, U.A. 12,
Headwaters of Indian Creek, 64° 35' N., 144° 55' W.: U.A. 1871, U.A.
1873, U.A. 1875, U.A. 2820, U.A. 2779, U.A. 1872, U.A. 1867.

1960

Guilday and Bender: Yellow-cheeked Vole

325

Northwest Territories, Canada

Fort Anderson, Mackenzie River: M.C.Z. 5590.

Fort Resolution, Great Slave Lake: U.S.N.M. 7702.

Peel’s River, Mackenzie River: M.C.Z. 5589.

Mackenzie River, 30 miles up Willow River: N.M.C. 4715.

Alberta, Canada

Peace Point, Peace River: M.C.Z. 18894, M.C.Z. 18896, M.C.Z. 18897,
M.C.Z. 18939.

Natural Chimneys, Mt. Solon, Augusta County, Virginia
Late Pleistocene: C.M. 5861.

New Paris, Sink-Hole #4, New Paris, Bedford County, Pennsylvania
Late Pleistocene: C.M. 5534-5544, 5546, 5547, 5550, 5553-5562, 5566, 5570,
5571, 5573, 5576, 5578, 5584, 5589.

Definition of Measurements

Cranial:

1. Length of nasal bones.

2. Greatest width of nasal bones.

3. Least interorbital breadth.

4. Least rostral breadth, taken immediately anterior to the zygomatic

plate of the maxillas, and above the incisor swellings.

5. Length of incisive foramina.

6. Transverse width of single upper incisor, taken just back of

occlusal surface.

7. Alveolar length of upper tooth row, M1-M3.

8. Occlusal length. Ml

9. Occlusal width, Ml.

10. Occlusal length, M2.

11. Occlusal width, M2.

12. Occlusal length, M3.

13. Occlusal width, M3.

14. Greatest width across skull, taken at the maxilla-jugal sutures.
Mandibular:

1. Total length of mandible, from most anterior point of symphysis

to most posterior point on the angular process.

2. Depth of mandible, from the most anterior point of the alveolus

of ml, project the angle of that tooth through the bottom of
mandible. (Hard to take with any degree of consistency.)

3. Diastema, from margin of alveolus of incisor to anterior margin of

of the alveolus of ml. (Also hard to take consistently.)

4. Alveolar length of lower tooth row, ml -m3.

5. Occlusal length, ml.

6. Occlusal width, ml.

7. Occlusal length, m2.

8. Occlusal width, m2.

9. Occlusal length, m3.

10. Occlusal width, m3.

11. Transverse width of lower incisor.

All measurements are given in millimeters.

Cranial Measurements— Microtus xanthognathus (Leach)
New Paris fossil— unaged

326

Annals of Carnegie Museum

vol. 35

^ lilil I I I

CO CM | | | | | CM

CM | | | | | lO

■— 1 CM 1 1 1 1 1 CM

= 2 i 2 2 2 i

O © | od © © |

^ CM | CM CM | |

00
CM CM

I I

©

CM

I I I I

« I

I I I

*>11111

^ I I I I I I

**11111-1

<*111111

H I I I I I I

CO

I I

CO CM |

CM |
r~l CM 1

© CM <
1—1 CM CM

© 2 i

m i

lj 00

bo

S3

3

^ ©
'2
'3d

I > * I

c

c>

o

I I

co I |

<* I I

- I I

3d

o

o in in q .cm

© no iri r-1 r~l 1 oo

IffJ © HO ifS lO
HO CO CO GO HO

GO 0"5 © CM CM © CO CM © 00 00 © 00

cm’_h. o<Mi>5Ccot>oqwi>i>i>

1—1 CM C'icMCMCMCMCM^'cMr— I.— If— h

CM © CM i— i >— i © CM © © 00 00 ©

© C" C"»

oqa)05oqqc5i>oot^in

i *> I | g % *>

(MOhphhWWCMO©©©

iq © iq GO l>» CM CO CM © CM
CM CM CM CM CM CM CM CM CM CM

—h©t^CM©©©h0^h©

C^t^t^C^C^©©©©©

s ^ ^ ^

CM CO iq CM | OO q j © |

< QJ . 00 >

J. ■§ ^

e S

<U ^

CO

CM

oqqqoqcMqqqq

iri©irj©©^^d'^'^'^

ncOiO^©.— icqCM©©

CO CO GO CO CM CO CO CO CO CM

©t^uO©'^-|©OOCOCM©

co go co ^ co co oo co co c-i

-q © © © CO I© | CM 1— < 00 00

CO CO CM CM CO 'co ' GO GO CM CM

ho no

cq © CM | . © © © ©

i-^odoo 'oo ' ©” © ho ho

Je? ^ S ^ S

U U U U U U

00 © © © TO HO

if5 © 1> 00 OO

iO if"5 iT5 iO iO iO

if} ifi iO i£S iO iO

U U

a rO

© ©
oo oo
no 110

N N N N N g

U U d d U Z < < < < < <

<u

©r-^©©©HO.—iI>-CMCO©© >

© © CO 00 © 1 i— i © [''* C*» 10* CM C3

00 00©iOif)t^00 X 00 00t^00
OOOOOOinm^^^-^CMlM

Cranial Measurements. Continued

1960

Guilday and Bender: Yellow-cheeked Vole

327

I 5 i I

w OfO | CM

CM CO CO | OO

^ CM CM 1 CM

s 1 1 t>do6cr-aiacr!^

OffiOOClCDOO
CM r— i CM CM <— 1 — h CM — i CM

o w ci oo ci

CM i— i CM h

iq 00

CM CM CM CM

. I NJ.
I>* £""» t"-*

CO ^ | ^ 2

O r-H 00 O
CO M3 in to

^ «? I «

CO CO 1 CO

co °1 °1 I ^

CO CO 1 CO

CM ~ I I I

CO 1 1 1

U

GO CM CM CM CO CM CM , CO , CO CM CM j CM CM CM CM CO

GO CM CM CM CO CM CM | CO |

fciO

©

-s;

x 3

» T3
5 a

l c

Cj J-l

1 I

C

<u

wt>ioc^ooooiqc)ocit^'^to , iq ^ in o

1—1 CM CM CM CM CM CM CM CM CM CM CM CM ' CM CM CM CM CM CO

CM —< CM CO CM CO |CO | CO CM CM CM CM CM — CM ^

O O — i — — i r-H CM CO CO CO CM —i O GO i-h O — 1 — 1 O cO

1—1 CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM

CM CO CO CM CO CO | ^ | CM CM CM —< CM CM CM CM CM CM

csoo©®soo^©r'~CMOoor^©r^c''*oi>-cr>o

CM CO CO Tj-1 CM CO 00 SO CM — i CO

OS O CM — < no -h

oooooocooooooot-'coooooi>- 1 N 00 00 a X Ol

l>> | | ^ | 1> O O) j | | j j | | ^. | t-

ifiino«o^»nco>i>int'CMCMCMoopHinci^
co G t>> co co i> co co co co co G co co t>- co co G

oooocococo»ooococoiO i ^co^oocOincMin

Gooocoooasiooocoooas .cm , oo oo co ^ <o m

i no l

CMcooooco^ocqcoco^oqcqoqiocqcqcqcq

^coco^cocoooco^^cocococococococo

oocoooocm— HCOHTtjocMOcqoqioos'^H in

dcicicioicidcidcidooooooocicici Mai

&

u u u u

co — i co oo
m co i>-
ns m in m

if) if) if)

s. d s S S ;

<<<<^dddqqqqqqqqqu

oo-HWTfMCwoow^MJO^mcooHT}!

^Hrt^HOOOco — ooooooooosooco
ooi>cot^cocococoinmincoooci

cooi>oooinif)if)if)if)if)fHpHi>a)

—H —< CM t"* — 1 — <

— CM

Cranial Measurements. Continued

328

Annals of Carnegie Museum

vol. 35

Tf ^ O I tO © LO I | <

06 06 1 cn cri oo 1 1 ci oo

GO CM CM G1 CM | CM | | |

CM ^ | ^ | | |

CM CM CM CM 1 CM 1 !

r— i CM CM © OO CM CM | 50 |

O © © CTO CM CO i— h
— ' cvj ^ CM — < CM

I I

^ oo CM W !M so W | CM |

qq oo | xo oo oo oo | co |

CM 1 CM CM CM CM 1 CM 1

3 t-

co to cm —(

oo oo oo

I I

| q iq so q ^ | CO

ft,

O q so l>» CM CM o
iri 1 co cd to o cd d

I I I

to o o

w oo q I q oo cm q , ©

OO GO OO ^ OO rJH

I I I I I I I

I I I I I I I

s

-c

«

S

ho

o

-sc

_ tO CM OO CM CM CM OO

o o oo co o o

Cl oo iO o i>

oo oo o o

© SO t" CO tr~ C"»

CM — O CM CM CM

tO

X u

w, a

S cS

O S
ts^
5 T

T c

to i — i <— i <— h ;

SO Of) OO OO OO SO

i> o cm so 05 oo
i> to to co to

OO I I I I I

CM I I I I 1

to to © cq ©
oo oo oo oo

N

N

N

N

N

s

s

s

s

s

s

S

s

s

<u

ho

U

U

U

u

U

z

q

he

U

U

U

U

U

U

U

U

U

r3

in

s'

§

s'

s'

U

ct

QJ

qj

uo

oo

o

CM

CD

oo

>

oo

o

CD

I>»

CO

to

>

to

to

to

to

MO

oo

!>■

03

00

CO

CO

CO

oo

i— <

C3

to

to

to

to

no

lO

iro

to

to

00

00

CO

to

to

to

to

to

iO

to

»ro

1C

to

00

00

oo

Mandibular Measurements. Continued

1960

Guilday and Bender: Yellow-cheeked Vole

329

I | | ■— h _ |00

©©—h— j^H^HOOCTii-jO-HOCM | j

CM i—3 CM i— h >— 3 CM CM CM CM CM > — 3 CM

^ © CM CM O CM © © CM CM CM © CM |

p P ^ °) P o © © p p

1—3 CM CM >— 3 C^i i— 3 — ’ CM CM CM CM c^i i— 3

^ CM CM OO CM CM i-H i— h <50 CM CM CM p | p

cooowoowMMOTwwwwmm

i * r-

P r> p p oo or) p cs cm p ip p tp oo

ooVi>i>o6i>t^^06t>i>06i>t>3

^ 0C | , | CO . | , | OT Tf a w M

00 id 1 I lid1 1 1 I id id id id id

SO. , , p GO i I . 00 t>- © CM

^ d 1 1 1 to d 1 1 ' id d id d ©

if) | | | o o o o m m

© q ©

cm d ^ —3 cm —3 1 od cm — 3
CMCMCMCMCMCM cm cm cm

p o

2 111

I I I

00 d p I

^5 I I

CO 2 ! 2

I

2

*1

I i I

^->3 I I

^5 I I

- I I

;;;;;;;;; ^ ^ S S

NNNNNN1^NN^^7'7'7

udqduuuuqss^^^ g,

<V

m^cMi>oOTti^'~iTi^(£i(MOO >
OiOOOOooOOOCTiOOOoO^ 03
iocooo«,icoir5lo©oo©ot^'0or^

rHpHlfllOlfJlDlOtNOOOCOt^OO-H

— • — ^ — 1 — I CM O

— i CM

§ S S
u u u

u T5

© 3

m m
m m

5561e

vol. 35

330

Annals of Carnegie Museum

to in c io j in | oq i w if) m to to a

i i i 2 i i i i i i i 2 i i i

i i i 2 i 2 2 I i i i 2 i i

°o s i i i 2 i i i 2 2 2 2 i i 2

3

-3

3

3

©

3

3

* -H

g I

© .ss

' ~ i-l

S a

ifc
2 S
E H

ez

<jj

^ 05 C5 | | O 05 05 05 | © 05 ©

^2 I

05 ^-H

©i

^ 2 i 2 2 ~ i i i i :• 2 2 i i 1

to - oo to to | in . to co . . to in

GO GO CO CO ' GO GO GO CO GO * GO GO

in i« .woo ,^Hts
1 i>- 1 'co't^ooc^'t^t^r^1 1

GOC5ootqo | go <o tq go oo © . . oo

in in in in to * in in in in in in ' ' in in

o i-h oo go oo | i— h i— j tjh in oo >- j | | oo to

co to in in to 1 in to in in in in ' ^ in in

o , o o m | o oo m , . .

©j 1 O) — I -h 1 — I oo oj 1 1 1

©[ . — i ©x ©j ©r -h — <

uuuuduuuuuuuuuuu 2

^lCO^OOOO'HiMOO^O^O©!^

0O0OG(50<5M0O'^l'^l'^TtlTtlTtlTfHin(OH

inminininininininininininminto
minininininininininininininmin

<u

>

20.5

So 7 7 S
P4- PH<~f'oz.

ART. 15. NOTES ON SOME NON-PASSERINE BIRDS
FROM THE PHILIPPINES

By Kenneth C. Parkes
Associate Curator of Birds, Carnegie Museum

In August and September 1956 I collected birds for a virus research
project in central Luzon, Philippine Islands. A general report on this col-
lection and a more recent Luzon collection acquired jointly by Carnegie
Museum and the American Museum of Natural History will eventually
appear. In order to keep the final report manageable in size, I am publish-
ing separately certain of the more extensive taxonomic discussions arising
from these Philippine studies. Several such papers have already appeared;
the present paper treats five non-passerine species at greater length than
would be practical in the final report. These findings are based primarily on
the combined collections of Carnegie Museum and the American Museum
of Natural History; I am grateful to Dr. Dean Amadon of the latter institu-
tion for many courtesies. Certain additional specimens were examined at
the United States National Museum through the kindness of Mr. H. G.
Deignan, and others were borrowed from the Museum of Comparative
Zoology through Dr. Raymond A. Paynter, Jr.

The field work in 1956 was carried out as part of a project of the Graduate
School of Public Health, University of Pittsburgh, under the sponsorship
of the Commission on Viral Infections, Armed Forces Epidemiological
Board, and supported in part by the Office of the Surgeon General, U. S.
Department of the Army. Specimens were collected and prepared with the
invaluable assistance of Mr. Telesforo Oane and the generous counsel of Dr.
Canuto G. Manuel, both of the National Museum of the Philippines.

Ptilinopus occipitalis. Yellow-breasted Fruit Dove

Several authors have discussed this species since Manuel (1936, p. 331) wrote
that he was unable to find any valid subspecific characters in the series of 23
specimens then in the National Museum of the Philippines. Some of these
authors, including Manuel, have used the generic name Leucotreron for this
species, but I follow the most recent monographer of this group (Cain, 1954)
in adopting the broader genus, Ptilinopus.

Much of the discussion has centered on two names given to specimens
from Mindanao: incognitus Tweeddale, 1877, and brevipes Hachisuka, 1930.
Hachisuka himself (1932, p. 185) synonymized his brevipes with incognitus,
and was followed by Ripley and Rabor (1956, p. 289) and de Schauensee
(1937, p. 4). Only one author, Peters (1939, p. 84) seems to have appreciated
the significant fact that Tweeddale’s type came from Butuan, Agusan, a low-
land region, and was thus not necessarily the same thing as the Mt. Apo
bird described as brevipes by Hachisuka. Manuel, who was unable to recog-
nize any races, had only lowland specimens from Mindanao. De Schauensee's
comparisons were made between two lowland Mindanao specimens and one
specimen from Luzon ( occipitalis ).

Mt. Apo birds are definitely smaller than true occipitalis, as stated by
Ripley and Rabor, but the color differences are even more striking than
would be gathered from their comments. They compared their Mt. Apo birds
with specimens from Negros rather than with topotypical occipitalis from

331

Submitted for publication, December 22, 1959
Issued November 7, 1960

332

Annals of Carnegie Museum

vol. 35

Luzon. From the latter, Mt. Apo birds differ even more than from Negros
birds. Two specimens from Ayala, Mindanao, a lowland locality, are inter-
mediate both in size and color between Luzon occipitalis and the Mt. Apo
birds. De Schauensee's two Lake Lanao specimens are also somewhat inter-
mediate in size (The “striking” color differences listed by de Schauensee
are individual variations in this species, possibly correlated with age but not
with geography).

In view of the differences between highland and lowland Mindanao speci-
mens, it would seem preferable to follow the treatment of Peters (1939, p.
84), recognizing the highland race under the name brevipes Hachisuka; the
name incognitus, used by recent authors, is based on an intermediate popu-
lation which is perhaps nearer occipitalis than brevipes .

In true brevipes the throat is heavily washed with yellow, this color invad-
ing the whole frontal area to some extent. The yellow breast patch is deeper
in color than that of occipitalis, and is washed with red in extreme speci-
mens. The red abdominal patch averages more deeply colored and also
larger than that of occipitalis. This is contrary to the statement of Peters
(loc. cit.), but Peters had, for comparative material of occipitalis , only two
specimens from Mindoro. I have examined these two skins, and find that
their red abdominal patches equal or exceed those of any of the Luzon speci-
mens seen. They are, however, nearest occipitalis.

Specimens from Negros and Samar (wrongly taken by Ripley and Rabor
[1956, p. 289, and 1958, p. 30] as typical of occipitalis ) are intermediate be-
tween occipitalis and brevipes. They are so distinctive in appearance as
possibly to be considered worthy of subspecific recognition. I do not propose
to take this step, but will point out that if recognition of a third subspecies
were to prove desirable in view of the large range occupied by these inter-
mediate birds, the race should bear the name incognitus Tweeddale, as low-
land Mindanao birds resemble those of Negros and Samar quite closely.
A single Cebu specimen in Carnegie Museum also belongs here; the species
may be extinct on that island (Rabor, 1959, p. 40). These birds differ from
topotypical occipitalis as follows: throat and forehead with deeper yellowish
wash; yellow patch on breast brighter and richer in color, and extending
to the whitish throat instead of being separated from it by a gray band;
raspberry-colored abdominal patch averaging more extensive. These charac-
ters are all obvious trends toward the conditions found in brevipes. In size
these specimens average slightly smaller than occipitalis, and are thus larger
than brevipes (See measurements given by Ripley and Rabor, 1956, p. 289).

In summary, then, the name brevipes Hachisuka should be used for the
Yellow-breasted Fruit Dove of the Mindanao highlands. Islands from which
I have examined specimens which may be called occipitalis include Luzon,
Mindoro, and Sibuyan. Specimens from Negros, Samar, Cebu, and the
Mindanao lowlands represent an intermediate population, occipitalis X bre-
vipes, for which the name incognitus Tweeddale is available.

Cacomantis variolosus. Rufous-breasted Brush Cuckoo

The extreme taxonomic positions that may be taken by contemporary
authors are well illustrated by opinions on the status of the genus Cacomantis.
Delacour (1951, p. 19) synonymizes Cacomantis with Cuculus on the grounds

1960

Parkes: Some Birds from the Philippines

333

that size is the only difference, with the species pyrrhophanus (usually as-
signed to Cacomantis) bridging the gap. On the other hand, Verheyen (1956,
p. 25) places Cuculus alone in a subfamily Cuculinae, placing Cacomantis in
a subfamily “Surniculinae.” No species of Cacomantis , however, is listed
among Verheyen's examined material. After having examined skins of all
species of Cuculus and Cacomantis , I feel that Delacour is almost certainly
correct, but I use Cacomantis for the time being, pending a thorough study
of this group of cuckoos.

Cacomantis variolosus is one of the many East Indian species that badly
need revision, but for which carefully labeled material is still grossly inade-
quate. The series I have examined strongly suggests that some authors who
have discussed this species have not taken certain factors into sufficient
consideration:

1. Post-mortem color change. My two specimens, taken in 1956, were
compared within the following year to older Luzon specimens. The shade
of brown on the underparts of the freshly collected birds was decidedly dif-
ferent from that of the older skins.

2. Individual variation, particularly in the color of the underparts (in-
cluding relative amounts of brown and gray). Such variation is apparently
extensive.

3. Migration. Some forms of Cacomantis are known to be migratory. A
specimen collected by Zimmer at Antipolo, Rizal Province, Luzon, on Decem-
ber 14, 1913, differs from any other specimen of C. v. sepulcralis examined
in the richness of the rufescent color of the underparts, approaching the more
southern C. v. everetti in this respect. That this color has nothing to do with
the age of the specimen is indicated by the strikingly paler underparts of a
specimen taken by Zimmer at nearby Manila only a month later. Both speci-
mens were marked adult male by the collector.

Some combination of the factors listed above may account for certain dis-
crepancies in the literature. For instance, Ripley and Rabor (1958, p. 38)
listed an unspecified number of Negros and Mindanao specimens as having
the palest and dullest underparts in their series, while Rand (1951, p. 581)
found that one Mindanao and one Negros specimen were the darkest and
richest chestnut below, while another Negros specimen was as pale as Luzon
birds.

Hartert (1925, p. 166) separated everetti on the basis of a single specimen
from Tawitawi with no gray on the throat; however, as Rand (1951, p. 581)
has indicated, this is a highly variable character. Peters (1939, p. 90) ex-
tended the range of everetti to Basilan on the basis of another single speci-
men which agreed with Hartert’s description in having “the entire under
surface chestnut to the base of the lower mandible.” However, A.M.N.H.
625758, also from Isabela, Basilan, has just as much gray on the throat as
any Luzon specimen. Even discounting this character, the only one used by
Hartert, everetti appears to be separable from sepulcralis on the basis of
deeper color of the underparts, and perhaps somewhat darker, less clear
gray head. Ripley and Rabor (1958, p. 38) listed Basilan among their
localities for sepulcralis without commenting on the fact that Peters assigns
Basilan specimens to everetti. A single immature specimen of everetti from

334

Annals of Carnegie Museum

vol. 35

Tawitawi in the American Museum has the dark transverse bars of the under-
parts much broader than any of a series of five of similar age from other
Philippine localities.

Whether everetti is separable from all of the many non-Philippine races
of Cacomantis variolosus, and whether the Philippine population here called
sepulcralis is actually inseparable from that race (described from Java and
Sumatra) can not be determined until a much more thorough study, utilizing
many more specimens than hitherto assembled by any author, has been
accomplished.

Rand (1951) has summarized the tangled history of the confusion of the
two species Cacomantis variolosus and C. merulinus in the literature of Philip-
pine birds. He established that variolosus occurs on Luzon, Negros, and Min-
danao, and Ripley and Rabor (1958, p. 37) added Mindoro and Cebu. To
this list may be added Masbate, on the basis of the Bourns and Worcester
specimen from that island listed by McGregor (1909, p. 374) under “Caco-
mantis merulinus,> , and now in Carnegie Museum.

Examination of a composite series of all available races of Cacomantis
variolosus in the American Museum of Natural History has permitted a few
deductions about the plumage sequence in that species. Even in the well-
known European Cuckoo, Cuculus canorus, the plumage sequence is not fully
understood (Witherby et al. , 1938, p. 300), but Cuculus appears to have a
cycle including both a prebasic and a prealternate molt (for terminology see
Humphrey and Parkes, 1959). I have not attempted to work out the sequence
in Cacomantis variolosus in full, but the following observations may be of
interest.

Two specimens were taken during the 1956 expedition; C.M. 137164 on
September 6, and C.M. 137287 on September 29. The latter is marked “skull
fully ossified”, the former “windows in skull.” The ovaries of both specimens
were about 5 mm. in diameter. The specimen with the ossified skull has a
more “immature” look about it than does the other, having faint brownish
edges to the wing coverts and secondaries, and indistinct barring on the
posterior underparts, particularly the flanks. C.M. 137164, with “windows”
in the skull, has no discernible edgings on the wing feathers, and merely the
faintest trace of ventral barring. The base of the mandible of the latter
specimen was “dull orange”, that of C.M. 137287 (skull “ossified”) “yellow-
ish horn.” Gilliard (1950, p. 483) gave the color of the base of the mandible
of the adult as “lemon yellow”, of the immature as “pale orange.” I have
examined four of the five specimens collected by Gilliard at Lamao, Bataan
Province, Luzon; the fifth was apparently deposited in the collection of the
National Museum of the Philippines. Gilliard lists these specimens as “two
male immatures, two females, one (?).” Collating this list with the labels, it
appears that the specimen left in Manila was one of the two “male im-
matures.” The one bird so labeled in the American Museum is molting out
of a barred into an adult-like plumage. The other three specimens are all
barred, and it thus appears that Gilliard was under the impression that the
adult female of variolosus, like that of certain other Cuculinae, is barred and
unlike the male. This, of course, is incorrect, and the distinction between
“adults” and “immatures” in Gilliard’s soft-part color notes is meaningless.
He had no adults.

1960

Parkes: Some Birds from the Philippines

335

Examination of certain non-Philippine specimens of this species indicated
that the barred juvenal plumage is succeeded by a second barred plumage,
which is followed by the first “adult” type of plumage (I avoid using definite
plumage names for these until the cycle has been worked out). The latter
is not the definitive plumage, and may, especially in females, be more or
less barred on the underparts. Both of the 1956 specimens are apparently
at this stage. The discrepancy in skull ossification may be explained if
Cacomantis follows the schedule of Cuculus; Verheyen (1953, p. 16) found
that in Cuculus canorus at least 10-11 months may elapse before the cranium
has become completely ossified. The two 1956 birds would then represent
specimens taken near the end of this period; it will be noted that the speci-
men with “fully ossified” cranium was taken over three weeks later than the
one with “windows”.

C olio c alia esculent a. Glossy Swiftlet

With several important series of specimens before me that were unavail-
able to Mayr at the time he discussed the Philippine races of this species
(in Delacour and Mayr, 1945, p. 109-110). I find that my observations differ
from his in several respects.

I am unable to see any consistent difference in either the color or relative
glossiness of the back, or in the relationship between crown and back color,
in isonota, marginata, or bagobo, although Mayr mentions these characters
several times. There is a certain amount of individual variation in blueness
or greenness of iridescence within a series from a single locality, and it is
also possible that, as in other iridescent birds, the apparent color may change
with wear.

As for size, there is more variation than indicated by Mayr’s figures, which
were based on an admittedly small series. There is also more overlap between
isonota and marginata than Mayr believed:

Mayr isonota wing 102-108 mm.; marginata wing 95-103 mm.

Parkes isonota wing 98-107 mm.; marginata wing 93.5-106 mm.

Mayr named the race septentrionalis , which I have not seen, from the
small islands north of Luzon. He characterized it as “larger [than marginata ],
paler and more greenish on the back”, and gave the wing lengths of two of
his five specimens as 105 and 106 mm. Comparison with the figures above
will show that if septentrionalis is a valid race, it must be based on color
characters alone. I do not think size is of much significance among these
populations, and certainly it is misleading to describe, as Mayr does, a single
Mindoro specimen with a wing of 101 mm. as “smaller” than isonota , and a
single Mindanao specimen with a wing of 100.5 mm. as “rather small.” The
six Mindoro specimens measured by Ripley and Rabor (1958, p. 43) had
wings of 101-104 mm.

I have seen two Palawan specimens, probably two of the three examined
by Mayr. Although their wing measurements (95, 98 mm.) are equivocal, I
agree with Mayr that on the basis of color these probably represent an un-
described population. Although Mayr does not mention the underparts, these
two Palawan specimens are darker ventrally than any other Philippine speci-
mens of this species seen.

Mayr states that the white area on the abdomen of marginata is less exten-
sive than in most specimens of isonota. With an unprecedented series of 42

336

Annals of Carnegie Museum

vol. 35

specimens of marginata before me, I can not see anything in this supposed
character.

Cebu specimens (topotypes of marginata) are rather consistent in having
broad white margins to the rump feathers. Good series from three central
Luzon localities (Lamao, Bataan Province, 11; Sacobia River valley, Pam-
panga Province, 5; Mt. Donald McDonald, Pampanga Province, 15) illustrate
Mayr’s statement that “the expression of these margins is quite variable, as
illustrated by every series of this race." Actually these highly variable central
Luzon birds probably represent intermediates between isonota of northern
Luzon and typical marginata of Cebu; as indicated above, I do not believe
the other characters listed by Mayr as separating these two subspecies are of
any real value. Cebu specimens do have tails which average slightly less
forked than isonota , but some of the Bataan birds show no fork at all (tail
square), although geographically intermediate.

Megalaima haemacephala. Crimson-breasted Barbet

I follow Peters (1948, p. 31) in the spelling of this generic name; many
recent authors have used Megalaema.

When Gilliard (1949, p. 276-277) discussed this species, he had available
from northern Luzon only two specimens from “Locos Norte" [=Ilocos
Norte Province]. He noted that “generally speaking, northern birds have
the shortest bills . . .” Examination of additional Luzon material not seen
by Gilliard not only supports this point, but indicates that subspecific sep-
aration of the northern and southern Luzon populations might be possible.
However, Gilliard has restricted the type locality (“Philippine Islands") of
Bucco haemacephalus P. L. S. Muller to Lamao, Bataan Province, a locality
from which specimens are almost exactly intermediate in size between those
of northern and southern Luzon, as suggested by Gilliard’s own measurements.

Gilliard based his new race celestinoi on one specimen each from Samar
and Leyte, stating that celestinoi differed from haemacephala only in size,
not in color. The size differences shown in Gilliard’s table of measurements
are small, and might disappear if larger series were measured. I have ex-
amined Gilliard’s two specimens. He mentioned a general north-south cline
in increasing depth of color of the yellow throat. Although he did not so
state in his description, the throat of celestinoi is rather brighter than that
of haemacephala, as would be expected. In addition, the two specimens of
celestinoi are among the most heavily pigmented individuals of this species
examined, in that the ventral green streaks are deep, almost black, along the
flanks; far darker, in fact, than those of Mindanao birds, although Rand
(1948) used this as a character of his mindanensis. It is hard to evaluate the
dorsal color of celestinoi, as both specimens are rather worn, but they appear
darker than comparably worn specimens from Luzon. These color characters
indicate that celestinoi is probably a more strongly differentiated race than
would be apparent merely by reading the original description.

I have examined the same five adults from Mindoro that were listed by
Gilliard. As indicated by his table of measurements, these average slightly
smaller in wing and tail than Luzon birds (although not in bill), and thus
indicate an approach to celestinoi. This is confirmed by color, the Mindoro
birds being deeper in color, both dorsally and ventrally, than the average of
Luzon birds. As a series, however, the Mindoro birds are best assigned to

1960

Parkes: Some Birds from the Philippines

337

haemacephala . Ripley and Rabor do not discuss this species in their Min-
doro paper (1958).

I agree with de Schauensee (1957, p. 7) that mindanensis Rand is a very
weak race. I have examined 8 Mindanao and 33 Luzon specimens, and find
that the only characters that hold up at all are a slightly deeper yellow throat
and slightly deeper green back in Mindanao birds, the latter difference much
obscured by wear. The other characters ascribed by Rand to the underparts
of mindanensis can easily be matched in the Luzon series. Although, as noted
above, Gilliard did not mention color characters in his description of celes-
tinoi, the latter race differs (as far as can be judged from the two specimens
seen) more from haemacephala than does mindanensis , which was based en-
tirely on color. In view of the extent of variation in color and size admitted
in haemacephala in Luzon and Mindoro, it would appear that the slight dif-
ferentiation of Mindanao birds is not worth nomenclatorial recognition.

Chrysocolaptes lucidus. Crimson-backed Woodpecker

As is well known to the committee that prepared the most recent edition
of the American Ornithologists’ Union “Check-list of North American Birds”,
it is difficult to select a single appropriate English name for a wide-ranging,
highly polytypic species. The name used above for Chrysocolaptes lucidus
is that adopted by Delacour and Mayr (1946, p. 149), although, as these
authors realize, this species is a mosaic of crimson-backed and golden-backed
subspecies. Most of the Philippine races belong to the former group.

As Gilliard (1950, p. 472) has pointed out, several species of birds exhibit
marked geographic variation on the large island of Luzon; his list of 14
species is somewhat incomplete (Parkes, 1958, p. 1-2). After having studied
the birds of Luzon for over three years, I find that the most frequent pattern
of geographic variation within the island does not agree precisely with Gil-
liard’s division of Luzon into northern, western, and southern “subregions”.
Gilliard gave the boundary of his northern and southern subregions as “the
isthmus between Tayabas and Lopez bays or . . . the region of Laguna de
Bay.” This does not accord with the pattern of geographic variation I have
found most frequently. Leaving out of consideration the birds of the Sierra
Madre of northeastern Luzon, which are all but unknown, most northern
subspecies or differentiates of less than subspecific degree range south to about
16° N. Lat., or approximately to Pangasinan and northern Nueva Ecija
provinces. Between this line and, approximately, Manila, is an area in which
either of two things may happen. There may be a rather broad zone of

intergradation between a northern and a southern form (as in Collocalia

esculenta and Rhipidura cyaniceps); or the birds from this area may be

referable to the southern form, so that there is a rather abrupt transition

between the northern and southern races ( Parus elegans, Dicrurus balicassius).

For many species of birds Manila is either the original or the subsequently
restricted type locality. Manila is far enough south so that the birds from
the general vicinity of the city are almost always identical with or close to
the populations from southern Luzon. In such cases, a name has often been
given to a north Luzon population. I have pointed out above that Gilliard’s
restricted type locality (Lamao, Bataan Province) for the nominate race
of Megalaima haemacephala lies near the midpoint of a north-south cline
on Luzon. Salomonsen (1953, p. 240) has shown that Gilliard (1949, p. 278)

338

Annals of Carnegie Museum

vol. 35

was equally unfortunate in his choice of Lamao as the restricted type locality
of Picus haematribon Wagler. There are north-south dines in both color
and size in Chrysocolaptes lucidus on Luzon, and Bataan lies squarely in the
middle of these dines. Gilliard might better have chosen a type locality
south of Manila for haematribon (Salomonsen suggests Cavite Province;
equally good would be Mt. Makiling, which Gilliard did select for the
nominate race of Mulleripicus funebris). He could then, as Salomonsen
points out, have described the northern mountain race as new. Salomonsen
solved the dilemma by recognizing, albeit reluctantly, three subspecies on
Luzon: his own montium in the north, ramosi Gilliard in the south, and
haematribon Wagler in between.

While it is true that northern Luzon specimens tend to be grayer and
larger than those of southernmost Luzon, there is extensive overlap. Salomon-
sen suggested “mixing of labels” as an explanation for what he considered
to be incongruities in size, but examination of his table of measurements,
as well as Gilliard’s, suggests that the matter is not one of discrepancy but of
variability. As for color, overlap occurs here as well. As implied by Gilliard
in his description of ramosi, the color characters are scarcely visible in
females. In males, only two of four northern Luzon specimens in the Ameri-
can Museum conform to the description of montium, the other two being
indistinguishable from specimens from southern Luzon (=“ ramosi”). The
large series available from central Luzon is, as would be expected, extremely
variable. It seems best to reunite all Luzon populations under the name
haematribon, acknowledging the variation mentioned above. De Schauensee
(1957, p. 7), although having examined only five specimens (all, contrary
to de Schauensee’s statement, from within the putative range of ramosi), also
synonymized ramosi with haematribon, but did not mention montium.

Delacour and Mayr (1946, p. 150) considered C. 1. grandis Hachisuka of
Polillo to be “of doubtful standing”. I agree with Salomonsen (1952, p. 350)
that grandis is a good race. In addition to being larger than haematribon,
it is more brightly colored (in adults). The brownness of the underparts
mentioned by the describer (Hachisuka, 1930, p. 179) and by Salomonsen
(1952, p. 350) is too variable in haematribon to be a certain character for
grandis. However, in adults of the latter race the whole chest area is
faintly washed with red, and the olivaceous areas of the wing (basal half
of outer web of primaries, primary coverts, metacarpal edge, alula) are
heavily washed with red. The small crown spots of Hachisuka’s female type
specimen represent a sign of immaturity rather than a subspecific character.

SUMMARY

Geographic variation in five Philippine non-passerine birds is reviewed.
Ptilinopus occipitalis brevipes (Hachisuka) is revived for the population of
the Mindanao highlands, incognitus Tweeddale being applicable to an inter-
mediate population inhabiting lowland Mindanao and certain central islands.
The status of the Philippine races of Cacomantis variolosus is reviewed, em-
phasizing the difficulties of studying this species. Information is also pre-
sented on the relationship of plumage stages, skull ossification, and soft-part
colors of this species. In Collocalia esculenta, study of an exceptional series
indicates that some characters used by previous authors to define subspecies

1960

Parkes: Some Birds from the Philippines

339

within the Philippines are of no geographic significance. The subspecies
Megalaima haemacephala celestinoi Gilliard is tentatively upheld; M. In.
mindanensis Rand is not considered separable from the nominate race. Both
Chry so colap tes lucidus ramosi Gilliard and C. L montium Salomonsen are
considered synonyms of C. L haematribon (Wagler), but C. 1. grandis Hach-
isuka is upheld.

REFERENCES

Cain, A. J.

1954. Subdivisions of the genus Ptilinopus (Aves, Columbae). Bul-
letin of the British Museum (Natural History): Zoology, v. 2,
p. 267-284.

Delacour, J.

1951. Commentaires, modifications et additions a la liste des oiseaux

de 1’Indochine Franchise (II). L’Oiseau et la Revue Fran^aise

d’Ornithologie, v. 21, p. 1-32.

Delacour, J., and E. Mayr

1945. Notes on the taxonomy of the birds of the Philippines. Zoo-
logica, v. 30, p. 105-117.

1946. Birds of the Philippines, 309 p. Macmillan, New York,
de Schauensee, R. M.

1957. Notes on Philippine birds. Notulae Naturae, no. 303: 12 p.

Gilliard, E. T.

1949. Five new birds from the Philippines. Auk, v. 66, p. 275-280.

1950. Notes on a collection of birds from Bataan, Luzon, Philippine

Islands. Bulletin American Museum of Natural History, v. 94,
p. 457-504.

HachisuJka, M.

1930. Contributions to the birds of the Philippines, no. 2, part 6.
Ornithological Society of Japan Supplementary Publication, v.
14, p. 139-222.

1932. The birds of the Philippine Islands, v. 1, part 2, p. 169-439.
Witherby, London.

Hartert, E.

1925. Review of the genus Cacomantis Mull. Novitates Zoologicae,
v. 23, p. 164-174.

Humphrey, P. S., and K. C. Parkes

1959. An approach to the study of molts and plumages. Auk, v. 76,
p. 1-31.

Manuel, C. G.

1936. A review of Philippine pigeons, 2. Subfamily Ptilinopodinae.
Philippine Journal of Science, v. 59, 327-336.

McGregor, R. C.

1909. A manual of Philippine birds. Bureau of Science, Manila,
769 p.

340

Annals of Carnegie Museum

vol. 35

Parkes, K. C.

1958. A new race of the Blue-headed Fantail ( Rhipidura cyaniceps)
from northern Luzon, Philippine Islands. American Museum
Novitates, no. 1891, 5 p.

Peters, J. L.

1939. Collections from the Philippine Islands. Birds. Bulletin Mu-
seum of Comparative Zoology, v. 86, p. 74-128.

1948. Check-list of birds of the world, v. 6, 259 p. Harvard University
Press, Cambridge.

Rabor, D. S.

1959. The impact of deforestation on birds of Cebu, Philippines,
with new records for that island. Auk, v. 76, p. 37-43.

Rand, A. L.

1948. Five new birds from the Philippines. Fieldiana: Zoology, v. 31,
p. 201-205.

1951. Birds of Negros Island. Fieldiana: Zoology, v. 31, p. 571-596.

Ripley, S. D., and D. S. Rabor

1956. Birds from Canloan Volcano in the highlands of Negros Island
in the Philippines. Condor, v. 58, p. 283-291.

1958. Notes on a collection of birds from Mindoro Island, Philip-
pines. Bulletin Peabody Museum of Natural History, v. 13,
83 p.

Salomonsen, F.

1952. Systematic notes on some Philippine birds. Videnskabelige Med-
delelser fra Dansk naturhistorisk Forening, v. 114, p. 341-364.

1953. Miscellaneous notes on Philippine birds. Videnskabelige Med-
delelser fra Dansk naturhistorisk Forening, v. 115, p. 205-281.

Verheyen, R.

1953. Contribution a l’etude de la structure pneumatique du crane
chez les oiseaux. Bulletin Institut Royal des Sciences Naturelles
de Belgique, v. 29, no. 27, 24 p.

1956. Contribution a l’anatomie et a la systematique des touracos
(Musophagi) et des coucous (Cuculiformes). Bulletin Institut
Royal des Sciences Naturelles de Belgique, v. 32, no. 23, 28 p.

Witherby, H. F., F. C. R. Jourdain, N. F. Ticehurst and B. W. Tucker

1938. The handbook of British birds, v. 2, 352 p. Witherby, London.

ri

_ m w 40

inXIXSNrNVIN0SHllWs"s3 i HVB a IUIB RAR I ES^SMITHSON IAN ~ INSTITUTION

■a 3 3 /#% §

8 8 TOT 8 i» 311 8

2 5 v ^ > N^jgS^X ^

RARIES SMITHSONIAN INSTITUTION NOIifUliSNI NVINOSHillAIS^Sa I BVB 8 IT

2 w _ =; . <o

miliSNI NVINOSHIIINS S3IUVU8n LIBRARIES SMITHSONIAN INSTITUTION

> 2 y“ 2: r-

» iOH I

w ' w 2 m xgyosv^ g rn

RAR I ES^SMITHSONIAN^INSTITUTION^NOlirUllSNI^NVINOSHIIINS^SH IBVB 811

2: 5 \

«P 1 tJLlf I 1 I life-

s ;.^v > '* s

LfUllSNI^NVINOSHilWS^Sa I B VH 8 II^LI B RAR I ES SMITHSONIAN ^INSTITUTION

“ 5 co =

RARIES SMITHSONIAN INSTITUTION NOIifUliSNI NVINOSH1IWS S3IHVM9IT

Z r* _ z r~ ^ I

« - - — - 5J

XfUIXSNt NVINOSHillAIS S3iaVUaiT LIBRARIES SMITHSONIAN INSTITUTION

2 ___ W 2 _ . CO 2

SS *^*v*rx

X

CO

o

2

2 > 'W* 2 > X^oxii^X 2

RAR I ES^SMITHSONIAN INSTITUTION^NOlIfUIXSN! NVINOSHJUWS^SB I BVB BIT

m ^ co ~ </>

o

LfUllSNI^NVINOSHlIIAIS SBIBVBQIl^LIBRARI ES^ SMITHSONIAN^ INSTITUTION

2 r- 2 r-

±1

m '<ass>' £ NjftaiBgX m ^ jjj NffuiBqy m

iliSNI NVINOSHJ.IINS S3IUVH a n~L IBRAR I ES^SMITHSONIAN-INSTITUTION^N'

21 _ CO Z . CO 2 w CO

CO 2 CO ^

< 2 ,,. < 2 < ^S v, S

x
co
O

z

2 S ^ > XtOiiiAr^ 2 ^ >

CO z <o * z <o fc z

^RIES SMITHSONIAN INSTITUTION NOIinillSN! NVIN0SH1IWS S3 I H VH 8 II L

c n _ <n — m

o v “ N^Q-tus^y q

IliSNI NVIN0SH1IINS $3liJVHan LI B RAR I ES^SMITHSONIAN^INSTITUTION ^Ni

I v Z r* z r- z

s m ^ 2 00 8 £ ^otbx o

rn v -V??*’ 2 x^asvv^ frj x^osv^x m

\R I ES SMITHS0NIAN~INSTITUTI0N^N0lifUli.SNrNVIN0SHillMSt,>S3 I aVB a 11 ”LI

LU.SNI_NVINOSHJ.IWS S3 I UVB 8 IT LIBRARIES SMITHSONIAN INSTITUTION
r-^ 2 \ ^ 5 < /> . = €/>

^RIES SMITHSONIAN INSTITUTION NOflfUIlSNI NVINOSHIIWS S31BVH8I1 L

z ^ _ z r~._z *“

fn 'v^.dc^ <o m '^v z m

iliSNI NVIN0SH1MS S3 I H VH 0 n~LI B RAR I ES^SMITHSONIAN”lNSTITUTION^ N'

Z _ _..- <2 2 V..' • CO z .V' W

< 2 .< ^\v, 2 xtera-K < AN - 2