tiie iG }
a ie Hi HE

8

Heys

witeeh

f He oF
eatats
bas abies sett
”* eT

fore)
rer erss

7,
sartryeenit

.
.
.

-
‘

“.
.
es
.
<

drome oe
aren he art

4060-6 -6bebs 58k

roo yh Sarees

iz

porere ra
ies

4

pert

Frtee

: bh atytin
Sa

: riteatte et
j sisteTsttiaaate:

rerires

wecrtasi rs

th
adit

— Fee UG EE Ue 6 CllUe weve es Fe lew eee be eeovrwwervweeoweFeewsY Ct Fae PCa tate wt ss PF Fee rE tae FE FOE AD wre OY OS

i! e us Z us
oe 2 a a oc
<x c < = <x
ce 4 = as o !
= . S = 3 ee 5 i
NOILALILSNI NVINOSHLINS 4 1yYVvVug aah LIBRARI ES) SMITHSON ee |
fe = a fs) Yy = 2
Be = 2 as 4, ® .
> — > -_ os >
mi i e a C7 2 E
_1.BRARI ES SMITHSONIAN INSTITUTION NOILNLILSNI _NVINOSHUINS = Iyvusit 3 -
4 = i = : < = :,
4 = 4 o an
Z e 2 Z. a
> = >’ > =
= 7) z z on.
NVINOSHLIWS LIBRARIES INSTITUTION NO
= os: 5 an
” a ” Ww
= oc rz oo
va _< cr. <
Cc ce c ft a
= co = oo
2 A = < a
.I.BRARIES SMITHSONIAN INSTITUTION NOILONLILSNI NVINOSHLINS S31INWY git Lt

“yy

NOILALILSNI NVINOSHLINS

INSTITUTION NOILNLILSNI

INSTITUTION
“
INSTITUTION

S S3iuvusiy

SSIYVYEIT_ LIBRARIES INSTITUTION

i

NVINOSHLINS S3I1YVUgIT LIBRARIES SMITHSONI

NVINOSHLINS S3IYVYGIT

K&
NX. 5
<<
SMITHSONIAN

SMITHSONIAN

SMITHSONIAN

: &

SMITHSONIAN _INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYSIT

“
¥
\
Y

LI

NY

NOILNLILSNI NVINOSHLIW

LIBRARIES

NOILALILSNI INSTITUTION NO

SaIuYvud!IT LIBRARIES

17 LIBRARIES

NVINOSHLIWS

INSTITUTION NOILNLILSNI

INSTITUTION NOILNILILSNI

Saluvyugit
saiuvua
INSTITUTION

INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVUSIT Lit

NVINOSHLINS S3IYVYEITLIBRARIES

z 7) z n z

< s Sy, te = Y; =

a) iM i fy, ee s z=

S \s > yy 5 = Y, Yo. O

” \ . WD YY ” wn . G&

EN ro MOAN oO die = 0 GA Ls

s ESN 2 ee i ey i

ea: 3 - z
VOILNLILSNI_NVINOSHLIWS $3 1yYVvVuYg ae BRARI ES SMITHSONIAN _INSTITUTION | N :
ASTRO a re <: YON a ELLY co Vy ; = % 22 No> = :

ary A Kw OT GATS 3 Kem —- VY Jy 3 {SVP =

2 MS 5 ey 2 ea).

. = w —

ep MITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S31iyyvyugiy
= ae = a
nl z ar z

” ” ” ome 72)

oO 2 ae: ode

= = Z, =

> = >" =

= ” Pe ”

NVINOSHLIWS LIBRARIES SMITHSONIAN

S

Sait

Zz
A Sy
FiINosB

INSTIT!
SC

e

NG ON

NVINOSHLINS SJIYNY

LIBRARIES

B 2

INSTITUTION NOILLNLILSNI

NVINOSHLIWS
SMITHSONIAN

¢

.

= = ” == ” =
a = 5 uw a W o
= o = oo “i = oc =
= _< = << ac = < c
: : : : : : :
O = S) Se S) = fe)
a <i Fa = a —! z
SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S23] YUVUGIT LIBRARIES |
= - aes oO
= tif a * a E a = = =
=] es we > 2 a 5 = =
- “GY, pid F fea = on oS
2 i = m is m :
Ee = ” a= _
_NVINOSHLIWS (S23 Iuvug gh op °MITHSONIAN _ INSTITUTION |, NOILALILSNI |
< = ee = ae <
5 a ul fy? : Sa as
S aY lp > e EUS 5 SHY 3
= Oo Ui: xr fe) s fe . cE
2 “iy = = = 2 =
= = j = * = = as > -
pO OMAN INSTITUTION NOILNLILSNI_NVINOSHLIWS $3 IYVYSIT_ LIBRARI ES
4: = = ~ ASS eS x -
ie c te: ra RW Pear ek a <i
: : : =e ae ? :
a Yi re) a oO = ) z=
=| = ol z rs a3 Fas
—NVINOSHLIWS | S3 R- a ees A LIBRARI ES_ SMITHSONIAN _ INSTITUTION = NOILALILSNITN
; : z — Gy? = 3
> a > = LEE = ‘= Eas
. = 2 — © £7. R = 2
4 z m Eon nee. = m
~~ w~ = wm ea ~
p> MITHSONIAN INSTITUTION NOILALILSNI NVINOSHLINS SJ fuvyeai? LIBRARIES 2s
- ea w | = 27) Po ”
= < = vy < = < =
— os QA —
: S = YY 2 ; 3 :
3 D & Uf Uf 2 g f g
= ES 2 Ge Z Ee =
; = > = >" = >
= 77) —- ”o z ” “ yA
-NVINOSHLINS | LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI
Z => > “” z
a iw wW , Lu
: c = = Uy? : :
2, << = <x YY a] <x zs
= ow e ow we, ox S
5 on” = = Sie a a =
=~ Oo Bile ro. aa Oo
= = = a3 Fo a z
SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3I1yNVHSIT LIBRARIES
c 5 : : : g
= ae oO a md
: = = 3 = a F
¥ : : = : : 3
™ — = za = Zo =

THE QUADRAT)/ METHOD
OF /STUDYING’

SMALL /MAMMAL POPULATIONS»

SCIENTIFIC PUBLICATIONS

OF THE
CLEVELAND MUSEUM OF NATURAL HISTORY
Vol. V, No. 4, pp. 15-77
Issued, December 28, 1939
CLEVELAND, OHIO

Ma ny if) '

THE QUADRAT METHOD
OF STUDYING
SMALL MAMMAL POPULATIONS /

CLEVELAND, OxIO
1939

(Some of this work was submitted in partial
fulfillment of the requirements for the
degree of Master of Arts at the Graduate
School of Western Reserve University)

GMAT OP 7

NOV 16 235

4

LibRAT >
=

a

ty. 7
~ A

oes

Ly ‘ mf i Al a
ry i We ik
:) na Rif

ta

a
iy
|
?

a

if

a Mi ay i"

NACE

i es : eh SILA ae

TABLE OF CONTENTS

Page
ieisemouaiales amd PPOUGeS syle oe kee oi ais adie) wiel lo va ee aes a 18
Hem CCLOME | ere yy ate ici ee ri eases eae et aaa Maa esy oe tiatala 19
| LAV ISIPENECIITE 5, Metal eR Mn fe are ee ee es eee pen 20
Quadrat Work of Investigators
momebnc Cleveland Museum...) 22. 04. cen. Aaa: 23
SMILE OGS Cue nine. tye leet Nil, gra cil alee Nan Shai ahi Ge 25
Comparison of the Yields of Different Sizes and Types
of Quadrats
1. The Yields of 10-Meter and 150-Foot
@nadtats; Compared a cy.1iirga mage ne 2 Baas aa Wk: 26
2. The 50-Foot and 10o-Meter Quadrats Compared........ AG
3. The 50-Foot and 150-Foot Quadrats Compared........ 35
4. Experiments with 50-Foot Quadrats Insulated with
SLD tae Lefoy a RI ten Gene ee SE a Ae Ra NERO 44
Pomuhe To°0-Foot Quadrat,,o:22.9 ACte. 2... dice. ae 2 yng 48

6. Circular Units: Their Advantages and Disadvantages. . .50

Discussion
iiwaunicaters Law Of Frequence) 2). liso. ea ee. 51

2. General Considerations

GD Grie Rates, (Baitimes ete). sho cecuy ie wre noetae oe 51
Populations (Annotated List with Preliminary Discussion)...... 64
Sree RASAGHIS MAE ee pe ee yaa ye MY Lk hale Sd sae wt NaN dich Sg 74

TasBxe I.

Taste II.

Taste III.

Tasrez IV.
TABLE V.

Tasie VI.

Taste VII.

Taste VIII.

Taare IX.

ABER, OX.

Tasie XI.
Tasie XII.

LIST OF TABLES AND FIGURES
Page

Summary of the Total Yields from 10-Meter Quadrats, North Chagrin
Metropolitan Park, Cuyahoga County, Ohio..-.....-.....5..:ea: 27

Populations-per-acre based on Total Yields from Five 1o-Meter Quad-
rats and the 3-Day Yields from 150-Foot Quadrats in Beech-Maple

Forest at North Chagrin Park. 22.3... 0... es.505 52s eee 30
1935 3-Day and Total Yields from the Five 10o-Meter Quadrats at

North Chagrin Park. | 5.2 0'). 9 Ga. Le 31
1935 Record of the 150-foot Quadrat at North Chagrin Park........... 32

Comparison of Population Estimates based on Yields of the North
Chagrin 1o-Meter and 150-Foot Quadrats in August, 1935......... 33

Population Estimates based on the Yields of Fourteen 150-Foot Grids
(each of nine 50-Foot Units), and on the Yields of the Central
Quadrats of these Grids... 5. oc Deb eeea woke cee oe ee 34

Summary of Quadrat-Trapping Work on Little Mountain, Geauga
County, Ohio, in 1933)... . i234 sent oe pe ce ale ee 39

The Yields of Three 150-Foot Quadrats on Little Mountain for the
Years 1933 and 1935) )2 .. oo 3e'2 oe deste eiolde ee ee 40

Populations-per-acre in Different Forest Communities on Little
Mountain. Figures Based on the 3-Day Yields of 150-Foot Grids
and their Central So-Foot: Quadrats.:> 5. 02522 eee: 2 eee 41-42

Checklist of the Mammals Mentioned in This Report............... 43
Yields of Control and Insulated Quadrats (50-Foot Size) Compared... . 47

Climax Forest Percentages of Frequence of Small Mammals in the
Cleveland, Ohio Region, 1935-39, and their Distribution into Groups
According to Raunakaiet’s Law. J:5, 600-0 oe eso eee 53

TaBie XIII. Drift Rates of 1o-Meter and 150-Foot Quadrats Compared to Each Other

TABLE XIV.

FIGURE I.

FiGure 2.

FIGURE 3.

and to Certain Basic Quadrat Relationships...................... 55

Small Mammal Populations in the Cleveland Region, 1935-39...... 72-73

Population-estimate-per-acre plotted Against Quadrat Size, the
Yield: considered Gonstaats / 205 2,22 Se ee Pee 54

Average Daily Yields of Small Mammals from Four Different Sizes of
Quadrats: Operated. 1932-39 Inclusive... .ic..2555. 2200. Shee 58

Average Daily Yields from 150-Foot Quadrats for each of Six Different

SCIENTIFIC PUBLICATIONS

OF THE

CLEVELAND MUSEUM OF NATURAL HISTORY
Vou. V Issu—ED, DECEMBER 28, 1939 No. 4

INTRODUCTION

In 1932 The Cleveland Museum of Natural History began a
series of intensive ecological studies of different habitats in the
Cleveland region. In each of these, the problem of small mammal
abundances immediately presented itself. Williams (1936) and
Aldrich CMS.) attacked the problem by using trapping quadrats
of small size. In the following year, 1933, the writer and his co-
worker, Mr. P. N. Moulthrop, took over the quadrat trapping
portions of their investigations and started a comparative study of
the yields and value of different sizes of quadrats. The study has
been carried on exhaustively since that time. Conclusions have
been carefully checked through several years of experience and the
results are now ready for publication.

The writer is deeply indebted to Dr. S. Charles Kendeigh, of
the University of Illinois, for his guidance and advice during the
course of this work, which was partially completed while Dr.
Kendeigh was a member of the faculty of Western Reserve Uni-
versity; and to Drs. Arthur B. Williams and John W. Aldrich, of
The Cleveland Museum of Natural History, for their advice and
generous permission to incorporate some of their data in the present
paper; and lastly, to the unflagging interest and zeal of the writer's
co-worker, Mr. Philip N. Moulthrop, and his several field assist-
ants, especially Mr. Scott R. Inkley and Mr. Winston Jesseman.
This paper would have taken many more years to attain even its
present preliminary status had it not been for their co-operation and
industry in the field. Further acknowledgments are certainly due
the host of property owners who have permitted repeated invasions
of their lands by field workers from The Cleveland Museum of
Natural History. In this connection, our especial thanks go to the
Cleveland Metropolitan Park Board, to the Little Mountain Club,
to the Baldwin Bird Research Laboratory, and to Messrs. Woods
and Charles King, Trustees of the Cleveland Museum.

Sci. Pub.

2O THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

LITERATURE

The need for some method of estimating mammal populations
has long been recognized. Various workers have in recent years
emphasized the lack of standardized quantitative methods; others
have gone farther and have suggested different ways in which
counts may be made. The first concrete suggestion for a basic unit
for population estimates was that of Grinnell (1914, p. 92) who
advocated the use of ‘‘trap-nights’’. The total number of traps set
in each locality is multiplied by the number of nights the trapline
functions. The yield for each species per trap-night is then figured
out, or the yield for all trap-nights. This method is admirably
adapted to the needs of the field collector of small mammals. It has
been frequently followed by various writers in subsequent years in
determining the relative abundance of species in given districts; but
there are several difficulties connected with its use, as suggested by
Dice (1931, p. 377), to wit: the skill of the individuals setting the
traps varies considerably; the yield decreases after the first night in
most cases; the weather, make of traps, age of traps, and the species
of animal, all afford factors tending to affect the significance of the
total yields. Another difficulty might be noted from the ecologist’s
point of view: trap-nights give no indication of the actual abundance
in numbers per unit area unless combined with other methods. Dice
Coc. cit.) himself used ‘“‘first trap-nights’’ only in some of his work,
but concludes that, owing to the fact that certain species come to
traps earlier and more readily than others, they do not give fair
estimate of the relative abundance of the various species in the area
trapped.

Besides using trap-nights per unit area, Taylor (1930, pp. 523-
542) made counts of different species along lines of specified lengths
and also made counts of the animals killed around standard amounts
of poisoned grain. In addition, he used quadrats in making counts of
hares and rabbits. These quadrats were designed to prevent ingress
and egress of only the larger rodents. He also counted fecal pellets
in given areas, and the numbers of used burrows and runways, a
method also advocated by Phillips (1931, pp. 633-649), who sug-
gested the basing of population estimates on the average number of

DEC:

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 21

animals in certain sample lairs or burrows, the average number per
sample lair being multiplied by the total number of lairs per unit
atea. Recently MacLulich (1937, pp. 25-39) has shown that counts
of droppings in a series of small quadrats can give a reliable index
of the relative abundance of certain species. Grinnell and Storer
(1924, p. 22), suggested using time units in place of space units, the
species or individuals captured or seen per day or hour being substi-
tuted for areal units. The writer has found this useful in obtaining
figures on the abundance of diurnal species, but the method is
obviously under limitations when used at night, and is subject to
the full brunt of varying conditions and possible interpretations,
due to the weather and other factors above mentioned. Dice (1931)
suggested that the use of indices affords the best method of express-
ing the relative abundance of species. He would list the number of
individuals of each species observed or taken in each ecological
habitat, and would base his indices on the frequency of each species
in any one habitat. He stresses that the indices must be specific for
each habitat and that the different techniques necessary for the
capture of different size-groups of mammals must be taken into
account. Subsequently, he has abandoned the use of indices in favor
of population estimates based on the computation of home-ranges
from the data yielded by large-size quadrats (1938, pp. 119-130).
Hamilton (1937, p. 780) secured figures on the populations of
Microtus pennsylvanicus about Ithaca, New York, by trapping all
the mice out of areas “‘more or less defined by natural boundaries’’,
varying from one-half to twelve acres.

Quadrats have long been used as a means of determining the
abundance of plants in ecological studies. For an account of the
development of the quadrat method, the reader is referred to Kenoyer
(1927). Thus Pound and Clements (1898) developed a scale of seven
degrees of frequence, based on quadrats five meters on a side. Jaccard
(1912) used adjacent quadrats as bases for a frequence index. Using
much of the early work in addition to his own, which involved the
use of circular units instead of quadrats, Raunkaier (1918) developed
his ‘‘Law of Frequence’’. Briefly, this is explicable as follows: Given
a number of quadrats, perhaps ten; and a number of species, ten for
instance; if one species occurs on all ten quadrats, its percentage of
frequence is 100; if on but one, 10. There will be a higher number of

Sci. Pub:

oly 8 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

species occupying a small proportion of the quadrats than of species
occupying most or all of them. The ratio of these frequences will be
5 :2:1:0:1, where each number refers to the frequency for each of five
classes of abundance (classes A, B, C, D, E, referring to the occur-
rences of a single species in 1-20, 21-40, 41-60, 61-80, 81-100 per
cent of the quadrats, respectively). As Kenoyer (1927, p. 343) points
out, this may be expressed by a graph, wherein the curve represent-
ing the numbers of frequences has two crests, one much lower than
the other. The first crest represents the number of least frequences
and the second the number of the greatest. Raunkaier himself based
this concept on over 8000 percentages of frequence, which arranged
themselves according to the above formula. Kenoyer found that his
surveys of vegetation in the Lake Michigan region supported the
Raunkaier formula. From the point of view of the developmental
history of the subject of this paper, it is of importance that he (1927,
p- 236) suggested the use of this formula and, by inference, the
quadrat method, with regard to birds and mammals. He found that
the formula held very well for sweeps of the insect-net over known
grass-land areas, and for micro-organisms in hay infusions. The
applicability of the formula to small mammals will be discussed
shortly.

A number of other workers, in addition to Taylor Coc. cit.),
have used quadrats in studying small mammal populations, among
them Elton, Middleton and their co-workers (several papers),
Townsend (1935), Chitty (1937), Williams (1936), and Hatfield
(1938). Townsend’s paper appeared after the initiation of this work.
In it is presented a somewhat novel method of obtaining actual
figures for small mammals per acre of ground. The method devised
was that of the ‘“‘moving quadrat’’ that consisted of three parallel
trap-lines one rod apart. Each trap-line consisted of a row of triads
of traps three feet apart. Thus the trap-lines were in all 33 feet long
and about of the same depth, forming a quadrat of a size inter-
mediate between the 1ro-meter and 50-foot sizes, later to be con-
sidered in this report, but much nearer the 10-meter unit. The trap-
lines were “‘cascaded’’ by shifting the rearmost row to a new
position one rod in front of the first row. This was done every day
for fourteen days, during which time a stationary trap-line was run
in the same habitat. The idea was evidently suggested by the work

Dec.

1939 B. P. BOLE, JR.— THE QUADRAT METHOD 23

of Elton (1925). Populations per acre were estimated on the basis
of the yield over a three-day period.

Chitty (1937) used blocks of traps (Tring live traps) spaced so
as to form a quadrat go feet on a side. He concluded that there was
a large amount of drift into units of this size from adjacent terrain,
and later used areas ‘‘9 or 10 times as large’, concluding that even
these larger units are subject to considerable drift. Hatfield (1938)
used small quadrats (24 feet on a side) in studying the habitat
preferences of certain small mammal species in Minnesota. Williams
(1936) made estimates of population per acre on the basis of the
yields during several days from quadrats 10 meters on a side. It was
his four-year study of the North Chagrin forest near Cleveland that
stimulated the present investigation. His original population studies
were made with but one size of quadrat. As the writer took over the
quadrat portion of Dr. Williams’ researches in 1933, all of the
quadrat trapping data secured by workers from the Cleveland
Museum is incorporated in this paper, including that which has
already been published by Williams (1936).

Dice (1938, pp. 119-130) used quadrats in the course of studying
small mammal home-ranges. His quadrats were of single chains on a
side (a chain = 66 feet) and of multiples thereof. From the data
obtained through the use of live traps and the marking of his
animals, he was able to derive an interesting formula for the rough
calculation of an individual home range: NX* = (A + X) (B+ X),
where N is the number of adults of the sex or age group being
considered, known to be living on the quadrat, X the mean width
of an individual home range, A the width of the quadrat from out-
side trap to outside trap, and B the length of the quadrat. The difh-
culty arising from this method is that the quadrat must approximate
the actual home range or the results are subject to considerable
error.

QUADRAT WORK OF INVESTIGATORS FROM
THE CLEVELAND MUSEUM

Dr. Williams Cloc. cit.) was primarily interested in the ecology
of sixty-five acres of virgin beech-maple forest situated in the north-
eastern corner of Cuyahoga County, Ohio at the North Chagrin
Reservation of the Metropolitan Park System of Cleveland. His

Sci. Pub.

24 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. a)

trapping of small mammals there constituted but a small part of
his investigations. He used five quadrats of small size — ten meters
on a side, apparently following the suggestion of Shelford (2929,
Pp. 55). These are referred to hereafter as 10-meter quadrats.

Aldrich CMS) also used the 1o-meter quadrats, singly and in
blocks of nine (thus forming a quadrat 30 meters on a side), in
studying the small mammal populations of various communities
of the hydrosere at Aurora Pond, Portage County, Ohio. His work
began during the summer of 1933 and continued through 1934.
No trapping was done at Aurora Pond during 1935, but since that
year further studies, with larger quadrats, have been made annually
by the writer and Dr. Aldrich.

Believing that the use of units as small as the 1o-meter quadrats
greatly increased the probability of error when they were used as
bases for the estimation of actual numbers, the writer tried a number
of different methods. In 1933 fifty-foot and 150-foot quadrats were
operated synchronously at a number of stations and the results
carefully compared. In 1934 the 50-foot and 1o-meter sizes were
compared, and in 1935, the 150-foot and 10-meter sizes. In 1936 the
conclusions arrived at were carefully re-checked by another ex-
haustive comparison of the 1o-meter and 150-foot sizes, on a far
wider scale than heretofore. The smaller sizes were discarded there-
after, and work was continued with the 150-foot units. In 1938 the
desirability of a “‘one-man quadrat’’ was fully recognized and trial
was made of 1oo-foot units. In this year, and 1939, circular units of
approximately the same area (1% acre) as the 150-foot sizes were
tried, upon the recommendation of Dr. S. C. Kendeigh.

Various other modifications were also attempted: protective
‘‘shelter-belts’’ of traps and quadrats, insulated from the in-wander-
ings (drift) of small mammals by sheet-iron sunk into the ground
to a depth of several inches. Full study was made of the yields of
central quadrats in blocks of nine. A similar system was used by
Elton, Davis, and Findlay in estimating the population of voles in
Scotland (1935, p. 284). All these studies were made in many dif-
ferent habitats, and the quadrat-behavior of thirty species of small
mammals was noted. The scientific names of the various species
concerned are given in Table X. Elsewhere they appear as binomials.
For the sake of uniformity and brevity the subspecific names are

Dec:

1939 B. P. BOLE, JR._-THE QUADRAT METHOD 25

omitted. In a few cases, as in the daily yield tables, more than one
subspecies of the same animal are concerned in the averages, and in
several others, the actual identity of the subspecies in question is
still in doubt.

The population estimates as derived from each quadrat are given
to the nearest whole animal; thus, a yield of 1 Blarina from a 150-
foot quadrat is interpreted as 2 animals per acre instead of 1.94
animals. The average figures are given, however, to the nearest
tenth or hundredth of a point.

FIELD METHODS

The method of quadrat trapping followed by workers from the
Cleveland Museum of Natural History is as follows: a square of land
of known area is marked off. Traps are set over the entire square, the
number varying according to the needs of the terrain. They are
placed in such a way that they will as quickly as possible drain the
area of its resident and transient small mammal population. The
traps are kept in operation for three days or until there is no further
yield, or until the daily yield has reached a minimum. A field map
of the quadrat is made, showing trap-sites, rocks, logs, etc. The
record of each trap is kept on the map, while the daily catch is
added up and tabulated on the back of map or on a separate sheet.
When specimens are saved, the quadrats they came from are re-
ported on the field tags along with the date and other data. Many
different sizes and groupings of quadrats may be used, and the
relative efficiency of the different methods varies. From the yield
obtained per quadrat the number of animals per acre, per square
mile, or any other large area, may be computed, providing the
ecological conditions of the large area are uniform and similar to
those in the quadrat.

In forest quadrats, ninety-five percent of the traps are set in
underground runways which may be found by rapid probing or tap-
ping with one’s forefinger, and which can often be detected by an
experienced field worker simply by walking over the terrain of the
quadrats.

The number of traps set per quadrat depends entirely upon the
terrain, but under all ordinary circumstances at least 12 traps are

Sci. Pub.

26 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

placed in each 10-meter unit and at least 100 in each 150-foot (half-
acre) area. Traps are mever set at fixed intervals within the quadrat;
they are placed where they will catch the mammals. Every effort is
made to reduce the bait factor to a minimum by setting the traps
in such a way that the mammal will be caught whether the trap is
baited or not. Local conditions — that is, the underground run-
ways, in which most traps are set — make bait unessential for the
capture of all the insectivores and all the voles considered in this
report. Bait (rolled oats or dried fruit) is usually used, however,
except in the rainiest weather, in order to attract certain other
species, notably the small sciurids and jumping mice, which do not
habitually use runways. For the bait preferences of certain small
mammals, the reader is referred to Townsend (1935, pp. 22-28).

The specialized technique used when we operated quadrats
insulated from drift is considered later on in this report in con-
junction with the data yielded by that type of quadrat.

COMPARISON OF THE YIELDS OF 10-METER
AND 150-FOOT QUADRATS

The five 10o-meter quadrats used at North Chagrin Metropolitan
Park were operated for the purpose of studying small mammal popu-
lations per acre and population fluctuations from 1932 to 1936 in-
clusive, by Dr. Williams. Data obtained in all the years except 1936
have been published by that worker (1936, pp. 46-47). That for 1936
appears for the first time in this paper.

North Chagrin Metropolitan Park is very fully described as a
Beech-Maple climax community (loc. cit.). Suffice it to say here
that two of the five 10o-meter quadrats (B and D, see Tables) were
situated on points of land overlooking the valley of the Chagrin
River, in situations where hemlock was an important member of
the plant community. Conditions of soil and cover were typical of
the forest as a whole in all but one of the quadrats (A) where there
was a deficiency of cover and humus. Quadrat C was encompassed
by a grid of nine 50-foot quadrats in 1935 and 1936 in an effort to
study the effect on its yield of a shelter-belt of traps designed to
exclude non-resident animals that drift into such small units during

B. P. BOLE -__THE QUADRAT METHOD gx
)

se By
Taste I. Summary of Total Yields from 1to-Meter Quadrats, North Chagrin Metro-

politan Park, Cuyahoga County, Ohio. [In Part after Williams (1936, p. 46).]

Bec:

sopadg ITy
sndvzoavde Ny

sndez

skur Arg
SNIOIOTY

‘onoy snosAuro19g
“spny snintog
SUTUI ],
sdojvosvivg
SN9IOUTI X9IOS
snourny X9109

eulielg

IwoX

qwaipene)

18
13
16
Io
II
72.
Iz
5 fe)
fa
67

18

13
53
13

12)
(e)

: NO rat traps used.

LzEO
are e
bay
sbee ED.
one
Bau

007
Le BS
See.
aa.
20)S ao eee a
EOP Hse.
ESB) =<
LOR AR i.
SRS). 2h
EQ36.. «..
E532: .:.
POSE 5... -
EGIA. 2s
Ee Seo se «2
ERG)...
Cy ie
EQIGr «<= -
1939*...
BGI.»
533". .
EQ3A.-.
EQ35.-~
rear...
Ope 0 ae
ne ae
EQZA i. :
BORG. «
EOZG. 2.

All Quadrats
1 oy eae
eee a eee
BORA se bie cies
SCOT ERRS CRE es eee
EQ3O55 <2. 62 =
* See p. 26-28.

A.
B.
C.
D.
E.

Sci. Pub.

28 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

trapping operations. Hence, its yields are not strictly comparable to
those of the others during these two years, but are of interest
nevertheless.

Table I gives a summary of the total yields during approximately
twelve days time from these 10-meter quadrats. Total yields are here
given instead of the yields during a shorter period because the table
is an enlargement of one used by Dr. Williams (1936, No. 12, p. 46),
who based his population estimates on the sum and total of all
specimens taken during the period of trapping. Evidence is given
subsequently in this paper that estimates should be based on the
yields during only the first three days of trapping. It will be noted
that the lack of humus and cover characteristic of quadrat A had no
apparent effect on its yields, probably for the reason that it was
well within the normal wandering range of numerous small mam-
mals whose incursions into the area were of temporary nature. It
will also be noted that the yields of quadrat C from 1935 on are
much lower than the yields for the corresponding years from the
other units, because of the “‘shelter-belt’’ surrounding it.

In these figures summarizing the results of all the quadrats
it may be seen that small mammal populations were great in 1932
and 1935, indicating that these years probably marked peaks of small
mammal abundances. Closer inspection of the data, however, re-
veals that the character of the peaks was determined almost entirely
by two species, the short-tailed shrew, and the deer mouse. The
more unusual species do not seem to follow the same curve.

In Table II are found population-estimates per acre for each
species, based in part on the figures in Table I. There the numerical
changes may be more easily visualized. It will be shown subse-
quently that these figures are in all probability far higher than they
should be, but as the sources of error are common to all, the figures
ate comparable to each other with the possible exception of those of
quadrat C from 1935 to 1939. They show two definite peaks in the
populations of at least two species, and provide the first evidence
of a short-wave cycle of about three years among the woodland
small mammals of our latitudes in North America. Subsequently,
as is shown in Table II, evidence was gathered that seems to show
that Blarina is either non-cyclical or has a cycle of a length as yet
not definitely determined. The figures of Townsend (1935, p. 48)

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 29

give population-estimates per acre for the years 1931-1934, but do
not entirely bear out the existence of such a cycle in all communi-
ties. The increased populations in certain habitats during the years
1931 atid 1932 he ascribes to factors influencing the moisture and
hence humidity of the community, and cites figures showing that
in certain habitats, such as very moist woods, there is a remarkable
constancy of populations from year to year. He believes that certain
small mammals may migrate into the moister areas during dry
periods, a feature noted by numerous other workers, including the
writer. It is to be expected, however, that cyclical changes in small
mammal populations are more likely to be noted in areas not
permanently suited to small-mammal occupancy; and the writer
believes that Townsend’s figures lend considerable support to the
fast-accruing evidence of the existence of cycles for certain species.

In 1935 a grid of 50-foot quadrats, numbering nine in all, actu-
ally constituting a quadrat 150-feet on a side, was operated at the
site of 1o-meter quadrat C, which formed the center of the central
unit of the grid. Such grids had been used as early as 1933 by the
writer, but it was not until two years later that grids were operated
at the same station with 1o-meter units. The purpose of the grid was
to study the yield from the center quadrat, which it was hoped
would provide accurate figures on small mammal populations from
an area protected from drift (by the outside quadrats). However, it
was soon discovered that it was virtually impossible to get a fair
cross-section of a given population from such central quadrats,
while the entire grid usually gave a very complete picture. The data
obtained in 1935 from the North Chagrin grid are given in Tables II,
DVenand), V.

The yield for the 10-meter quadrats (Table IID and the 150-foot
grid (Table IV) are compared in Table V. In all such grids treated
in this report, No. 5 is the central 50-foot unit and Nos. 1, 3, 7, and
g are corner quadrats. It will be seen that the five 10-meter quadrats
(Table III) captured nearly two-thirds as many animals within
three days as the grid (Table IV), although their combined areas
amounted to less than one-quarter of the grid’s area. Consequently,
the figures per acre for each species are very much greater when
based on the smaller units. Since drift into the grid has not been
completely stopped by any means, the figures obtained from it are

Sci. Pub.

30 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

Tasxz II. Populations-per-acre based on Total Yields from Five 1to-Meter Quadrats
and the 3-day Yields from 150-foot Quadrats in Beech-Maple Forest at North Chagrin
Metropolitan Park.

v Sie TAN Wee a ee: a
“ g Ss S : 2 5 9 a N
5 a OQ Gera. (BB Ce 6) See
3 oS 8) 2, B.Jhoe 8. 9 oi.
o a 88 & On 6 a £ 2 oe
1o-Meter.

EQRD AN Upskieadts stale se 296 P PGP GR aega 8 8
LOE We Mere Kobra ois 5 96 8 Pi GPiy CPIAB | Be aor ee

EQRA oa torsesouiag cre 48 8 P ee aes ere em es =

51(0 eB tc SOIR ES 184 P ZAP 16/216. 8 eae

LQRO. fe Seer as eeepc Bont. Gb 6S Oe Se 8 16
150-Foot

LORI erate ee cee 42 2 Oo woe te 32 PME
LOSS et aa. cease Ly 2 2 eit inee Mme 6 aS

LQ SAP Ss Amwns s anteriaterere Fei h2: 2 oe Ae aha 5 3

EOS SS 0. be ave sion n es 29 2 ARs (Re hea 17

BORO 6. Ais. she otis Says eyes 18 6 (24 ae Di 5 DE ake 2 2. as
TS), 17 Ne Er, AQ 0) 4) to PP ve 8 4
EQS Ose ecu eee eee 29 3 P mt Bah Tee 20 4
FOLQT Tae oe sae 12 5 a0 oe 2 4

* Figures from corrections made on 1o-Meter yields, the proper corrections
determined from the relative yields of 10-Meter and 150-Foot quadrats in 1935
and 1936 at this station.

** 2 quadrats in average.
Blank : norecords for year. P : population below 8 per acre (10-Meter) or 1 or 2
(150-F t.).

2 : present, but reliable figures lacking.

certainly not too low, even taking into account the fact that a small
number of animals escape, or drag traps into places from which they
cannot be recovered. Such occurrences are rare and are usually
omitted in compiling the data here given, although in the case of
moles, these are carefully taken into account. Elton (1935) and co-
workers, Hamilton (1937) and other workers have overcome this

Dec.

1939 B. P. BOLE, JR. THE QUADRAT METHOD 31

difficulty by the use of traps that captured their animals alive, a
method preferable from many points of view, but too time-consum-
ing for the worker who wishes to cover large sections of the country
in a short space of time. The grids, as was mentioned above, were
set up with the idea that the central quadrat would have the drift
eliminated and that an accurate picture of the inhabitants of the
quadrat could thus be obtained. This is undoubtedly true, at least
as compared to quadrats unprotected from drift; but an attempt to
base the total population of a large area on only one such unit gives
the anomalous results indicated in Table V. It happened, in the case
of quadrat C-5, that an infestation of Pitymys was centered in that
locality, and that the population of Blarina was somewhat denser
than usual. But in the first three days, no Peromyscus was taken,
indicating a total lack, in that small unit, of one of the most abund-
ant and characteristic mammals of the North Chagrin woods. How-
ever, if many such central quadrats are averaged (See Table VD),
from similar woods in different localities with the same general

Taste III. 1935 3-Day and Total Yields from the five 10-Meter Quadrats at North
Chagrin Metropolitan Park, Cuyahoga County, Ohio. Total Yields in Parentheses.
For the distribution of these quadrats, see p. 26.

Species A B & D E Totals
Blarina brevicauda..... 3 (8) a (Ga) Tce) 3 (5) 3 (4) Gs)
Peromyscus leucopus... 0 (3) 265) o (0) ING) 2 (8) 6 (24)
Pitymys pinetorum..... 0 (Co) 0 (0) Ge) 0 (0) L(G, 2G)
Sciurus hudsonicus..... 1 (2) 0 (0) o(@) o (0) 0 (0) HG)
Mamias sthiatus......... 0 (0) Ee) 0 (oc) o (0) o (0) HG)
Microtus pennsylvanicus 0 (0) o (2) o (0) o (0) 0 (0) o (2)
Totals, All Species..... AeGra)) sro) * 2G) 5G) 7G) 3G

Quadrat C was situated at the center of the 150-foot grid (See Table IV and
pp- 26-28). The drift it would normally have captured was collected by the outer
portions of the grid; hence, the yields for this particular quadrat were much lower
than those of the others. There is evidence to show that some of its residents drifted
outwards and were taken in the grid.

The area of the five 10-meter quadrats is 0.123 acre. The total trapping time for
each of the five quadrats was nine days in 1935.

Sci. Pub.
Vol. V

THE CLEVELAND MUSEUM OF NATURAL HISTORY

32

= “AVUIOTA DIVIPOWIUIT SIT UT UDYV} SvAA

‘s}vipenb pr3z ay Jo Aue
JO 3sa83v] 9y2 a3v S-> Wo sppatA dy2 IVY ‘JOAIMOY ‘9dULYD JO JO1IeUI & ST IT “TIT 2qeL, UE U9ATS se D ywspenb sJax9Y\[-OF JO} soy?
Qpnpour usar sy samy uy, “(TI Tq¥L pue 9z “d 299) 5 yespenb sax2y\J-o1 9y3 sopnpour avspenb styy, ‘prsd jo awspenb sayUID,

Gee Om Of. Or — 36" Coe. - © OY © t Ger srs iy sro,

(1) 0 (1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)O ****surjoa sAurooneT

(%)z (0) 0 (0) 0 (QE) 0 (0) 0 (0) 0 (2) 0 (OO HG) (C)O *"""**"snyeTIIs svTUTe Y,
(7%) © (1)0 (0) 0 (0) 0 (0) 0 (1) 0 (0) 0 (0) 0 (0) 0 (0)O ****snoruOspny sninqog
(t1) 6 (0) 0 (0) 0 (0) 0 Gz (S$ Cm (0) 0 (0) 0 G)1r +++ wmsojsurd skurAr1g

(9%) v1 (v) ¢ (1) 0 (¥) & G1 (v) 0 (Ge) Core (0) 0 (¥) € + -sndoonay snosAw019g

(1%) $1 (4)1 (Fy) t Ga (QD) (¥) @ya (CD) (0) 0 (9)¥ ++" epneorAsiq vurselg

s[eqOL, 6-9 8-D ae) Ce) «5D V9) £-9 ia) 12) satoodg

_——__—_——— ee

‘919% 91S'0 “prlIy jo vary ‘sasoyjuored ut 3039"] 2y3

‘spjarx (Avq-6) [ero], pu Avq-£ aya ‘(sa1uq 100,J-0$ ouTU Jo pIId)) IvIpeNd 100J-OS1 YIeg UlIsLYD YON 2Y3 JO pioda7y GC61 “Ay ATAVY;

Dec.

1939 B. P. BOLE, JR.— THE QUADRAT METHOD 33

Taste V. Comparison of Population Estimates based on Yields of the North Chagrin
to-Meter and 150-Foot Quadrats in August, 1935.

Blarina Peromyscus Pitymys Tamias Scturus huds.

Population per acre, based on
the 10-Meter 3-day Yields. 97 49 24 8 8

Population per acre, based on
the 10-Meter Total Yields. 184 216 24 24 16

Population per acre, based on
the 150-Foot 3-day Yields. 29 27 17 4 3),

Population per acre, based on
the 3-day Yield of C-5 only
(the center quadrat of grid) 53 Ce 85

Population of the 65-acre
area, based on the 10-Meter
2301257 Ate ea ee 6341 3171 1585 538 538

Population of the 65-acre
area, based on the 10-Meter
Motal Waelds:......5..... 11097 12602 1585 1076 1076

Population of the 65-acre
area, based on the 150-Foot
Bedyields.3..066 2.5.5 1890 1754 1224 254

Population of the 65-acre
area, based on the 150-Foot
MiaealVieldse iii o. bs ce'es 2856 3536 1632 254 254

population figures, the picture becomes more normal, and, as is to
be expected, the estimates per acre run on the average below those
obtained from 150-foot quadrat data; in fact, for several species in
Table VI there are no records at all for the center quadrats, a situa-
tion which may be partially duplicated if any other quadrat of the
grid is taken in place of the central one. However, the fact that the
outside quadrats always show a slightly better picture than the
centers indicates that drift is at work, and is an argument for a very
short trapping period. Thus, the number of grids necessary to pro-
vide good average data for the center quadrats is very great. Most
workers using quadrats have indicated the necessity of obtaining

Sci. Pub.

34 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

figures from a large number of quadrats when working on the rela-
tive abundance of organisms, as were Raunkaier (1918), Gleason
(1920), Kenoyer (1927), and Dice (1931 and 1938). This holds
equally true for workers interested in actual abundance. A difficulty
arises in the fact that in order to obtain sufficient records of central
quadrats from one and the same tract of woodland or field, which
is to be preferred over quadrats in different districts, a great deal of
time must be spent and the resultant depopulation arising from the
use of many grids would be considerable and conceivably might
seriously alter the ecological picture. Furthermore, as Dice points
out (1938, p. 125, it seems wasteful of time and data not to use the
information accruing from the sheltering tier of quadrats them-
selves. The utility of a central quadrat might be greatly increased if
enlarged from 50, to let us say, 150 feet or more, but modifications
in the shelter-belt must then be made or the usefulness of the method
will speedily collapse. Some further observations on shelter belts
will be made later on in this report.

Taste VI. Population Estimates based on the Yields of fourteen 150-Foot Grids,
each of nine 50-Foot Quadrats, and on the Yields of the Central Quadrats of these
Grids. All Data secured in 1938 in Climax Forests.

nn

=

“=
2. & r=

—
SS 3 & wn 2 Q, > S| 8D
oJ B » S = fo) rs = aA =
= Ss iss} &Y 12) na wi ps} a
3 uU A n = ¢ 3 S jo) = om
oO 7) ~~ 4 =) ° ro} 5) ~ q

= a 3 QQ a bard > SH = =) vo 'o) n
ra S vo ~” o acs ” ro n a, & ” Ss
Sem Ee area) A) en eas a
E ca US ey eee St oe gs 4 2 = on nay)
ee Sein - Oi ek nae 8 3 S Ps) = n o
QA c 77) : os
A Ou Ohaus al Ron ies Ss So = Se Bs) Wes
p OS Bier oe LON Amat aan es o ra BS ge ,s
Oo Ba! Oa pau (Onn eae CO len ov = Ay N- 4
Centers): 1256 10.19" 01) Tez OP) er, On 1G soy Mie er Ms te Oe wc Oa pS) 2's

Grids ss). (34. 29:7! 01142211 50328) 4e0) | 2-6) loeTA 22:2: 0128 22 QommAnanS

Dec.

1939 B. P. BOLE, JR.—-THE QUADRAT METHOD 35

THE YIELDS OF 50-FOOT AND 10o-METER QUADRATS

In 1933 the writer believed that the accuracy of a quadrat could
be greatly increased if the size of the unit used were increased by
using 50-foot instead of 1o-meter units. The area of a 10-meter
quadrat is but 1/41 of an acre, while that of one 50-feet on a side
is approximately 1/17. The increase seemed quite important and a
great deal of data was secured during 1933 and 1934 with the use of
the new size of quadrat. It seems unnecessary to compare in detail
the yields of the two sizes; it will suffice to report that in any given
year the yields from the 50-foot units were in general slightly the
greater, the yield of the larger units being to those of the smaller
as 6 to 5, instead of 2.5 to 1, as one might expect from the relative
areas of the two units. The average daily yield from the two sizes is
compared in Fig. 2. The fact that the 50-foot line is usually below
the 1o-meter line is due to the fact that the 1o-meter line covers two
““mouse years’’, 1932 and 1935, while the 50-foot data was almost
all gathered in 1933 and 1934, years of low small mammal popula-
tions. The point to be gained from this may be much better empha-
sized by comparing the 50-foot yields with those of the 150-foot
quadrats, the contrast in which is of similar order as that between
the zo-meter and 150-foot units. At Little Mountain, Geauga
County, Ohio, a series of 50-foot quadrats and 150-foot grids were
operated synchronously. As a check on the North Chagrin data, let
us consider that obtained at Little Mountain.

THE YIELDS OF s50-FOOT AND 150-FOOT QUADRATS
COMPARED

Trapping was carried on at Little Mountain with these two units
in 1933, an ebb-cycle year further characterized by great drought.
Since 1935 only the larger units have been used. The actual field
technique was the same as was used at North Chagrin and elsewhere.
The presence of Mr. Philip N. Moulthrop as field assistant, who had
done the bulk of the trapping at North Chagrin, tended to standard-
ize the trapping technique and eliminate error due to the varying
experience of different workers. This error is chronically brought

Sci. Pub.

36 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

up as an objection or a difficulty to be encountered in assembling
quadrat and other trapping data, vide Dice (1931) and Ruhl (1932).
In this connection, it might be well to mention that any differences
in trapping technique speedily manifest themselves in the yields of
the quadrats. In the course of this work, every field assistant is
given a rigid course of training before being allowed to trap inside
quadrats, and the yields of each trapper are carefully and con-
tinuously examined. The most efficient trapper is he who gets the
most small mammals of al] species in a given area in the shortest
time. No great deviation below this can be tolerated, and it has
been found that well-trained workers do not deviate in their out-
puts to any appreciable extent.

The numbers of the different sizes of quadrats used at Little
Mountain were: in 1933, eight isolated 50-foot units and three 150-
foot grids; and in 1935, four 150-foot grids only, and three grids
only from 1936 to 1939 inclusive. In 1933 and 1935 trapping was
done for comparable lengths of time in the early summer. From 1936
on the quadrats were worked for but little longer than three days.

Grids of quadrats were maintained as follows: one in a nearly
pure hemlock community, the ‘“‘hemlock block’’ hereinafter men-
tioned; one in white pine woods, the “‘pine ridge’’ block; one in a
second-growth oak-chestnut area on the central ‘‘saddle’’ of the
mountain where pines are invading, the “‘pine seedling’’ block; and
one in a beech-cucumber area, the ‘“beech-cucumber’’ block. Except
for the lattermost, all were worked in both 1933 and 1935. The
beech-cucumber block was completely worked only from 1935 on,
although two of its corner quadrats were worked in 1933 as isolated
50-foot units. During 1936 and 1937 the hemlock and pine-seedling
grids were discontinued, and in their place there was substituted the
“‘Daley’s Field’’ grid in a woods-edge community of sumacs, aspens,
young red maples, pine, and Kentucky coffee trees. The hemlock
quadrat was revived in 1938 and 1939 in place of this unit.

A chart summarizing the results obtained in 1933 and contrast-
ing the yields of 50-foot and 150-foot units constitutes Table VII.
It will be seen that the average per 50-foot solitary quadrat is higher
in all cases than the averages for the grids, for the two species that
appeared in the three-day totals in both the grids and the solitary
units. This also raises the estimate per acre and hence the estimate

Dec.

1939 B. P. BOLE, JR.— THE QUADRAT METHOD 37

for the population of the seventy-five acre area under consideration,
as indicated. Thus further evidence is provided that the estimates
based on small units are always larger than those based on larger
ones.

Remembering that the estimates per acre given in Table VII are
based on a three-day, and not on a total yield, the population per
acre of Blarina and Peromyscus on Little Mountain in 1933, as fur-
nished by the data in this table, may be compared to the figures
given by Williams, for 10-meter quadrats as recorded in Table II,
for those species at North Chagrin in the same year. It will be noted
that the divergence in figures per acre between North Chagrin and
Little Mountain is tremendous. At that time the writer felt that
the North Chagrin woodland naturally had a much higher number
of resident small mammals than did Little Mountain, even taking
into account the sources of error already mentioned. However, the
results of our 1935 and subsequent work with large grids indicate
no actual fundamental differences (see Tables V and IX) between
the two communities.

Table VIII gives the actual yield from each of the three grids
used in both 1933 and 1935 on Little Mountain. They help to give
an idea of the size of the series of specimens on which the total popu-
lation is based. This total (Table VID) is obtained from the average
figures for the three grids, here considered as three large quadrats.

Corroborating the North Chagrin records insofar as relative
abundances and increase are concerned is the fact that the popula-
tions of Blarina and Peromyscus underwent an increase of roughly
400 percent between 1933 and 1935. The appearance of many more
Species in the three-day totals for 1935 than for 1933 is another
indication of a general increase not confined to any one species. The
fact that chipmunks and flying squirrels were not taken in 1933 is
due, however, to our not having used rat-traps regularly in that
year.

The writer feels that the estimate for Pitymys is approximately
correct, but that the one for this vole obtained by the grid-census
at North Chagrin in 1935 is somewhat too large owing to the fact
that the grid apparently encompassed a colony. The 10-meter
quadrats (Table I) at North Chagrin tend to show that Piétymys is
less generally distributed than either Blarina or Peromyscus and that

Sci. Pub.

38 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

the population is not generally as dense as in the 150-foot quadrat,
further indication of the fact that more than one quadrat, of any
given size, should be used if population estimates for a larger area
are to be made. In this connection, attention is called to Table IX
(pp. 41-42), wherein the summarized estimates for all the 150-foot
quadrats on Little Mountain are recorded. It will be noticed that
in certain cases, notably those of Blarina in the pine ridge and beech-
cucumber grids, the yearly records from a single 150-foot quadrat
do not indicate the same cyclical tendency that the averages from
several quadrats do. Similarly, it will be seen that the estimates
based on the center quadrats are irregular and hopelessly out of
step with those based on the grids. The figures, to the writer’s mind,
entirely remove the central-quadrat method from the field of re-
liability except under certain conditions already enumerated. A
50-foot quadrat, even if sheltered from drift, can give no picture
with any degree of completeness of the small mammal population
of a large area, and if sheltered from drift from the outside has
nothing to prevent drift out of itself. The estimates tend to be
inflexible and either too high or too low as compared to the esti-
mates based on the large units. Thus, the smallest possible figure
obtainable from a single 50-foot unit is 17 per acre, on the basis of
one animal captured. The averages from a very large number of
centers give a better picture, but the difficulties accruing to this
method have already been pointed out. From the central quadrat of
the hemlock block it would appear that the Peromyscus population
diminished by half from 1933 to 1935, the exact reverse of the situa-
tion as recorded in all other quadrats, protected and unprotected.
Another conclusion that might be pointed out from the above and
that has already been indicated is that the population of the more
generally distributed species seems to be more or less constant in
northeastern Ohio, during a ‘“‘mouse year’’, or increase of the cycle,
from woodlot to woodland and from one climax forest type to
another (See Table IX, pp. 41-42).

Dec.
1939

in in 1933.

Tasre VII. Summary of the Quadrat-Trapping Work on Little Mounta

B. P. BOLE, JR.—THE QUADRAT METHOD

Population per acre based
on 3-day yields from
isolated units Cleft)

and from grids (right)

Population of area based
on center quadrats of
the three grids

Population of area based
on 3-day yield from grids
Ga ca)

Population of 75-acre
area, based on 3-day
yields of isolated 50-foot
units

Ave. number per 50-foot
quadrat of Hemlock
grid

Ave. number per 50-foot
quadrat of Pine-
Seedling grid

Ave. number per quadrat
of Pine-Ridge grid of
nine 50-foot quadrats

Ave. per quadrat of
the 8 isolated units
(based on 3-day yield)

3-Day and (Total)
Yields from 8 isolated
50-foot quadrats

Species

324 6.4

490

0.33

Blarina brevicauda...

444

o (1) 0.44 279
0.67

Sorex fumeus fumeus......

Peromyscus leucopus......

17.2 9.3

693 1333

1308

0.33

I.00O

8 (16)
13)
o (1)

Dake

163*

On 125

Napacozapus insignis.....

Tamias striatus. .

o (1)

Sciurus hudsonicus...

*Too high an estimate, as the Woodland Jumping Mouse is limited to wet ground.

**For wet, cool areas only.

Attention is called to the estimates based on the center-quadrats and the isolated 50-foot units as compared to those based on the

150-foot grids taken as units.

39

Sci. Pub.
Vol. V

THE CLEVELAND MUSEUM OF NATURAL HISTORY

40

“S¢6r ur Lr “€€6x ur skep $91 :syespenb q]e Joz O9vIJoay °S$€61 ut Lt
“£61 ur shep 11 “yooTWOP £$€61 ur ox “€E61 ur skep Zr ‘Sulppoog-ourg ‘$€61 ur gr “€€6r ur shep zz ‘aspry-surg :sown-Surddesy,

(Fx) 9 (0) 0 (S)z (0) 0 (Sz (0) 0 (hr) z (0) 0 “****-supro0a sXusoone[5
(£) 1 (0) © (1 (e)° (e)° (0) 0 (Ho (9) =" -snoruospny snsnqog
(LE) bx (0) © (8) + (0) 0 (9) % (0) 0 (£2) g (o)o “ooo s-sngeraas serurey,
(xo (oe) 0 (0) o (oe) (0) © (0) 0 (oe (oe) 0° “ues ss snnosnur sn
(rr) $ Ox (oe (e)° (3) ¢ (0) 0° (St (o)o "++ wngojaurd sduréarg
(z11) gb (£2) Sx (6h) 1% (11) 9 (27) o1 (ye (1b) Lr (6) 9 -*-sndoonay snosXuros9g
(F) 1 (©) 0 OL (0° (zo Ol (£)1 Or "**"Tyomorq sdoyvosvseg
(bz) £ (hx) 9 (br) + (or) + (Sz (z) % (S)x (z) 0 "+ -snaumny snoumy xo0g
(vor) 7 (4%) Z (of) ox (9) £ (of) $1 (8) z (bb) Z (fx) z "+ *-epnvoraaiq vurielg

$61 €€61 6£61 €€6x 66x €£61 66x €£61

s[v30], oor Woy Sur[ps2g-ould os pry-oulg sotadg

A SS TS See A A TT

SS _evveoes=oo0O©060601.}8———eeeeeeeeeeeeeeeeeeeeeeeeeoeeeeeeeeeeeeee eee

"S€6x pue ££61 ssvdX SY) JOy UTEIUNOW 21331] UO SIeIpeNd 100,J-OS1 95], JO SPIFIA PY “TTA aTav |

Dec.

1939 B. P. BOLE, JR.—-THE QUADRAT METHOD 4!

Taste IX. Populations-per-acre in different Forest Communities on Little Mountain,
Geauga and Lake Counties, Ohio. Figures Based on the 3-Day Yields of 150-Foot Grids
and their Central 50-Foot Quadrats.

&
= 2
> 3
aS] a = =o n
S = na = <3 nw * =
: ES as Bal Se Ne
s 4 es = a fo) 2) > iS Se S ro)
a S > a = = 2 =| 9 = = Q
s Ei 2 q Bea oe eS Ea inate ey glial geiko
of ra] Ce Oue PF a co f. eeee Gi) eas tae
Ping Ripce
Grid
1933 4 iN Iz
1935 14 p: 2 16 4 34
1936 18 4 4 2 4
1937 20 16 : ab at ae Io ae 5 fe
1938 22 Iz i 2 4 ae 10 ah 4
1939 26 18 4 a
Center
1933 ms mi He by ae ye ye
1935 17 ae Ae us] Ae Ae 17
1936 is a ne i zt. xt a8
1937 17 Be: Ns me phe eas 17
1938 ye x7, ae sis ie ahs as
1939 } me
Ping SEEDLING
Grid
1933 4 4 os a ae ne 6 ae Es
1935 28 4 es 4 4 nh 20 f. 8
Center
1933 .-
1935 17
Grid
1933 6 8 ss 23 ae ¥e 12
1935 20 8 ae 8 4 2 42
1938 8 12 4 2 16
1939 12 18 2 4 2
Center
1933 oF x: re - oe ae 34
1935 an ae Si - ¥ »: 17
1938 17 a a
1939 17 17
Datezy’s Freip
Grid
1935 4 2 ~ 8 2 2 il) Sid + 6 4
1936 12 wi 24 pk) Be ey a8 7 4 Pe ere ec
1937 34 of 12 2 4 bie 18 ne Si ae eee z:
Center
1936 17

1937 i ae io Ss ae os 51

Sci. Pub.

42 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V
m
Pe -
S)
2 5
La © | ~” = = ”
q 3 <) 2) 3
e 5) a. a S| 3 > Q.
H § ie) q ie} 4 72) | BS
< =} —_ 7) xs ” te) N
5 a a MS See R Mea ae slaps ele: 3
< = oO cs g 3 2) 3) = x as
= = 5 = Os = im = =) Ss = an
Oo ma) “A A. = O 7) av = a a aie

Brzcu-CucuMBER

Grid

1935 38 8 2 6 4 4 39: 4
1936 32 16 6 4 4
1937 24 24 4 2 16
1938 30 58 8 4 4 22!
1939 22.) 32. 4 oe 2
Center
1935 ck n 17 be if ue 17
1936 : me Bs ape ah be e
1937 a 17 ti fe ue be 34
1938 ap 85 a sh As ve 17,
1939 “i : ae ss e
AVERAGES, ALL QuADRATS
Grid
1933 409 | bao.” SP od OP Pp P ; Pe?
93.4)" (Goro)! (4-5) (G29) 1) es) Gs) G5) )-2 eae
1935 24:8 4.8 ‘2.6 ‘Si4) 3-6 36° \30.4) O58") ANS)" pons e
TO36 | p52) | 6.3) 4e7 oP P P elee Batlle T3. ee “id
TO370 | M25 Oy p12 7 ALF) Wey 12) Pe a7 ee P 0.6
HOLS, | 2oro W273" 27) wecG Wt ge sng) CEOL Os mers Foy hee Poe
TORO: W/ZPOLO! 29-7 WBN7i ean eae P BON Ws Je Of7
Center
1933 S260) 2556 17.0
TORS Tr SS ee D27
EQZ00 | 1720) 0/2: 5-6
O37 deo) Se 34-0
1938 561340 5-6

*Figures for 1934 are hypothetical when parenthesized, and are based on yields
at other stations. Otherwise the 1934 figures are based on spot censi.
**Peak Population not duplicated at other stations than Little Mountain.
Elsewhere, 1937 was peak year for Sorex fumeus.

***Peak much lower than elsewhere, due to failure of key food crop in 1937
(acorns). At other forest stations, Beech and Sugar Maple predominate; they
are rare on Little Mountain.

****Does not include figures for Daley’s Field, which was not a grid in this year.

P Present, but in populations too low to be measured by only three or four
quadrats.
Key to Forest Communities: Pine Ridge, Pénus-Acer Facies of Pinus-Tsuga
Association; Pine Seedling, Quercus-Castanea Association with understory of
young white pines; Daley’s Field, a Woods-Edge Community (Acer-Populus
Assoctes); Beech-Cucumber, Fagus-Magnolia Facies of Acer-Fagus Association.
See p. 36.

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 43

Taste X. Check-list of the Mammals Mentioned in This Report. Names for the most
part are taken from ‘‘List of the Mammals of Ohio,’’ Ohio State Mus. Mimeographed
Bull. No. 1, by E. S. Thomas and B. P. Bole, Jr., February, 1938.

Common or Virginia Opossum. . Didelphis virginiana virginiana Kert

Sniglay shrew. ...2-..--.2. 06+ Sorex fumeus fumeus Miller
Canada or Masked Shrew...... Sorex cinereus cinereus Kerr
Thompson Pigmy Shrew........ Microsorex hoyé thompsoné (Baird)

Common Short-tailed Shrew... .Blarina brevicauda brevicauda (Say)

Old-field Shrew or Little Short-

PBUCAUSHECW sit o:- ahalaiotse doi 2 Cryptotis parva (Say)
rae C1110) ean Scalopus aquaticus machrinus (Rafinesque)
Mainy-tailed Mole............. Parascalops brewert (Bachman)
Stat-niosed Mole. .............. Condylura cristata (Linnaeus)

Eastern or Common Raccoon... . Procyon lotor lotor (Linnaeus)

Dagedicts)| a0) ne Mephitis mephitis nigra (Peale and Beauvois)
Allegheny Least Weasel........ Mustela rixosa allegheniensis (Rhoads)
Mew York Weasel. 2 .......4.5: Mustela frenata noveboracensis (Emmons)
Pastesmihed Fox... f..2 606.0685. Vulpes fulva fulva (Desmarest)

(This species and the next are not distinguished in the text)
Basccen Gray Pox............-. Urocyon cinereoargenteus cinereogenteus (Schreber)

Middle-Eastern Chipmunk...... Tamias striatus fishert Howell
(In a few cases the form /ysteri is also included under Tamias)
Middle-Eastern Red Squirrel... .Sciurus hudsonicus loquax Bangs
(Unless otherwise specified, all records of Scéurus refer to this form)
Northern Gray Squirrel......... Sciurus carolinensis leucotis (Gapper)
Middle-Western Fox Squirrel... .Scéurus niger rufiventer Geoffroy
Middle-Eastern Flying Squirrel. .Glaucomys volans volans (Linnaeus)

Prairie White-footed Mouse..... Peromyscus maniculatus bairdi (Hoy and Kennicott)
Northern Deer Mouse.......... Peromyscus leucopus noveboracensis (Fischer)
Stone’s Lemming or Lemming-
VG LO: 2 5 ee ae Synaptomys cooperi stonet Rhoads
Eastern Meadow-Mouse or Field-
VOUS ce SR ee eee Microtus pennsylvanicus pennsylvanicus (Ord)
Mole Pine-Mouse or Pine-Vole. . Pitymys pinetorum scalopsoides (Audubon and
Bachman)
Common House Mouse......... Mus musculus musculus Linnaeus
Norway or Barn Rat........... Rattus norvegicus (Erxleben)

Hudson Bay Jumping Mouse... .Zapus hudsonius hudsonius Zimmerman)
Woodland Jumping Mouse...... Napaeozapus insignis insignis (Miller)
Eastern Cottontail Rabbit...... Sylvilagus floridanus mearnsi Allen

Sci. Pub.

44 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

EXPERIMENTS WITH QUADRATS INSULATED
BY SHEET-IRON

The method of enclosing quadrats with galvanized iron sheeting
to exclude drift was first attempted with the aid of Dr. S. C.
Kendeigh at the Baldwin Bird Research Laboratory in August, 1934.
A 50-foot quadrat was selected, and a trench dug completely around
its borders to a depth of from six to ten inches. Galvanized iron
sheeting three feet in height was set in these trenches, which were
purposely dug below the level of mole and pine-mouse runways. The
sheeting was supported by being nailed to small posts that did
not reach the top of the sheeting. The sheeting was nailed to in-
ward side of the posts so that any mammal desiring to leave the
quadrat could not jump up to the tops of them or climb them. Pre-
viously Dr. Kendeigh had determined the height to which deer mice
(Peromyscus leucopus) could jump. This he accomplished by placing
live mice in garbage pails whose sides were one foot high and two
feet high respectively. Two placed in a one-foot container escaped
at once. Of ten animals, both adults and young, placed in a two-foot
container, only one succeeded in escaping and apparently did so by
running centrifugally around the rusty walls of the can until the top
was reached. The highest jump made by excited mice did not exceed
eighteen inches. Thus it was believed that the additional nine or
ten inches afforded by the iron sheeting constituted defense against
drift.

After the sheeting was installed, the trenches were refilled and
the ground about the sheeting was tamped down hard, to make
excavation as difficult as possible. Finally collars of tin sheeting
were placed about every tree inside the quadrat and all those
immediately outside, to prevent species of arboreal tendencies, such
as Tamias and Peromyscus, from entering or leaving by this route.
Traps were then set throughout the quadrat and about the bases of
each post supporting the sheet-iron, while a few were placed about
the bases of nearby trees unprotected by sheeting.

The result of all this elaborate preparation was exactly zero in
1934. It seemed quite possible to the writer that no small mammals
would stay in a 50-foot area while posts were being pounded and
sheet-iron rattled for the better part of a day. However, the experi-

Dec.

1939 B. P. BOLE, JR.— THE QUADRAT METHOD 4S

ment seemed well worth trying, and it must be said that there was
virtually no recent small mammal sign whatever in the quadrat.

The attempt was resumed in a different place next year, complete
in every detail. It should be remembered that the purpose of the
sheet-iron was to exclude drift and that therefore, if the sheeting
shut off all runways, the total yield should give the exact popula-
tion of the 50-foot unit. The traps placed outside the quadrat along
the bases of the posts and trees might be expected to catch the drift
animals that would otherwise have been taken in the quadrat in the
absence of the sheeting. Some idea of the proportion of drift to
inhabitants should thus be obtained, realizing that any animal hap-
pening to be in the quadrats at the time they are set up is considered
“‘resident’’.

From Table XI, it will be seen that just one Blarina was taken
inside the sheeting, on the third day of trapping. During the five
days that the traps were kept working, four Blarina, four Pero-
myscus, and one Pitymys approached the quadrat and were captured
outside the sheeting and in its immediate vicinity, all these exterior
traps being less than ten feet from it. Of these nine animals, six were
taken in the first three days and two on the last two. It is possible
that some of these were animals dispossessed during the installation
of the metal.

One Blarina per 50-foot quadrat indicates a density of 16 or 17
individuals per acre, rather lower than the average of figures ob-
tained from other sources. The fact that no other species whatever
were taken in the quadrat seemed to suggest again that the animals
might have left due to the disturbance attendant upon the installa-
tion of the sheeting. It was therefore decided to leave the quadrat as
it was except for removing four eight-foot sections of iron, one on
each side, to permit animals to go in and out. It was decided to
allow the quadrat to become ‘‘naturalized’’ in this way for a few
months, and then to return and quickly install the few remaining
sections which would entail a minimum of disturbance.

Accordingly, in November, 1935, Mr. Moulthrop returned to
the quadrat and installed the four missing sections. There had been
a fall of leaves since the summer work, and conditions in general
were mote favorable for small mammals. The trapping was compli-
cated by several falls of snow, which covered the traps to a depth of

Sci. Pub.

46 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

six inches on one occasion. The traps were run from November 27th
to December 5th, inclusive. During this time no less than 13 Blarina
were taken inside the quadrat; 2 Parascalops and 2 Pitymys also were
captured before all signs of activity ceased. In the peripheral traps
on the outside of the sheeting, the ‘‘drift’’ consisted of 8 Peromyscus
in the same length of time. The astonishing yield of shrews came
almost entirely in the first five days, the trapping on the second,
third, and fourth days having been almost entirely stopped by snow
and hard rains. Ten of the Blarina and 2 Pitymys were taken within
three days of actual trapping. Both moles eluded capture until the
seventh and ninth days of trapping, and Mr. Moulthrop believes
that a third animal was left uncaught.

These results are not surprising, in consideration of the in-
ability of 50-foot quadrats to give a fair sample of the population
of a large tract of woodland or field. The Blarina yield would give
a population of 217 animals to the acre, a concentration highly
unlikely in view of the summer’s trapping records. Furthermore, no
Blarina were taken outside the quadrat in November. This seemed to
mean that the number of shrews within the quadrat was for some
reason unusually high, perhaps because of the insect concentration,
which was very conspicuous as compared to the ground outside
the quadrat.

It may be suspected that the sheeting insulation “‘leaked,’’ and
that the Blarina were really drifting in under it. All that can be
said in answer to this is that not a sign of tunneling under the
sheeting was discovered either in 1934 or in 1935. By 1936, however,
there was definite evidence that considerable tunneling had taken
place under the sheeting, and it was reset. In addition to the animals
captured inside the quadrat (see Table XI), several Peromyscus were
captured outside the iron sheeting, as was the case in 1935. In 1937
further evidence of leakage was found and the tunnels were plugged.
The results obtained in these two years in general seem to corrobo-
rate a deduction that was obvious as early as the fall of 1935 — that
the metal quadrats acted as a trap of sorts for insectivores, especially
for Blarina. It is difficult to see just why a shrew should be less able
to find its way out of one of these quadrats (open in four small
sections) than into it, but such seemed to be the case, to judge from
the records. The explanation lies in the fact that moles plough up

Dec:

47

B. P. BOLE, JR.—THE QUADRAT METHOD

1939

“[eurrue UO jo SSUTYIOM JO p4O591 941 UO PSP a»

‘surddny “y “Y pur ystopusy ‘9 *¢ ‘siq Aq ouop Surddesy Z£6r,.

O= 000 (0) aj (6) oO =i 20 (oye [es 0) (o} fo), fo) Oe, (0 Oar as see snowing x10
oy eee) prclegatie © Ah ede i ey) (oho) [Ss 6 OS 0) SO Gay ae aa, sdojvosvseg
(Oy (eh Yo) On ta 170 (o] Sie) Te) (Ok (0) (0) Vir te fe) Be IO) oe ae er ae a skurArtg

Ver el -T De geet: (e) = 0) 19) VO te Oo %GF 6 ip Rs ir ee el ace a a snosAuosog

OZ Tae Ommeat ZI 1s 0 aH ts) Vien Can =e Lve 20) (OF LO I south LS aOR Nn vulivig

3) 3) VB) oO oO CB)

ol te) 8] Oat. oe or orc) ad Gece oe a = Spe tae BSI

abe. & a. ites © <4 8 <0 8 4 8 =| a oo

Tp alta) ee ee Dire 2 ce te ese 2
Cyemes by ie Sn Ss Coe FS ed (pea a On a bs By. Sees 0
oS ee aa ae et as 8 a. os Sai ane 5
% 6 A 2 6 A Be te A 5 6 A Bo 2 8 6 A 2
mM BH ral oy eo al oy HR al oo eal oo HH ew oOo BH eal o)
poinsuy jouu05 poiv[nsuy jo31u07) poqvpnsuy josu07
ee ee ee Se a ee eee

4Lt61 9t61 St61

SS. oe
‘SIIU(] P2I¥[NsU] IY UT Spat X [v3], uO pur
s[ouUOD O42 UT spjarX Avq-£ UO pase ‘oI0v-Jod-soqvuTasg *(2Z1g 100,J-0S) paseduro5 SIVIPENC prIv[Nsuy pur [OIIUOD JO SP[21X “TX A1avy,

Sci. Pub.

48 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. ig

the ground inside the metal units to an extent unapproached outside
the quadrat, form networks of runways three or four layers deep,
and thereby greatly increase the foraging terrain not only for Blarina,
but for soil invertebrates as well.

I believe that what was said about center-quadrats will also
apply to the metal quadrats — namely, that estimates for a large
area should certainly be based on several of them and not on just
one. In this case the expense of the metal, the manual labor involved,
and the length of time it takes for the quadrat to become repopu-
lated after the initial installation of the sheeting, renders the method
impracticable. If, as seems to be the case, the insulated quadrats
trap certain species and thereby create unusually favorable soil
conditions for still others, the insulated quadrat method would
appear to be the most inaccurate source of information on popula-
tions, giving a faulty picture of not only the actual, but also the
relative abundance of the species of any given small mammal fauna.

THE 100-FOOT QUADRAT, 0.229 ACRE

By the end of the summer of 1938 a total of seventy-eight 150-
foot quadrats had been operated since the beginning of the work. In
the course of all this a fundamental practical weakness in this
size of quadrat was made clearly evident: the 150-foot quadrat
must be set by at least two men. One man cannot properly set one
of these quadrats in the eastern United States in less than a full day's
time; and a serious error appears in the data if he takes this long. If
he sets half of a quadrat in the morning and the other half in the
afternoon, he will have caught several animals, under normal cir-
cumstances, before the quadrat is completely set. If, as is customary,
the first day’s yield is considered to be that of the first night follow-
ing trap-setting, he is “‘padding’’ his results with the yield of part
of an extra day. If he counts his trap-setting day as his first, he is
obviously getting little service from the part of the quadrat that
he last set. In short, all of his quadrat does not work for a com-
parable length of time, which is not to be desired. It is impractical
to try to end the trapping of the two halves of the quadrat at dif-
ferent times, 1. e. as they were set. Consequently, no less than two
men should be used to set a 150-foot unit. They can finish it in half
a day. They also tend to keep the movement of the small mammals

Dec.

1939 B. P. BOLE, JR.—_THE QUADRAT METHOD 49

at a minimum during trap-setting. There is usually very little
activity around the trapper for a radius of 50 feet during trap-setting,
and two workers virtually cover a 150-foot quadrat in this way.
But even when two or three men are used, some captures will often
be made during the trap-setting period.

In view of the fact that two workers are not always available,
there is a definite need for a ‘‘one-man’’ quadrat. The 100-foot
quadrat largely supplies this need. During 1938 the writer instructed
one of the Museum’s field-workers, Mr. Winston Jesseman, in the
use of these units. Mr. Jesseman operated six of them in 1938, and
two in 1939. The writer operated one in 1936 and two in 1938, all
three in New England.

As might be expected, the population estimates based on the
yields from these quadrats (which have to be multiplied by a factor
of 4.35 in order to give figures per acre) run noticeably higher than
the 150-foot estimates, and much lower than the estimates based on
yields from smaller quadrats. These facts, as expressed by the line
representing the average daily yield, may be observed in Fig. 2.
The yields after the third day in this series of quadrats are undoubt-
edly lower than they would have been if a series of length compa-
table to that of the 150-foot units had been used.

During 1938 most of the open-field 100-foot quadrats yielded
population figures for Microtus of over 50 per acre, which is con-
siderably higher than the average for the year for the same species
as deduced from a large series of 150-foot units. Only three times
since 1933 have the larger units yielded figures higher than 50 per
acre for this species. Similarly, Peromyscus leucopus populations
appeared to be about 4o per acre in 1938 on the basis of three-day
1oo-foot yields as opposed to the 150-foot figures, in the same
ecological environment (climax forests), of 22.2.

While the error would thus appear to be considerable, it is
far smaller than that incorporated in the data of smaller quadrats,
and under certain circumstances this may be compensated for by
the greater ease in operation that this quadrat affords in com-
parison to the 150-foot unit. One man cam operate it, but wher-
ever it is possible, two or three men should operate a 150-foot
quadrat instead. The relative efficiency of the 100-foot quadrat
may be visualized in Fig. 1.

Sci. Pub.

50 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

CIRCULAR UNITS

Dr. S. C. Kendeigh and his students at the University of Illinois
have used half-acre units of circular outline, corresponding to the
150-foot units used by the writer. A circular area with a radius of
83.2 feet that he trapped in an Illinois oak-hickory woods in
November 1938, gave populations for Peromyscus leucopus of 16 pet
acre (40 per hectare), 6 per acre for Pitymys, 2 for Blarina, and 4 for
Glaucomys. These figures are far lower than the averages for the
same species in the same year in Ohio, but well in line with some
other Illinois figures at hand. Unfortunately, the writer has no
data that is strictly comparable to this, but a 100-foot quadrat
operated by Mr. Jesseman in a swamp forest community at Bloom-
ington, Illinois, in the same year, gave figures of 17 per acre for
Peromyscus and 26 per acre for Blarina. These are well below the
Ohio figures for the same habitat as furnished by 1oo-foot quadrats.

Dr. Kendeigh prefers the half-acre circular unit to the 150-foot
quadrat because its perimeter, and hence its edge-exposure to drift,
is 67 feet less and its area exactly a half-acre. The fact that this
museum’s youthful assistants invariably prefer to trap in the corner
50-foot quadrats of each grid, and that the corner quadrats show -
consistently higher yields than the others after the second day of
trapping, indicates that Dr. Kendeigh’s preference is very well
founded. However, the difference in the yield-behavior of the various
parts of a 150-foot quadrat only becomes evident after the initial
three-day period is nearly over. If the three-day limit is adopted,
there is no apparent difference in the yields of the two types of
trapping areas in the comparative data so far accumulated by the
writer (in 1939, three circular units operated, two at Aurora Pond,
Ohio, one at the Holden Arboretum). However, there is no doubt
that the half-acre circular unit is preferable, where it can be practi-
cally operated.

In open fields the laying out of a circular unit is an easy matter
— far easier than laying out a grid or an unsubdivided 150-foot
quadrat. In thorn scrub it may take ten times as long as to run the
grid. The writer believes that the circular unit should be used
wherever a worker can describe a perfect circle about a central
point and at the same time lay a trail of string behind him. Where

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 5!

this cannot be done, the circular unit can still be established by
running at least four cross-lines outwards from the center, but the
inaccuracy of the circle thus described vitiates wholly or in part the
advantages of the circle over the 150-foot square unit. As the latter
is almost always easier to run in forested country or in heavy brush,
it should be used in most such cases.

GENERAL CONSIDERATIONS

An attempt to apply Raunkaier’s Law of Frequence to the small
mammal populations of northeastern Ohio upland forests appears in
Table XII. The average percentage of frequence given for each
species is that over a five-year period beginning in 1935. In compil-
ing the data it was noted that the cyclical species usually changed
their group classification from year to year, and that the Raunkaier
formula failed to hold for each season. It will be seen, however,
that the averaged figures closely approximate the Raunkaier for-
mula, and seem to indicate that the law holds good among mammals
as well as plants and invertebrates. Close inspection of the data
reveals that the one species falling in Group D (60-80%) is Blarina
brevicauda and that a gain of but 2/10 of a point would cause its
classification in Group E. This would make the figures even more
similar to the Raunkaier formula. Similarly, a gain of one point
would take Tamias from Group A into Group B. It will also be
seen that at least four of the species in Group A (Synaptomys,
Microtus, Mus, and Zapus) are not strictly animals of the upland
forest and are not represented by any more than one or two speci-
mens per season; if these are deleted, the forest frequences conform
still more strictly.

Fifty-foot quadrats (in all cases components of grids) are used
here instead of the larger units, as the problem at hand is the
relative rather than the actual abundance, and small quadrats seem
to be as useful as the large. Actually, a 150-foot unit almost invari-
ably is large enough to include most if not all of the small mammal
species of an Ohio woodland. Thus, the smaller quadrat seems
actually preferable to the larger for the purpose of testing the
Raunkaier formula.

It is of interest to note that in Table XII we see a recapitulation

Sci. Pub.

52 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

of the periodic population maxima that we found in the other
tables. In this case, however, we are dealing with the distribution
of the species rather than population density. In general, it would
seem that widespread dispersal goes hand in hand with high density
of population. That the former may sometimes occur before the
attainment of the latter is indicated by the figures for Peromyscus
leucopus. Thus, the frequence of this species was almost as high in
1937 as in the ensuing peak year 1938; but if we turn to Table XIV,
we find that the 1937 populations were not nearly so great as those
of 1938.

It is often asked whether the use of quadrats in counting motile
organisms is justifiable. The method was originally devised for
counting stationary living things like plants. The only way in
which an animal such as a small mammal may be considered sta-
tionary is to think of it in terms of its ““home range’’ in the sense
suggested by Seton (1909). Insofar as its home range is concerned
small mammals may be considered as fixed organisms, at least
during the breeding period, if not during most of the year. If a
quadrat is used that is large enough to include several home ranges,
we ate obviously removing “‘units’’, in a sense fixed units, in the
course of trapping, and we can get a close approximation to the
actual number of individuals in the area. It is equally obvious that
if a small enough quadrat is used we are dealing with a part of one
or more-home ranges. Eventually the animal, whose home-range
includes the small quadrat, is captured, and forms part of the basis
for estimating the population of a larger area than the quadrat.
This population is based, however, on the yield for the small
quadrat, and five animals for each 100 square meters gives a very
much larger population figure per acre or hectare than one based on
five animals per area ten times as great, although the larger area
may more nearly approximate the actual home-ranges of the five
animals than the smaller area. The accompanying graph, Fig. 1,
shows quadrats of different sizes yielding the same number of ani-
mals apiece, and how the estimate per large area, based on this
yield, increases as the size of the quadrat decreases. The time
element is here considered constant for all sizes of quadrats.

From the above, it is evident that the smaller the quadrat, the
larger is its yield in proportion to its area. Figures obtained in the

Dec.

B. P. BOLE, JR.—THE QUADRAT METHOD 53

ue ie)

I LC A EL enero

I fo) I re § 7 MET $ JoeyoneY 03 Zurpsoody ‘uorssodsiq souenbezq jo onvy
fe) I I z €1 Se aA aera at ae ‘*dnoaZ sad ‘ssouanbasy 3s2I10,5 xewyyD Jo soquinyy
0/,001-08 0, 08-09 %09-OF %ob-O% %OT-O CC rr er) eee eee eee eee oe oe sdnosy) 99 ¥IUI9I9g
a es ee
a ‘al 5 q Vv souanbary

SuOle 1G) Ole yt 9 On GO. Siar 07S Som gig) tor ey On 1 Oe ye | 6s got g 62 --ssvax [Ty ‘o8eI9Ay

(Seep 1604 (syh(o) Pg eet PO ee OC : GiO se 2. 6) Te ah Te kb Suey gpome = fees ein 6661
or ee eee. cree br ber L'o o'hr 9'oL eeee L‘o 1'6 LL + I 9°S L'o 9° gz 8°1g Sis Wells iat epereiahetelicienens gt6r
eevee O’II 77 CABG d eevee eoree Cds d b'b9 cree cree JL) L'9 £9 8° Qt 1 a €-¢€¢ Z'7%® 0080-0 2-0-0100 0 O00 OG LE6r
g'I L:¢ ores eevee gii ores L°¢ 9°91 ove l'Il L:¢€ g Lz 9 ¢ 1° S€ QI o' bz Hho fble CONC ECE CIC SC CIC Ci ir Cir 9t61
eee. eee esos gir econ. eree bre L:06 gr TTL bre bob cee I‘II cee 6°Sz 9°76 UUOU 0-00 0-C-ORO OO C61
Paeia fe Ss St SS Ss Se SSeS
3
g : = 2 ie:
he 7) : OD wa

g ae ee he ae Sie ee oe =e

ao) oe ey BP RB 4 v O Qo ae Asien! nu

SY oH ” ro) a, f= g 3 Fa e) 8 Pp 2 rv, ba a

Bee ee ne ed a :

ue eee Nee oe em Oe OO eo ee D4

"Iva YO" Ur (Splsy jo syusuoduo7)
S¥IPEND 3100,4-0$ paspuny suo Ajayeurtxosddy jo sps0d2y wo poarsop ssouanbosy 2a1F sonB1q “MET S JoleyunEY 03 Jurpsiooe sdnossy
OIur UOTINqrasiq] s19y3 pue “661-SE6r “vorGoy puelras[D 243 Ur speumepy [[eWIS Jo DUaNbas7 Jo soFeIWEDI0g 38930, XBW “[[X AAV],

ESTIMATE—PER ACRE

QUADRAT SIZE

Fic. 1. Estimate-per-acre. Plotted Against Quadrat Size, Yield Constant
(a animal per quadrat).

field bear out this theorem, particularly when estimates are based
on total yield rather than on a limited yield after two or three days.
Figure 1 indicates that there is an enormous error in estimates based
on the yield from single 50-foot quadrats as compared to the “‘large”’
quadrats of 150 feet. For instance, a yield of five animals for a 50-
foot quadrat gives a population of 87 per acre; for a 150-foot unit,
between 9 and 10; while a yield of five per 10-meter unit gives over
200 per acre! Actually, however, it is not meant to imply that othér
things being equal a small quadrat will give the same yield as a
large area. Over a long period of trapping, the yield of a small
quadrat is to the yield of a large one more nearly as the sum of the
lengths of the sides of the small one are to the sides of the large,
rather than as the area of the small is to the area of the large (Table
XIII). The longer the traps are kept in operation the truer this
becomes, as indrifting animals which enter by the four sides of the
quadrat, become a larger and larger percentage of the total. Drift
is never entirely eliminated in an open quadrat unprotected by a
shelter-belt of some sort, but in proportion to the ‘‘resident’’ popu-

ee:

193 9 B. P. BOLE, JR._—THE QUADRAT METHOD SS

lation, it becomes larger and larger as the area of the quadrat is
made smaller, since the length of the perimeter of the quadrat
decreases more slowly than the area enclosed. In using ‘‘total
yields’’ with small quadrats, we are therefore confronted with
two errors, the one arising from drift which cannot be legitimately
considered part of the resident population, and the other from the
fact that this error is greatly enlarged when it is a component of
the figure per quadrat area used in estimating the population of a
much larger unit.

The full import of drift in a small quadrat, even when the yield
for a single night is considered, may be made obvious by the con-
cept of a stump in the central quadrat of a grid of nine 50-foot units.
Let this be theoretically the home of five deer mice (Peromyscus

Tasig XIII. Drift Rates of 1o-Meter and 150-Foot Quadrats Compared to Each Other
and to Certain Basic Quadrat Relationships.

A. Sum of the Length of the Sides of a 10-Meter Quadrat................ 131.4 Ft.
B. Sum of the Lengths of the Sides of a 150-Foot Quadrat............... 600.0 Ft.
DOPE TEMTDENALOUDEN een ce ye ae R eines ce tine a elses overs « cccuneisenee 124.565
D. Ratio of the Area of a 10o-Meter Quadrat (1076.3 Sq. Ft.) to that of a

MAO ROUM Ouddrat (29), oo Sq.Ft. )in 4 fsc5e soe Meds e nie eete hestee: 1:20.905
E. Average Daily Yield per Quadrat from four ro-Meter Quadrats from

theyath to the oth Day of trapping injt935 -..... 22. 2---see2ce ete ee 1.2
F. Average Daily Yield per Quadrat from six 150-Foot Quadrats from the

Atmicontne oth Day of Trapping in! 1935. 2 < a. we. ess ono as yee es 5-8
Rr MED EI SOTCBE SCONE Cec Na Ses ymin Vict edan sao-giueend ty t-qrgya ee dude sistels sarees Se 1:4.83
H. Average Daily Yield per Quadrat from four 10-Meter Quadrats from the

Rumeco the 7th Day of Trapping iM 1936.5... oes icin iepeitinced oe 2 Be 0.25
I. Average Daily Yield per Quadrat from twenty-three 150-Foot Quadrats

from the 4th to the 7th Day of Trapping in 1936................... 2.19
MPMRETO ICH ELINLOME Stash: Lis teu et Sah ON als lala'a's co Ver cict dich vin medura suctes 1:8.76

K. Ratio if the Yields During the First Six Days instead of the Last Six are
used in Computing the Daily Average per Quadrat of the 10-Meter
BemCSrpTOsGh (Mere gets .ctae-o cee ies < kiSE Cw) Ac esta A Sots «Stale ete 1:404
L. Ratio if the Yields During the First Four Days are used instead of Those
During the last Four in Computing the Daily Average per Quadrat of
PR CREO-Metet CHES; TG 3G. a) epee this laden surement ns Wloehaehe 6 are Re 1:4.38

Sci. Pub.

56 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

leucopus noveboracensis). Let us consider that traps are set in only
one 50-foot corner quadrat, and that the large quadrat constitutes
the major part of the home range of the colony.* Eventually all five
will wander into the corner quadrat and will be captured. This may
happen in the first night and experience has shown that this is
usually the case with this species in favorable weather. If, on the
other hand, traps are spread over all of the nine quadrats, the
chances are greatly in favor of the five being caught elsewhere in
the large quadrat, perhaps one animal in each of five of the nine.
Whatever the arrangement, we now have five animals caught in
approximately half an acre, while in the case of the corner quadrat
we have five animals caught in only 1/17 of an acre. If there are
two flying squirrels (Glaucomys volans volans) in the district, they
are as apt to be caught in the 50-foot unit, if it alone is trapped, as
in the larger one if traps are set over the whole area. But the interpre-
tation of the yield would be quite different.

Chitty (1937, pp. 50-52) emphasizes the fact that not all small
mammal species are equally wide-ranging or sedentary in their
habits, and that a trapline or quadrat may draw some species from a
much larger area than it does others in the same length of time. It
is all the more important, therefore, to use large quadrats from
which the resident population can be removed at once through the
medium of a large number of traps. Unless the traps are numerous
enough to cut all channels of movement within the area trapped, the
resident population may not be removed before the yields become
seriously enlarged through drift, and this drift may consist largely
of active, widely-ranging species. The impression is then made that
these species are relatively more numerous in the habitat than is
actually the case. Obviously, population estimates based on total
yields are subject to an error from this source in addition to the
others already mentioned.

The determination of the exact time after which the yield from a
quadrat can be said to be of drift only is still a matter of some con-
jecture, but evidence is presented on the following pages in support
of an arbitrary three-day limit. Until the entire small mammal

* Murie and Murie (1931, pp. 207-8) indicate that the home-range of Peromyscus
maniculatus artemisiae, a related form, is in the vicinity of 100-yards in diameter.

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 57

fauna of a given area is rounded up with live-traps, marked, set
free, and then recaptured and the details of their movements care-
fully noted, the problem is not likely to be satisfactorily solved. A
host of American, English and Russian workers are at present
engaged in research that has already added much to our knowledge
of the problem. Certain species, however, are difficult to capture in
live-traps, notably, among the species considered in this report, the
insectivores and the lemming Synaptomys cooperi stonei.

It is the consensus of opinion among most mammalogists that
for most cases the great majority of mammals on a trapline are
removed in three nights. Townsend (1935), Hamilton (1937), and
Dice (1938), used three-day limits. Elton, et al. (1935, based the
estimates for their quadrats on a four-day yield. The writer feels
that if large quadrats and a sufficiently large number of traps are
used, the four-day limit is much too long.

The accompanying figures (Nos. 2 and 3) are indicative of the
fact that there is usually a sharp decline in the daily total of cap-
tures until the third day of trapping, after which the rate of decline
is much more gradual. It will also be seen that the general level of
the line for any given year as compared to those of other years is
an indication of the ‘‘population pressure’ during that year, as
evidenced by the rate of drift. A similarity in the rate of decline
from yeat to year may also be noted, although the rate for 1933
(Fig. 3) is irregular and resembles that of the smaller quadrats
(Fig. 2). Had a larger number of quadrats been used there is every
likelihood that the irregularities would have disappeared. The
peak occurrence on the fourth day of the 1938 trapping is an irregu-
larity that must for the moment remain unexplained, as a large
number of the 1938 quadrats contributed to the peak, and it cannot
be explained on the basis of too few quadrats having been used or
on any single atmospheric condition. In the average line for all
five years such peaks vanish. The similarity of decline, emphasized
in the average line, indicates a more or less constantly decreasing
rate of drift, although the zero rate is probably never attained if a
sufficient number of units are used. We may call this rate of decline
after the third day ‘“‘drift curve’’ and we note at once that in the
absence of a well marked initial peak in the case of the 10-meter
units (Fig. 2), the entire series of early yields seem to resemble a

NUMBER OF MAMMALS CAPTURED (YIELD)

THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

DAY

Fic. 2. Average Daily Yields of Small Mammals from Four Different Sizes of
Quadrats Operated from 1932-39, Inclusive. At each point are figures giving
the number of quadrats in the Average.

NUMBER OF MAMMALS CAPTURED (YIELD)

faa
Neen Ee ee
1 frill i ed Gal ol et
tl al as iil ad a i i
«LIT el fil i fa a CE a a 8
Lil SS cl a a a in SH
rat WS AE A
jo tee MI ae

6 7 89 10 i 12 13 if 15 ie 17 18 19 20

DAY

Fic. 3. Average Daily Yields from 150-Foot Quadrats in Six Different Years,
with the Average Daily Yield per Quadrat for All Years. At each Point are
numbers giving the number of quadrats in the Average at that point.

Sci. Pub.

60 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

drift curve. On theoretical grounds, it is legitimate to assume that a
ro-meter quadrat’s yield would consist largely of drift owing to its
small area, and the nature of the drift curves seems to support this.
Further support is provided by the last item of Table XIII, where a
substitution of data obtained on the first six days of trapping with
1o-meter quadrats, for the data obtained from the fourth to ninth
days Clast six days) actually makes the ratio of the small and large
quadrat yields more nearly as 1:4.5. This ratio, mentioned in Table
XIII, can obviously hold only if there is some drift, i. e., a regular
rate of influx. If populations are so low in any given locality that
there is virtually no drift, as was the case at North Chagrin in ebb-
cycle years (1933, 1936, 1939), the ratio breaks down. The weakness
of the data given in Table XIII lies in the fact that all of the 10-
meter quadrats were operated at but one place, whereas the 150-foot
series covered a large number of localities and hence more nearly
represent an average set of data. Had more 10-meter units been used,
it would doubtless have become apparent that even with this size
of quadrat the drift figure of zero would never be reached, and the
ratio would hold.

While it may seem possible that all the residents of a quadrat
might not be captured until after a three-day period, it must be
remembered that some drift is occurring from the onset of trapping,
and that in all probability this more than compensates for the resi-
dents uncaptured at the close of the three-day period (cf the data
obtained from insulated quadrats, p. 53, Table XID). Whether this is
true or not depends largely upon the thoroughness with which the
quadrat is trapped.

In compiling the averages from the daily yields of the various
quadrats, it became quite apparent that weather conditions greatly
influence the wandering of insectivores, particularly shrews, and that
a cool, humid night preceding or following rainy weather caused a
marked increase in the rate of drift of these animals, while unusually
hot and dry nights caused a sharp decline. Since Blarina is the most
abundant of Ohio’s mammals in nearly all habitats, it is obvious
that for any one quadrat rather spectacular peaks may occur in the
daily records far beyond the three-day limit, although in normal
years these do not usually reach the proportions of the third or
second-day yields. The sixth day peak in the 1933 record was caused

Dee.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 61

by a cool humid night preceding a light rain in the case of two of
the three quadrats comprising the record, and was made up of a
sudden yield of Blarina and Sorex fumeus. Where large numbers of
quadrats are used, however, these peaks are usually leveled off and
become lost, although in 1938 a well-marked peak occurred on the
fourth day (See Fig. 3). Even here, however, the ‘‘peak’’ figures are
inconsequential in comparison to the enormous first-day yield, and
the peak disappears in the line representing the average daily yields
Over a six-year period.

Admittedly the selection of a three-day limit as a standard, as
is done in this paper, is somewhat arbitrary. All that can be said
is that the characteristic drift curve becomes pronounced either at
or before the third-day, and rarely if ever after it, and that for
this reason even a two-day limit seems to be more accurate than the
four-day one used by Elton, at least for all the districts in which
the writer has trapped. It might be added that 150-foot quadrats
used in California in 1934 and in Maine in 1936 supported the data
here to be given. The western quadrats will be treated in another
paper.

Truly accurate data on the length of time necessary to trap the
entire resident population of an area can admittedly be obtained
only by the long use of quadrats thoroughly insulated from drift.
The anomalous character of the results of preliminary experiments
with insulated quadrats has already been indicated.

Dice (1938, p. 126) states that it is unnecessary to kill the
animals taken in the course of quadrat research. The writer will
be unable to agree with this statement until some satisfactory
method is devised for the taking alive of shrews and moles, which
often comprise nearly two-thirds of the small mammal population
in Ohio. Live-trapping for Blarina, Parascalops, and Condylura is as
yet an unfathomed problem, to the writer’s knowledge, and if
accurate data are required on a small mammal fauna these species
must be taken.

How close together may two quadrats be placed? The answer is
that in open fields they may be placed at least as close as their own
diameters. Three quadrats were operated on the Holden Arboretum
in 1938 along the hawthorn-covered banks of a shallow ravine.
These were placed in a row up the draw, the second but 50 feet from

Sci. Pub.

62. THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

the first, and the third about 150 feet from the second. All three gave
yields of all species normal for the habitat. There was not the
slightest evidence of there having been any drainage of mammals
from one to another; in fact, the first quadrat, somewhat less favor-
able for mammals, gave smaller initial figures than the second
and third. Synaptomys appeared in the second but not the first or
third. Sorex cinereus and Mus appeared in the third only, and S.
fumeus in the first and third, but not the second. Peromyscus leucopus,
Microtus, Zapus, Condylura, Parascalops, and Blarina were taken in
all three. It might be added that in these quadrats trapping was
initiated in sequence, and that the first was operated throughout
the trapping period of the second and third, and the second through
that of the third, which was run for four days only.

In forest habitats quadrats should undoubtedly be placed farther
apart than this. A 50-foot quadrat placed about 200 feet from a
grid operated at the Baldwin Bird Research Laboratory in 1935
gave every evidence of having been seriously affected. However,
trapping in the smaller unit was not initiated until the grid had
been trapped for six days. Three hundred yards is a safe distance;
no disturbances in the yields have been noted where the quadrats
have been this far apart.

The simplicity of operating break-back traps, as compared to
live-traps, and the simplification of the resultant record-keeping are
points in favor of the present method. There are also a number of
other practical points that recommend the quadrat system. For one,
a 150-foot quadrat is an excellent de-mouser for areas somewhat
larger than itself, and is a useful instrument for rodent-control under
certain conditions. Furthermore, there is no other way in which
examples of rare species in a fauna may be so readily secured. In
ordinary trapping, long trap-lines, no matter how long, usually
exhaust the more abundant species before the more desirable forms
are taken. This often takes several days. A 150-foot quadrat, if
properly set, yields its Napacozapus, Synaptomys, and rarer shrews
on the first night, along with the usual flood of Blarina or Microtus.
The taxonomist who is trying to build up a collection or who needs
series of certain rare species would do well to use the quadrat method
to the exclusion of ordinary trap-lines. For many years Ohio natural-
ists have pronounced the species of Sorex, Pitymys, and Synaptomys,

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 63

rare and local; the quadrat method reveals that these forms are
generally distributed and locally abundant.

The reader has already been referred to the work of Townsend
(1935) with regard to the bait preferences of certain small mammals.
On theoretical grounds it would seem that the use of bait should
very definitely increase the rate of drift at least for certain species.
With this in mind, two quadrats were operated in 1939 without
any bait whatever. No differences could be noted in the yields of
baited and unbaited quadrats, although the test was scarcely a fair
one for Peromyscus leucopus, which in 1939 was very low in numbers
anyway. During the last six years very few of the quadrats were
baited during wet weather. Another objection has been noted in
the course of baiting: it attracts insects to the traps, and hence
shrews; and attracts larger mammals also, — foxes, raccoons,
opossums, and fox squirrels, which snap the traps and break into
covered mole sets. This was found to be a serious source of trouble
im 1932 when meat baits were regularly used. Since that time rolled
oats have been the standard bait used by the writer and his assist-
ants for all mouse traps; and dried fruits for the rat traps used for
chipmunks, red squirrels, and flying squirrels.

Nearly all of the work heretofore reported was carried on in
climax forests of different types. During 1936, 1937, 1938, and 1939,
a large amount of data was obtained on the populations of many
species of small mammals in a large number of different habitats.
It seems unnecessary to record in detail the support given to the
conclusions deduced from the work in earlier years. Suffice it to say
that the following facts remain clear: (1) that the 150-foot quadrat
is far better to use than any smaller size, unless but one trapper is
available, in which case the 100-foot unit should be used; (2) that
the basing of estimates on center quadrats of grids is faulty due to
the difficulty of getting figures from a sufficiently large number of
units, and that in consequence it is both unnecessary and undesirable
to subdivide acre or half-acre quadrats into grids of smaller ones
except as it may facilitate mapping or trap-setting; G) that esti-
mates of populations for any habitat should be based on more than
one half-acre quadrat wherever possible; (4) that a// parts of the
quadrat must be of similar ecological constitution; (5) that the
three-day limit is under most conditions the proper time allowance

Sci. Pub.

64 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

and (6) that violent rain storms and severe drought may alter condi-
tions sufficiently to make a longer time allowance desirable. In
general, census-taking should be avoided during excessively dry
weather. If a heavy thunder-storm snaps most of a quadrat’s traps
before the third day, the work should be done over in a new place.

POPULATIONS

From time to time in recent years there have appeared figures
on the populations of small mammals indicating that certain species.
may run as high as 100 individuals to the acre, or very much higher.
Williams (1936) published figures on Blarina brevicauda and Pero-
myscus leucopus of over 200 per acre at North Chagrin Metropolitan
Park in Ohio, as previously discussed in this report. Townsend
(1935, p- 49) records 104 per acre for Blarina; Hamilton (1937)
reports that Microtus pennsylvanicus attained figures of from 160 to
230 per acre during the 1935-36 mouse year in New York State;
Elton, Davis, and Findlay (1935, estimate densities of over 100 and
less than 200 per acre for M. agrestis in Scotland. These figures are
minimal indeed compared to some others. Hall (1927) estimates
4,000 per acre for Microtus californicus during a mouse plague in
California, while Mus musculus was estimated at the colossal figure
of 78,000, an all-time high for such studies. Piper (1909) gives
figures of 8,000 to 11,000 per acre for Microtus montanus in Nevada.
That such figures are not reported solely by American workers is
indicated by an estimate of 6,000 per hectare for Maicrotus pelliceus
in the Soviet Union [Pidoplichka, reported by Vinogradov, (1934)].
A population of mice such as 10,000 per acre means one mouse for
every four square feet of ground, and it is obvious that no habitat
could stand such a population except during the passing of a great
migrating horde. The writer has never encountered any such popula-
tions while using 150-foot quadrats in Ohio, Maine, Massachusetts,
and California. Only once during seven years of quadrat work have
we found populations greater than too per acre for any one species.
At Aurora Pond, Ohio, in 1938 Microtus pennsylvanicus registered a
total of 118 per acre in a rich, wet meadow. This density the writer
believes to have been artificial, caused by a flooding of part of the
normal habitat through the raising of the pond’s water-level. In

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 65

the quadrat in question a Microtus was captured in every other trap
of the 150-foot quadrat during the three-day trapping period. It is
frankly not conceivable to the writer how populations could be
much denser, but this nevertheless made him more tolerant than
he previously had been of some of the published figures.

A summary of the populations per acre of some Ohio small
mammals follows. All the figures are based on the three-day yields
of 150-foot (45-meter) quadrats (See Table XIV). The species may
be treated individually. The cyclical aspects of the figures will be
treated in greater detail in a forthcoming paper.

Blarina brevicauda. Suort-Taitep SHREW

Blarina, the short-tailed shrew, is the most widely distributed
and most abundant mammal in Ohio. Although it is locally out-
numbered in hayfields by Macrotus, and in forests by Peromyscus
leucopus ot Sorex fumeus, its general average is higher than for any
other species. Preliminary records seem to indicate that Blarina is
non-cyclical, although severe drought may curtail the population
sharply over a period of years. Populations have run from o to 30
per acre in grassland habitats, with an average of 18; in forests, the
avetage is 26 per acre, with the highest population recorded 48 per
acre. During the 1933-34 drought period figures sank as low as 4 and
Io per acre in habitats that since 1935 have yielded from 20 to 4o.
Nothing approaching Townsend’s 104 per acre (1935, p. 49) has

been recorded under natural conditions.

Cryptotis parva. Oxp-Fie.p SHrew
This tiny shrew is a very rare animal in northeastern Ohio, and

has been taken just once inside a quadrat, in a fallow field habitat
at North Chagrin Park.

Sorex fumeus. Smoxy SHrew

Smoky shrew populations are very susceptible to drought, and
apparently are affected by cycles as well. During the 1937-38 Sorex
peak their numbers reached 15 per acre in upland forests, while in
one instance at Little Mountain, a population of 58 per acre was
recorded, one of the highest ever discovered for any species in the

Sci. Pub.

66 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

course of this work. In habitats other than climax forests Sorex
fumeus is distinctly uncommon, the average being less than 1 per
acre.

Sorex cinereus. Canapa SHREW

The transitory characteristics of Ohio Sorex cinereus populations
afford special problems in determining peaks and population figures.
During favorable G. e., cool, wet) years, this species is found chiefly
in old fields and on fallow land; during droughts the species retires
to sedge meadows, cat-tail marshes and swamp forests. Such migra-
tions may occur several times in one year. Where suitable dry-
weather habitats are lacking, Sorex cinereus does not occur in Ohio.
Figures available indicate that the species was at peak in 1937 with
populations of from 1 to 11 per acre. High figures (5 per acre) in
fields during 1938 were caused by excessively high water-levels in
bogs and marshes generally. During other years the population is
usually much lower than 1 per acre.

Parascalops breweri. Hatry-Tamep Moz

The hairy-tailed mole is one of the most important forest mam-
mals in Ohio from an ecological standpoint. Its numbers were at a
great peak in 1936 during which populations of as high as 24 per
acre were noted. The average for climax forests was 7.6, as opposed
to figures of 1, 2, or 3, per acre in other years. In other habitats
the population has usually been in the vicinity of 1 or 2 per acre.
The species avoids most communities of the hydrosere except during
drought years. Trapping Parascalops with mouse-traps, the only
practicable means in quadrat work, requires a special technique of
setting which is only learned through experience. The writer's
limited experience with Scalopus, which no longer occurs in the
region under consideration, indicates that the same methods will
not always work with that form.

Condylura cristata. Strar-Nosep Motz

Star-nosed moles offer few special problems in trapping, and a
population may be easily removed with mouse traps. Although
climax forests are not this species’ favorite habitat, they constitute
the best areas for cyclical studies. It is evident that they were at

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 67

peak in 1936, and the highest population noted, 6 per acre, was
recorded in a sedge meadow during that year. Ordinarily this mole
is more abundant in swamp forests than elsewhere. Its concentra-
tions there are dependent upon water levels. During very wet years
it freely invades hayfields and fallow land.

Tamias striatus. CxHipmunK

The cyclical nature of seed production of beech and sugar maple
seems to be the basic factor regulating the changes in chipmunk
population densities. In 1935, following a year of great seed pro-
duction by these trees, chipmunk populations reached a figure of
8.7 per acre in upland forests, a density that is undoubtedly ex-
ceptionally high, as similarly favorable conditions in 1937 resulted
in but a very slightly higher figure (1.14 per acre) for 1938 than the
lowpoint of the previous year (0.8 per acre).

Sciurus hudsonicus. Rep Squrrre.

Red squirrel populations in Ohio were at their highest in 1935
(1.7 per acre), since which time they have steadily dwindled to a
point where they no longer figure in quadrat records, a truly remark-
able decline. Careful counts made at Little Mountain over a period
of years corroborate the results obtained with the quadrats.

Glaucomys volans. Fryinc SquirReEL

Like the other herbivorous species treated in this paper, flying
Squitrels were at peak in 1935, when populations were generally
in the vicinity of 3 per acre in woodland habitats. A sharp drop
to 0.4 occurred in 1936, since which time there has been a slow
recovery. The largest population ever noted was one of 6 per acre
during the late summer of 1938, in a beech-maple forest. During
this year a low peak was registered. In 1939 the species suffered a
virtual eclipse, the writer's field assistants having been unable to
supply the Cleveland Museum’s Preparation Department with
enough specimens for a small habitat group.

Peromyscus leucopus. Derxrr Movusz

Great cyclical disturbances occur in the deer mouse populations
of any Ohio woodland, but the numbers of this ordinarily abundant

Sci. Pub.

68 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

mammal apparently do not approach the figures reported by other
workers for other localities. A density of 46 per acre, recorded in
1938 in an isolated beech-maple woodland, was the greatest noted.
Average figures for all upland forest habitats were 29 per acre
during the 1935 peak. In 1936, during the ebb of the cycle, Peromyscus
became positively uncommon, 5.2 per acre, a figure below that of
Sorex fumeus and Parascalops breweri for the same year. Recovery
progressed steadily after 1936 and an average figure of 22.2 per
acre was attained by the summer of 1938. Staple food crop failures
during this year curtailed the trend, however, and the figure of 22.2
in 1938 marks another maximum which completes a second three-
year interval between peaks. During 1939 Peromyscus virtually
vanished from Ohio’s forests.

In other habitats deer mice come and go, depending on conditions
within the upland forests. Populations are rarely as great as they are
there. In general thorn-apple scrub and fallow fields are not in-
habited unless populations are nearly at peak in the forests. Swamps
and bogs show fluctuations corresponding to changing water-levels,
and in open country, far from scrub or woodland, this species is
replaced by the next, P. maniculatus bairdi.

A curious population of 38 per acre was noted in a flood-plain
woodland along the Rocky River west of Cleveland in 1936. This
was more than twice as great a population as was noted in any
other habitat during that year, was more than six times as great
as any population at any other station, and was all the more remark-
able in that the habitat is periodically under from one to three feet
of fast-moving water each winter. In general it may be said that the
herbivorous species of small mammals were all at figures reminiscent
of the 1935 peak at Rocky River, where for some reason, to date un-
discovered, the factors tending to liquidate the 1935 peak were not
effective.

Peromyscus maniculatus. Wuite-Footep Mouse

Prairie white-footed mice are rare and local in northeastern
Ohio, except on the beaches and sand-dunes along the shore of Lake ©
Erie. Here, in 1935 transect trap-line records used in conjunction
with a 50-foot quadrat indicated that a 150-foot quadrat might have
yielded very high figures. By 1936, when the first large quadrat was

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 69

used, populations were still at 22 per acre, although far less numer-
ous than in the previous year. The habitat has not been trapped
since. Elsewhere, the highest population noted is 4 per acre, with
the average for fallow fields of 1.5 in 1938, 0.7 in 1937, and no
records whatever in 1936. These fragmentary records suggest that
1935 and 1938 both marked peak years for this form, as for other
herbivorous species.

Microtus pennsylvanicus. Fre.p Voz

As usual elsewhere with most workers, it is a species of Macrotus
that holds our Ohio records for the greatest number of individuals
per acre, as noted previously. Figures of 118 per acre in a rich
meadow in 1938 constitute an all-time high. A second quadrat
nearby reduced the average for that particular habitat to 85 per
acre, which is still higher than any recorded elsewhere. The species
was at peak in 1929, 1932, 1935, and again in 1938, although there
were no factors now apparent that should have tended to keep the
populations from further increasing in 1939. Actually, sharp de-
creases were noted in 1939. The 1935 peak was by far the greatest.

Microtus is generally distributed in all habitats in which
herbaceous vegetation is continuously abundant throughout the
year. Populations often show great numerical variations in any
given year, but the general tendency has been strongly cyclical in all
habitats save those in which water levels have caused migrations
of seasonal character. The 1936 figure of 28.3 per acre in fallow
fields is far higher than it should be as it includes in its record but
three quadrats and one was in a field at Rocky River, already
alluded to, in which a single quadrat yielded a figure of 76 per acre.

Pitymys pinetorum. Pine Vote

Pine voles are forest-edge animals in Northern Ohio. During
cyclical maxima they spread inwards into the forests and are then
to be found throughout the larger woodlands in the Cleveland
region. The North Chagrin Park figure of 18 per acre (1935) is by
far the highest ever recorded by means of quadrats, and brought the
avetage figure for that year to 6.4 as compared to the ebb-cycle
figure of 0.5 per acre in 1937. The 1938 record shows that Pztymys
was again at peak.

Sci. Pub.

79O THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

Synaptomys cooperi. Lemminc Vote

Ohio’s only lemming is generally distributed but not at all
common. The somewhat meager evidence available from small
quadrat and trapline sources indicates that a very high peak oc-
curred in 1935, when the species was first reported for the region
under consideration. Just how great the populations were at that
time is somewhat of a mystery as but one quadrat was operated in
a habitat suitable to this species. This gave a figure of 4 per acre.
The highest figure obtained since 1935 is one of 6 per acre at the
Holden Arboretum in July, 1938. The average for this year, 1.5, is
no higher than that for 1937. That the species was at peak seems
to be indicated by the fact that it appeared in a greater variety of
habitats in 1938 than in earlier years, although the figures in every
case are very low.

Zapus hudsonius. Hupson-Bay Jumprnc Mouse

Up to the present time (1939) the figures obtained through the
use of quadrats have failed to indicate the existence of a well-
marked cycle in the population-changes of this jumping mouse. The
highest figures recorded for this species are 18 per acre in a rich
meadow and 12 per acre in a swamp shrub (Alnus-Cephalanthus
associes) community. Both figures were obtained in 1936. In fallow
fields the maximum figure of 3.5 per acre was recorded in 1939, while
in brushy scrub (Rhus-Crataegus associes) the maximum was ob-
tained in 1938. Such figures indicate plainly that Zapus, like Sorex
cinereus and Miacrotus, is greatly influenced by local and seasonal
variations of water levels, relative humidity and general precipita-
tion as reflected by lushness of herbaceous vegetation. Zapus is not
a regular inhabitant of upland forests, the most stable of Ohio
habitats, and in consequence cyclical records will be difficult to
secure. That cyclical maxima do occur is very probable, in view of
the periodic outbreaks of this species in Ohio hayfields. It has not
yet been the fortune of the writer to be able to work in such a
locality.

Napaeozapus insignis. Woop.ianp Jumpinc Mouser

Trapping has been continued long enough in the case of the
woodland jumping mouse to establish the existence of a cycle, but

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 71

the length of it has not been determined. The records are very
much less numerous than for Zapus, but as this species inhabits
upland forests, they are relatively little affected by water-table and
climatic factors. A peak occurred in 1937, when populations reached
a figure of 2.8 per acre, while an ebb apparently occurred in 1934
and 1935, when the species was not taken in any quadrat. The high-
est figure noted is one of 10 per acre at Shady Lake, Ohio, in a bog
forest (Acer-Betula associes) in 1938. This is the first and only time
the species has been taken in this habitat.

Mus musculus. House Moust

House mouse records are scattered and irregular and show no
convincing evidence of peaks or other manifestations of cyclical
phenomena. A population of 4 per acre in a sedge meadow habitat
is the largest yet on record for their species in Ohio.

Other species

Besides the forms listed above, other species are occasionally
taken inside quadrats. In all cases these are larger mammals, and
the irregularity of their capture is in line with their very apparent
low density of population as compared to the smaller forms. The
Northern gray squirrel (Scéurus carolinensis leucotis) and the cotton-
tail rabbit (Sylvilagus floridanus mearnsi) ate occasionally captured
in the rat-traps set for chipmunks and flying squirrels, and at least
once during our work with quadrats of different sizes each of the
following have been taken: the skunk (Mephitis nigra), Virginia
opossum (Didelphis virginiana), least weasel (Mustela rixosa) and
Norway rat (Rattus norvegicus). The quadrat method is not effective
with these larger species, since there is a great disparity in size
between the quadrat and the home range of the animal.

Only four of the terrestrial small mammals known to occur in
northeastern Ohio have so far eluded capture in the quadrats. One
of these is the New York weasel (Mustela frenata noveboracensis) and
two, the Red-backed vole (Clethrionomys gapperi) and Common mole
(Scalopus aquaticus) are probably extinct today; the last is a very
rare shrew (Maicrosorex hoyi thompsoni). The fact that these species
have not been taken is evidence of their relatively great rarity.

he number per acre and are the averages

igures give t

Tasrz XIV. Small Mammal Populations in the Cleveland, Ohio Region, 1935-39. F

of estimates based on the yields of single 150-foot quadrats.

THE CLEVELAND MUSEUM OF NATURAL HISTORY

Number of Quadrats WO nAtTH ted di
in Average Car
3 3 . . . ead . . . .
Sylvilagus floridanus ee oe ae
. . . 2 Sap . . .
Mus musculus Th oy eal ee
Napaeozapus insignis sine Seats t ES
OFA a © a)
. 2 Lal “Ww O
Zapus hudsonius oe ;
alow 4 Ki

Synaptomys cooperi

Pitymys pinetorum

4 ES . . -_ on. wo Oo oO
Microtus pennsylvanicus hee as Lae Ss
Peromyscus maniculatus pate skehe

al fo)
Peromyscus leucopus Cy er ee Myon
a. wn Si as wewv rho
Sciurus hudsonicus ie oh: OMIT Ws
eH OOO Lan | .
Gl ] pe are per iuaeee
aucomys volans eae ie ice ares

. . I.00 COR MN Ww 2)
Tamias striatus BE iG ie ark ae
. -CO Went Nae) Ge)
Condylura cristata [6066 Ht 10

Parascalops breweri Oe eae Pa edo
ANON AY al H

Cryptotis parva

RENO eel oc0o0°0
Sorex cinereus HOU CONG Al Gl
NmMONhKRR Vaile}, (0) KO)
Sorex fumeus e009 > OV on Ommt
al PSNI etGN 0000
2 ) { ) ,
Blarina brevicauda i SSeS alban Woloe ear
2 e
I
BS NOSSO ON ire, DOTS EO IN
Hasitat and YEAR é ENON CNIS 9 AVE
O faa)

Boc Mgapows

Sci. Pub.
Vol. V

had

wy
13.0

I.O
ZO

oe
.

Dec.

eb

TasLe XIV Cont.

Number of Quadrats
in Average
Sylvilagus floridanus
Mus musculus
Napaeozapus insignis
Zapus hudsonius
Synaptomys cooperi
Pitymys pinetorum

Microtus pennsylvanicus

Peromyscus maniculatus
Peromyscus leucopus
Sciurus hudsonicus
Glaucomys volans

Tamias striatus

Condylura cristata

Parascalops breweri
Cryptotis parva
Sorex cinereus
Sorex fumeus

Blarina brevicauda

Hasitat and YEAR

Fattow Fieips

aes +
Isis 3°)
ome) 4

NDDWNWN

Rico Mrapows

B. P. BOLE, JR.—THE QUADRAT METHOD

=1@ - Oo
tomy Nts
=©
a
OROTOeO 30) O50
CO NOT clact i=) pyar
“© (9) {O)(e) 9 KO)
A THN se a
2
\O
02009 ©9000
\O w Wo Fra nd oO
me HOO -
(O)" to! Oy OC
a a) <= =) ©
AO - AR aA
Oo
a
Oi -24@
a ca
OoO0oO°0
CO mF
2 10)
La |
Oo 0000
a SS cg) I~ Si;
=
10,0 ac
Sales sin
Oo
—-
al
e' O70 © 00000
O Wevo tO CO TO
SS ge a
Q
jaa
\O KOO DW B@ nwo ROO A
AAAM AMMAMAS
ANANNDD ONDADADAASA
Pua SARS te

THorn ScrusB

Z

ab

Oo

2GOn 04

1938
ey)

6.0

Froop Puan Forest

Oc oO

27h (©)

1936
CoastaL SAND DuNES

eeoee

22

1936

Sci. Pub.

74 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

CONCLUSIONS

From a study of the yields of small mammals obtained from
several sizes of trapping quadrats over a period of eight years, it is
decided that:

(1) The quadrat method provides an accurate means of determin-
ing small mammal abundances;

(2) In general, the larger the quadrat the more accurate the
results, and that quadrats 150 feet or 50 meters on a side or circular
units of corresponding area are still small enough to be easily
operated and are large enough to provide a high degree of accuracy,
and hence should be considered the standard sizes for population
investigations of mammals whose home ranges are of the same size
or smaller than that of the quadrat;

(3) The use of shelter-belts or outer tiers of quadrats about a
central one is not justifiable because of the tremendous labor in-
volved if the quadrat is maintained at standard size, because the
belts themselves cause population disturbances within the quadrat,
and because the data the belts yield can be put to better statistical
use if the traps that furnished them are part of the quadrat itself;

(4) More than one quadrat should be used in any given com-
munity, but should be placed no closer to eachother in forests than
300 yards, and no closer than 100 yards in fields;

(5) All parts of a quadrat must be ecologically uniform and
similar to the larger area which it is meant to represent, if figures
on populations for one particular habitat are desired;

(6) A three-day trapping limit should be set, as the resident
population is practically exhausted in this time;

(7) Trapping within a quadrat must be thorough and executed
with the sole purpose in mind of getting all the mammals out of it
as swiftly as possible, for which reason traps should be set in under-
ground or surface runways or wherever the mammals are, and should
not be set at regular intervals within the quadrat, since the lairs
and runways do not so arrange themselves in nature;

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD 7)

(8) The use of insulated quadrats is futile as the metal sheeting
used will cause spectacular changes in the ecological picture of the
quadrat, particularly if moles are part of the small mammal fauna;

(9) The use of very small quadrats (smaller than 100 feet on a
side) is indefensible owing to the relatively high percentage of
animals drifting into them from adjacent terrain;

(zo) Quadrats are found to be particularly useful in studying
cyclical changes in small mammal populations and in collecting
the less common species of a given small-mammal fauna.

LIST OF REFERENCES

Aldrich, J. W. CMS.) Ecology of the Swamps and Bogs of North-
eastern Ohio. (Unpublished manuscript in the library of
Western Reserve University).

Chitty, D. (1937, May) A Ringing Technique for Small Mam-
mals. Jour. Animal Ecol., Vol. 6, No. 1, pp. 36-53, 5 figures.

Dice, Lee R. (1931, Nov.) Methods of Indicating the Abundance
of Mammals. Jour. Mammalogy, Vol. 12, No. 4, pp. 376-381.

(1938, July) Some Census Methods for Mammals.
Jour. Wildlife Management, Vol. 2, No. 3, pp. 119-130.

Elton, Charles (1924) Periodic Fluctuations in the Number of
Animals: Their Causes and Effects. Brit. Jour. Exp. Zool.
Vol. 2, pp. 119-163.

[amma (2925) Plague and the Regulation of Numbets in
Wild Animals. Jour. Hygiene, Vol. 24, pp. 138-163.

Elton, C.; Davis, D. H. S.; and Findlay, G. M. (1935, Nov.)
An Epidemic among Voles (Microtus agrestis) on the Scottish
Border in the Spring of 1934. Jour. Animal Ecology, Vol. 4,
No. 2, pp. 277-288.

Grinnell, Joseph (2914) An Account of the Mammals and Birds
of the Colorado Valley with Especial Reference to the Dis-
tributional Problems Presented. Univ. Calif. Publ. Zool. Vol.

TPM 5 E-2-04:

Sci. Pub.

76 THE CLEVELAND MUSEUM OF NATURAL HISTORY Vol. V

Grinnell, J. and Storer, T. I. 1924. Animal Life in the
Yosemite. Berkeley. Univ. Calif. Press. 752 pp., 62 plates.

Hall, E. R. (1927) An Outbreak of House Mice in Kern County,
California. Univ. Calif. Publ. Zool. Vol. 30, No. 7, pp. 189-203.

Hatfield, D. M. (2938, May) Studies on Rodent Populations of a
Forested Area. Jour. Mamm. Vol. 19. No. 2, pp. 207-11, 2 fig.

Hamilton, W. J., Jr. (1937, May) The Biology of Macrotus
Cycles. Jour. Agr. Res. Vol. 54, No. to.

Jaccard, Paul. (1912) Distribution of the Flora in the Alpine
Zone. New Phytologist, Vol. 11, pp. 37-50.

Kenoyer, L. A. (1927) Raunkaier’s Law of Frequence. Ecol.,
Vol. 8, No. 3, pp. 341-349.
MacLulich, D. A. (1937) Fluctuations in the Numbers of the

Varying Hare (Lepus americanus). Univ. Toronto Studies, Biol.
Ser., No. 43, pp. 1-136, 18 fig.

Middleton, A. D. (1930) Cycles in the Numbers of British Voles
(Microtus). Jour. Ecol., Vol. 18, pp. 156-166.

—_—_—_———— (1931) A Further Contribution to the Study of
Cycles in the British Vole (Microtus). Jour. Ecol. Vol. 19, pp.
190-199.

Murie, O. J. and Murie, A. (1931, Aug.) Travels of Peromyscus.
Jour. Mamm. Vol. 12, No. 3, pp. 200-209.

Phillips, J. F. V. (1931). Quantitative Methods in the Study of
Numbers of Terrestrial Animals in Biotic Communities: A Re-
view, with Suggestions. Ecol., Vol. 12. No. 4, pp. 633-649.

Piper, S. E. (1909) The Nevada Mouse Plague of 1907-8. U. S.
Dep't Agr. Farmers’ Bull. No. 352, 23 pp.

Pound, Roscoe and Clements, F. E. 1898. The Phytogeogra-
phy of Nebraska. I, General Survey. Lincoln.

Ruhl, H. D. (1932) Methods of Determining the Abundance of
Some Game Species. Trans. 19th Amer. Game Conference.

Dec.

1939 B. P. BOLE, JR.—THE QUADRAT METHOD G7

Seton, E. T. 1909. Life Histories of Northern Animals. 560 pp.
C. Scribner’s Sons, New York City.

Shelford, V. E. 1929. Laboratory and Field Ecology. The Re-
sponses of Animals as Indicators of Correct Working Methods.
pp. 1-55- Baltimore. Williams & Wilkins Co.

Taylor, W. P. (1930) Methods of Determining Rodent Pressure on
the Range. Ecol., Vol. 11, No. 3, pp. 523-542.

Thomas, E. S. and Bole, B. P., Jr. (4938, Feb.) List of the
Mammals of Ohio. Ohio State Museum Mimeographed Bull.
Movs. 2 pp. —

Townsend, M. T. (1936, July) Studies on Some Small Mammals
of Central New York. Roosevelt Wild Life Annals, Vol. 4, No.1.

Vinogradov, B. S. (1934) Materials for the Study of the Dynam-
ics of the Faunas of Muriform Rodents in the U.S.S.R. (Histori-
cal Review of Fluctuations of Mice in the U.S.S.R.) 61 pp.
Leningrad. (U.S.S.R. People’s Commisariat Agr., Assoc. Con-
Pa@lv@t bests and WDiseases, Rec. Serv.) (Text and Titles in
Russian, with English summary).

Williams, A. B. (1936, July) Composition and Dynamics of a
Beech-Maple Climax Community. Ecol. Mon., Vol. 6. No. 3,
pp. 318-408; also Sci. Publ. Cleve. Mus. N. H., Vol. 6, pp. 1-96.

+t

i
ig is
3 \) i}

, 44 :
an ecal
4
L.@ ey
F qs
ii :
Hy df’
7
= et ,
- a
,
j

5 Var

A

‘
i's
Ce
(
ae
ir

ce)

Dp gare A

FS ag B ge! ne

ne

a tae

I

a 5 a Lb Cl ~~ ey ~
= Oe = © G3. 3) =©
2 SAG 2 — OK) 2 ead) = X
m S\ | g m : < m FiNgsA>> Zz
ARIES) SMITHSONIAN INSTITUTION, NOILALILSNI NVINOSHLINS, Saluvuaia Stl
= eran 2 ce = ~~ =2
= = a Mi, = = z
= ro) si a My oO eo oO
YK wo a n Uff pe AD w 7)
\ SE E 24 Gt 2 =
ee ar thar ioe z = a
JLILSNI_ NVINOSHLIWS S3iY¥vus tot IBRARIES SMITHSONIAN _INSTITUTION NO
Z . = Pe z .
” 7) 77)
4 ee ~ ane = rena
= = = - eo < (3
ine e ZS e ae
= = ~ N
5 = 5 si rs} a
Zz a = a 2 a
ARIES SMITHSONIAN INSTITUTION NOILONLILSNI NVINOSHLINS S3I1uVvud ees 2
S as & ~ Ca z =
“i a6 ro) it A
= oO = = ‘ = o
5 2 5 2 NX 5 z
te : = a > Ss - > |
hy as = 2 = de
ALILSNI NVINOSHEIMS, 53 iuyvua roti BRARI ES SMITHSONIAN INSTITUTION | N:
i= = = NS = < NS =
z = z = z 5 =
2 B 3 z A Qe
= oO ¥ = 5 oO aa O ?
= Zz = = E- z
ee ee ; : 3 :
+ wa al
RARIES SMITHSONIAN _ INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVS Brice
af ' 2 ¥ Zz x“ . =
es Pa ze 7) WZ MS ani ”
. = ox a Nc S a
< 3 < 4 WKS < a |
: : : 2 {5 -
ee ie Seale 3 aes S
E ‘ z = er et ra
NLILSNI_NVINOSHLINS S3I1uva @iti LIBRARIES SMITHSONIAN INSTITUTION _N
= = Bi = ee =
oO oS om — @ —
= SENN = Pe) = ; = =
NS \ = = = 2 Z
a \ ad r =
RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NYVINOSHLINS S3I¥Vvesit L
a Fu ‘3 Ww z= wo . ad
= — ae = SS oe = =
& 5 s\n 22: : 5
oS ee 5 = 5
INLILSNI _NYINOSHLINS S23 iyvuag is Too BRARI ES SMITHSONIAN _INSTITUTION ei.
ee * WwW = - uw Ys, = =O | ul
oi - S = Weim = =
z5 = 3 © LH <
= ¥ < = < ey c <x
ear ao = oF = 4
3S se re) al 3 si
Fa : 3 z er z +
SRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS S23! YUvVusIT |
2 s z = z es
SHLD 2) x w Oo o AS = Oo res]
? 4 oe f ae) Sey >) 3 ke Ww
Lm) 5 If 2 = > Ke :
= — Vg a = 8) = a

=> “”) = nN = ¢
«8 us a ul b
a oc a ow pa
= < =A <x — «

es Cc
a. = ca = : = :
et oO -- ro) a e) 2
J = a Zz ee | Fa
-NVINOSHLIWS (S31NVUaIT LIBRARI ES_ SMITHSONIAN INSTITUTION _ NOILOLILSNI-
: : 2 gy = : 3
5 A > A > z
= a = <2 is
a a a =
Zz o = a =
INSTITUTION NOILNLILSNI NVINOSHLINS SJINVYSIT LIBRARIES
ee cy) z w Ze (
< = 7 < Sy ee :
z = 4, ee = yy Z :
oO i “Myf. OC a G; Li O S
” Oo Joma?) n Go. 7 ¢
= we @ | G Ul x O pf ic “3 D8 C
= 2, = Sent = aa
= ees 2 B 2 a Se
S3IYVYGIT LIBRARIES INSTITUTION NOILNLILSNI
ul 2 B 2 o
pe O = on a C
ce = faa Le st foe
| WN c = = = <
= SS Sa ro} pi oO zs
= a | a —! =z — =
Ss ITHSONIAN INSTITUTION NOILNLILSNI NWINOSHLINS S31y¥vuUgIT LIBRARIES
= Ei 3
e S z 5 ,
= - a - a F
17 wn = 7) San “ :
NVINOSHIINS Sa1YVYGIT_ LIBRARIES SMITHSONIAN INSTITUTION, NOILO.LILSNI_
| ee z . 2
; = < Ww RA = x WS = s
= 5 ANS a Zz 4 ed fi 2
oe °, oO ‘ QS - ae oO Bi Sts C
a: 2X 8 2 EY ff :
z = Y = = 2 tj :
f = 7 a » 2B A .~ 3 ; i
IMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SaIluVvVUugI7 LIBRARIES
za eh > 29) > U
aW? ao 7) aa ”
5 E = = =
<
5 : S : 2
fo) a Oo a fo) :
= oe 2 ag Zz as
S3IYVYSIT LIBRARIES NOILNLILSNI
5 = S — S :
= a = a - eg
E > ra 2} = 2
E Po] a 2 Aa = a
Ee eu ae is = a
_ = a Zz oD Zz pale
MITHSONIAN INSTITUTION NOILALILSNI_NVINOSHLINS SaluVudy LIBRARIES |
all ata z * = = = i z re) z
3 = yr, = =. = =
m7) ow Y yf a” o iD ¢
} =e oO Vf is a e) ae ; ¢
a A S aoe £ iS = ia mS 2
Wasi’ = >” = >" Ss be SY >
Se. ” z 7) z 7) oe 2
NVINOSHLIWS LIBRARIES SMITHSONIAN, INSTITUTION NOILNLILSNI_
& ul <Gll = — Uo NZ ame uw <C> FF
Oe = LMR 2 ES ' Gt ESE ' E® °

eee ee The