BOSTON PUBLIC LIBRARY

3 9999 06317 631 5

57

UNITED STATES DEPARTMENT OF THE INTERIOR

FISH AND WILDLIFE SERVICE

ATTWATER'S PRAIRIE CHICKEN
ITS LIFE HISTORY AND

MANAGEMENT ^^..^^

NORTH AMERICAN FAUNA 57

iiO'

PuSLJCLtBRARY^^.^i
n€UM£»TSUEPfcRTM£kT

I.I—' ' "I

1 2000 I

UNITED STATES DEPARTMENT OF THE INTERIOR

'Harold L. Ickes, Secretary

FISH AND WILDLIFE SERVICE
Ira N. Gabrielson, Director

North American Fauna 57

ATTWATER'S PRAIRIE CHICKEN

ITS LIFE HISTORY AND

MANAGEMENT

BY

VALGENE W. LEHMANN

UNITED STATES

GOVERNMENT PRINTING OFFICE

WASHINGTON : 1941

For sale by the Superintendent of Documents, Washington, D. C. - - - - Price 40 cents

ABSTRACT

ATTWATER'S PRAIRIE CHICKEN, a characteristic bird of the Texas
coastal prairie, is closely related to the now extinct heath-hen of
northeastern North America. Once abundant in an area extending from the
coastal tall-grass prairies of southwestern Louisiana and Texas west and south
to near Port Isabel, it has decreased in numbers as man has exploited its
habitat, until now it is threatened with the same fate as that of the heath-hen.

Important factors limiting the numbers of the bird include excessive or
persistent rainfall during the nesting season, heavy grazing, excessive pasture
burning, agricultural operations, and overshooting. Management will
usually involve protection from excessive killing, improvement of food and
cover, and control of predators and of the kill by hunters. Responsibility
for this rests with the landowner.

Optimum prairie chicken range apparently consists of well-drained grass-
land, with some weeds or shrixbs, the cover varying in density from light to
heavy; and with surface water available in summer; diversification within
the grassland type is essential. In the absence of ample refuges for the
species, probably all other favorable factors together will fail to save
Attwater's prairie chicken from extinction.

This number continues the series of the North American Fauna issued by
the Bureau of Biological Survey, of the United States Department of Agri-
culture, prior to its transfer and consolidation with the Bureau of Fisheries
on June 30, 1940, to form the Fish and Wildlife Service, in the Department
of the Interior.

CONTENTS

Page

Introduction 1

Former distribution of prairie chickens in Texas 2

Differences between Attwater's and the other prairie chickens 4

Attwater's prairie chicken 4

Lesser prairie chicken 5

Former abundance of Attwater's prairie chickens 6

Present distribution and numbers 7

Habits 10

Courtship and mating 10

Nesting 14

Growth and development of young 16

Brood size 18

Juvenile mortality 19

Family disintegration 19

Annual increase 20

Flocking 20

Seasonal movements 21

Spring 21

Summer 22

Fall and winter 24

Food 25

Habitat requirements 30

Kind of environment best suited 30

Character and density of vegetation 30

Topography 30

Water 31

Seasons of scarcity 31

Limiting factors 31

Natural factors 32

Rainfall during the nesting season 32

Floods 35

Drought 35

Hurricanes 35

Hail 35

Local storms 36

Disease 36

Spread of woody vegetation 36

Predation 37

Nests 37

Young 38

Adults 39

Review of natural factors 40

Artificial factors 40

Agriculture 40

Pasture burning 41

in

IV CONTENTS

Limiting factors — Continued.

Artificial factors — Continued. Page

Overgrazing 42

Oil development 43

Drainage 43

Pasture mowing 43

Mechanical accidents 44

H anting 44

Management 45

Protection 45

Habitat improvement 46

Evaluating conditions 47

Census methods 47

Spring counts on the courtship grounds 47

Rope count 49

Car-dog count 52

Using the census 52

Spring 53

Summer 54

Winter 55

General recommendations for habitat control 56

Predator control 57

Harvesting the surplus 57

Restocking 58

Summary 59

Literature cited 62

Index 65

ILLUSTRATIONS

Plate Facing page

1. Attwater's prairie chickens (Tympanuchus cupido attwateri) on boom-

ing ground Frontispiece

2. Dense cordgrass areas in Aransas County, Tex 4

3. Male Attwater's prairie chicken, showing vocal sacs 10

4. Nest and eggs of Attwater's prairie chicken; Colorado County, Tex__ 14

5. Concealment of nests by Attwater's prairie chicken; Colorado County,

Tex 15

6. Chicks of Attwater's prairie chicken; Colorado County, Tex 16

7. Wild indigo (Baptisia) in a closely grazed pasture; Austin County,

Tex 22

8. Diversified cover — excellent prairie chicken range; Colorado County,

Tex 30

9. Medium-heavy to heavy cover — excellent food-cover conditions in a

moderately grazed pasture; Colorado County, Tex 31

10. Shells of eggs at prairie chicken nest destroyed by house cat; Colorado

County, Tex 38

11. Native bluestem prairie — well populated by prairie chickens; Colorado

County, Tex 39

12. Excellent unburn ed cover at right of road; inferior burned cover at

left; Colorado County, Tex 42

13. Rope counting of prairie chickens on Matagorda Island, Tex 48

14. Fenced plot planted to hegari; Wharton County, Tex 56

FiGDHB Page

1. Distribution of Attwater's prairie chickens in Texas 3

2. Movements of a combined brood, Colorado County, Tex 23

3. Rainfall conditions in May in the range of Attwater's prairie chicken in

Texas Facing page 34

4. Diagram of the rope count 50

V

ATTWATER'S PRAIRIE CHICKEN

By Valgene W. Lehmann
Collaiorator, Division of Wildlife Research, Fish and Wildlife Service^

INTRODUCTION

Attwater's prairie chicken {Tympanuchus cupido atticateri Ben-
dire) (see frontispiece), might well be called the heath-hen of the
South. It is so closely related to the now extinct heath-hen {T. c.
cupido) of northeastern North America as to be classified in the
same species. Like the heath-hen, Attwater's prairie chicken once
inhabited a large area, its former range including the coastal tall-
grass (Andropogon) prairies of southwestern Louisiana and in Texas
west and south to Cameron County, near Port Isabel. In certain
areas the birds were abundant. Old-timers report that the deep
booming courtship calls of the males once reverberated from the
prairies with such force and monotony as actually to pain sensitive
eardrums. The bird, however, is no longer abundant. It has de-
creased in numbers as man has exploited its habitat until now it is
threatened with the fate of the heath-hen — extinction.

In his "Biological Survey of Texas" Vernon Bailey (1905: 19)'
places Attwater's prairie chicken at the head of the list of breeding
birds of the Texas coastal prairie. In addition to being a character-
istic bird of the region, this prairie chicken is probably the most
popular species wherever found. Most people who know it have a
genuine appreciation of its color and charm. Rare indeed is the
person who finds no esthetic stimulus in the sight of a strutting
male on the booming ground, or a brood of downy chicks on the edge
of a short-grass flat. Both ranchmen and farmers highly appreciate
the prairie chicken's appetite for grasshoppers, salt-marsh cater-
pillars {Estigmene acraea)^ and the moths of the cotton leaf worm
{Alabama argillacea). Under proper conditions prairie chicken
hunting provides a high type of sport, and the flesh of the birds,
especially that of the young, is highly esteemed as food.

1 Cooperative contribution from the Texas Cooperative Wildlife Research Unit, estab-
lished by the Agricultural and Mechanical College of Texas ; the Texas Game, Pish, and
Oyster Commission ; the American Wildlife Institute ; and the Fish and Wildlife Service.

* Publications referred to parenthetically by date (alone or followed by colon and specific
page) are listed in the Literature Cited, p. 62.

Explanation of Fkontispiecb

Attwater's prairie chickens (Tympanuchus cupido
attwateri) on booming ground

2 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

The real appeal of the prairie chicken, however, lies in its con-
nection with the colorful and eventful early days in Texas. The
prairie hen summons memories ; it prompts old-timers to recall when
the range was free of wire fences and oil derricks, and rich grasses
grew waist high. Thoughtful people deplore the passing of Att-
wat^r's prairie chicken, one of the last landmarks of the prairie as
it used to be. Highly appropriate was the selection of this bird as
a species of major interest by the Texas Cooperative Wildlife Ke-
search unit.'

FORMER DISTRIBUTION OF PRAIRIE CHICKENS IN TEXAS

H. C. Oberholser, in a letter to the present writer, states that in
his opinion prairie chickens once occurred at some time of year on
most prairie areas in Texas. In the main it appears that the differ-
ent kinds of prairie chickens in the State occupied separate ranges,
and that mixing and intergradation were confined largely to marginal
areas.

The principal former range of the greater prairie chicken in Texas,
as indicated by the records of F. M. Bailey (1927: 130), Gross (Bent
1932: 262), Strecker (1927: 321), and old residents with whom the
writer has conferred, was northeastern Texas southwest to the vicin-
ity of Waco. Likewise, records show that the lesser prairie chicken
was indigenous to northwestern Texas and the high plains region
in winter to about Bandera and westward through the "hill country"
to the arid plains west of the Pecos River (Bendire 1892: 355, and
others). Attwater's prairie chicken, it appears, was largely confined
to the better-drained prairies of western Louisiana and Southeastern
Texas (fig. 1, p. 3).

According to Oberholser (1938: 190-191) the eastern limit of the
range of Attwater's prairie chicken was in the vicinity of Abbeville,

* So many persons have assisted in the prairie chicken studies that it is impossible to
list them all. General supervision of the work was by Walter P. Taylor, leader of the
Texas Cooperative Wildlife Research Unit, College Station, Tex. Valuable editorial
suggestions were received from W. B. Davis, professor of wild game. School of Agriculture,
Agricultural and Mechanical College of Texas; and from William J. Tucker, executive
secretary, Texas Game, Fish, and Oyster Commission. The bulk of examinations of crops,
gizzards, and scats was done by Clarence Cottam, Clarence F. Smith, and their associates
in the Section of Food Habits, Division of Wildlife Research, Fish and Wildlife Service.

In his field work in 1938 the writer was assisted by H. R. Siegler, field biologist of the
Research Unit. Many Colorado County landowners cooperated ; among these, M. C.
Shindler, Emil Gleuck, Ed Koy, Adolf Renz, and I. V. Duncan deserve special mention.
E. P. Haddon, photographer of the Texas Commission, took some of the photographs here
reproduced. The assistance of the State game wardens was indispensable. Deserving
of special mention are T. S. Boothe, Beaumont ; J. C. Gardner, Hull ; R. Z. Cowart, Rosen-
berg; Ed McCloskey, Victoria; C. D. Tidwell, Bay City; G. P. Ferguson, Sinton ; and
T. T. Waddell, Eagle Lake. Waddell's contributions to the study were outstanding; he
gave most generously of his time, records, and extensive experience. To him, and to all
others, the writer is deeply grateful.

ATTWATER'S PRAIRIE CHICKEN

Opelousas, and Bayou Teche in Louisiana. There are no authentic
records of the occurrence of any species of prairie chicken in Texas
south of northern Aransas County, except for one bird reported
from near Brownsville by Merrill (1879: 159-160). Prairie chickens
did not occur near San Antonio, Tex., in 1890, for Babbitt, in Bendire
(1894: 130) wrote as follows: "The prairie hen is not found in the

LEGEND

Probable Former Ranqe
Prcaenf Pange

Figure 1.— Present distribution of Attwater's prairie chicken in Texas and
probable former range in the coastal section.

immediate vicinity of San Antonio, Tex., but exists in great numbers
south and southeast from here, all at an average distance of 100
miles. * * *" Simmons (1925: 82) submits the records of O.
Brinkman and C. D. Oldright as evidence that Attwater's prairie
chicken occurred as a breeding bird in the vicinity of Austin. Travis

4 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

County, and in Williamson County as late as 1878, but the accuracy
of the data is questionable. Apparently the limit was the northern
edge of the coastal prairie.

Roughly, the territory occupied by Attwater's prairie chicken was
south of a line extending northeast from Refugio through Fannin,
Thomaston, Provident City, Rock Island, Industry, Welcome, Bell-
ville, Prairie View, Tom Ball, Humble, Liberty, Devers, Cheek, and
Orange. All this area of approximately 8,500,000 acres in coastal
Texas, however, was not occupied. Deciduous woodlands near rivers,
as along the San Antonio, Guadalupe, Colorado, Brazos, and
Trinity, were used only to a limited extent and only along the mar-
gins. Prairie chickens did not occupy the pine forests in Harris
County and to the east or the thick mesquite-acacia brush that oc-
curred in considerable stands in Calhoun and other western counties
as much as 100 years ago. Brackish and salt-water marshes in Or-
ange, Jefferson, and Chambers, and less widely in other counties
to the west, and extensive cordgrass {Spartina spartinae) flats (pi. 2)
in Aransas County and elsewhere in low country bordering the Gulf,
probably always were little used by chickens except to a limited
extent in winter. There were, however, about 6,000,000 acres of
bluestem prairie that probably supported many prairie chickens in
favorable years.

DIFFERENCES BETWEEN ATTWATER'S AND THE OTHER PRAIRIE

CHICKENS

During the nineteenth century three kinds of prairie chickens oc-
curred in Texas: the greater prairie chicken {Tympanuchus cupido
americanus Reichenbach), Attwater's prairie chicken {T. c. attioateri
Bendire), and the lesser prairie chicken {T. pallidicinctus Ridg-
way). Differences between the greater and Attwater's prairie chick-
ens are slight; the lesser prairie chicken is somewhat better
characterized.

ATTWATER'S PRAIRIE CHICKEN

Bendire (1894: 130) described Attwater's prairie chicken as —

Smaller than T. americanus [greater prairie chicken], darker in color, more
tawny above, usually with more pronounced chestnut on the neck ; smaller
and more tawny light colored spots on the wing coverts, and much more
scantily feathered tarsus, the latter never feathered down to the base of
toes, even in front; a broad posterior strip of bare skin being always exposed,
even in winter, while in sximmer much of the greater part of the tarsus is
naked.

In weight Attwater's prairie chicken, however, is not perceptibly
lighter than the greater prairie chicken. The average of 10 males
(33.11 ounces, as shown in table 1, p. 5) exceeded by 2.11 ounces
the average weight of the greater prairie chicken (31 ounces), as

North American Fauna 57. Fish and Wildlife Service

PUATE 2

03

■3 >

« s

« +=

s

a

<J o

>i 00

X3 a>

S3 H

g

fl

o3

=3

0

s

0

o3

(H

O)

bO

03

'2

Hi

0

^

u

^

<tj

ATTWATER'S PRAIRIE CHICKEN

given by Gross (1930a: 40). The average weight of six females
(25.7 ounces) was only 0.6 ounce less than that of the female greater
prairie chicken (26.3 ounces), according to the same authority (loc.
cit.). The weight of Attwater's prairie chickens, especially that of
males, varies perceptibly from season to season ; three males obtained
at the beginning of the courtship season were appreciably heavier
(about 7.5 ounces) than those collected at other times. Darkness of
color, tawniness above, and the amount of chestnut on the neck are
other characters that vary greatly both seasonally and individually.
Winter specimens are generally lighter in color than those collected
in spring; they have comparatively little chestnut on necks and
backs. The feathering on the tarsus also varies with the season ; the
legs of specimens collected in winter are well feathered to the base
of the toes. The style of barring on the back and rump, according
to F. M. Bailey (1927: 130), is the same for both subspecies, that is,
the bars are single, broad, and solid black. Altogether, physical
differences between Attwater's and the greater prairie chicken are
minor and insufficient to allow accurate field identification. In a
series of skins, however, the smaller measurements of wing, tail, bill,
and total length and the differences in general ruddiness and buffiness
of the underparts are characteristic and serve to separate Attwater's
prairie chicken as a subspecies.

Tabui: 1. — Weights of 16 Attwater's prairie chickens

County

Date collected

Weight I

County

Date collected

Weight 1

Grams

Ounces

Grams

Ounces

MALES

Colorado

Do -

Apr. 17, 1939 2
Sept. 1,1937 3
Sept. 4,1937'
Oct. 23,1937
Nov. 3,1937

Jan. 6, 1938
Jan. 27,1938
Feb. 14,1938

1, 135. 20
682. 00
590. 07
760.20
874. 00

723. 69
715.24
726.80

40.03
24.05
20.81
26.81
30.82

25.18
25.22
25.63

MALES

Refugio

Do

Colorado

Feb. 15,1938

do

Mar. 18, 1938
Apr. 10, 1938 <
July 26,1938

Feb. 15,1938
Aug. 20,1938
Aug. 23,1938

1, 103. 70
1, 125. 20
1,120.45
1, 077. 26
921. 34

785. 60
722. 89
708. 72

38.92
39 68

Do.-

39 51

Do

38.00

Do

Colorado

FEMALES

Refugio

Colorado

Do

32 50

FEMALES

Colorado

Do -

27.70
25.50

Austin-

25.00

1 Average weights: Males, 938.94 gm. (33.11 oz.); females, 730.49 gm. (25. 70 oz.).

' Taken from hawk.

' Immature bird.

* Taken from poacher.

LESSER PRAIRIE CHICKEN

The lesser prairie chicken is somewhat smaller than either of its
relatives. Verne E. Davison, in a letter, reports that 20 mature males
from Oklahoma weighed 23.50 to 31.50 ounces and averaged 27.56
ounces. Five hens weighed 23.75 to 27.50 ounces and averaged
25.55 ounces. In other words, these male lesser prairie chickens were,
on the average, 3.44 ounces lighter than the male greater prairie
chickens (31 ounces) weighed by Gross (1930a: 40); these female

6 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

lesser prairie chickens weighed 0.55 ounces less than the female
greater prairie chickens that Gross obtained. According to Bailey
(1927: 131), the general coloration of the lesser prairie chicken is
paler than that of either the greater or Attwater's, and the color
and arrangement of the bars on back and neck also differ. Wliereas
the bars are single, broad, and solid black in both the greater and
Attwater's forms, they are treble, a broad brown bar enclosed by
two narrow black ones, in the lesser species.

FORMER ABUNDANCE OF ATTWATER'S PRAIRIE CHICKEN

Accurate information as to the former abundance of Attwater's
prairie chicken is difficult to obtain, although the data at hand
record their numbers in certain areas. Many old cattlemen of the
coastal prairie have told the writer that in early days the prairie
chickens were relied upon to furnish fresh meat for the cattle camps.
The task of killing 40 or 50 prairie chickens was menial, the cook of
the outfit usually attending to it.

In the Eagle Lake area, Colorado County, not more than 35 years
ago, prairie chickens were shot as clay pigeons are today. On ap-
pointed occasions parties of 10 to 20 or more men encamped in the
sandhill country along the Bernard River and hunted the birds for
periods varying from a few days to a week or more. At the end
of each day the chickens killed, or their heads, were tallied. At the
end of the encampment period the party having killed the smallest
number paid the expenses of the outing. Waddell and others state
that 10 or more piles of prairie chickens, each containing upwards
of 100 birds, usually were left at the camp sites to rot or to be eaten
by vultures. These encampments began about July 4 and continued
through fall and winter.

During the summer of 1893 or 1894, in Matagorda County, near
Bay City, V. L. LeTulle reports that 71 Attwater's prairie chickens
were shot in 2 hours; and that in 1895, at the site of the present town
of Van Vleck, he and 3 friends killed 72 birds in an afternoon, and
except for poor marksmanship would have bagged many more.
Near Wharton, in Wharton County, in the fall of 1894 or 1895,
LeTulle found 340 piled where hunters had camped.

Mendell Burrell of the Ray Pipkin ranch (Big Hill country,
Jefferson County) told the writer that as late as 1920 his domestic
chickens were fed under the ranch house in winter to prevent prairie
chickens from consuming the grain. In the same area it is said that
flocks of from 150 to 200 Attwater's prairie chickens often alighted
in the introduced chinaberry trees {Melia azedarach vmbraeulifera)
around the ranch houses and fed extensively on the berries. In
verification of this statement W. S. Boothe, State game warden at

ATTWATER'S PRAIRIE CHICKEN 7

Beaumont, Tex., presented the writer with a photograph, taken in
1910 on the White Ranch at Devers, showing a dozen prairie chickens
in a chinaberry tree beside a house.

C. H. Brosig, who hunted Attwater's prairie chickens in the Eagle
Lake area for more than 40 years, reported that the birds once
were so numerous in the sandhill country bordering the Bernard
River that a new covey frequently was flushed while singles from
one previously discovered were being pursued. Paul Mundelius noted
a similar high density in concentrations of prairie chicken^ in the
Sealy-San Felipe section in the eastern part of Austin County in
1873-75. These conditions are seldom found on areas where the pop-
ulation is less than one bird to an acre, and they show the former
abundance of Attwater's prairie chickens in favorable areas. Not
all the coastal bluestem {Andropogon) prairie, about 6,000,000 acres,
wag equally favorable for prairie chickens even under pristine con-
ditions; well-drained, well-populated country (one bird to an acre),
as along the Bernard River, did not aggregate more than 900,000
acres, or about 15 percent of the inhabited range. A little more than
half, 3,300,000 acres, or 55 percent, of the prairie country was only
fairly well drained ; these areas, protected for periods up to 17 years,
now have a maximum population of about one bird to each 10 acres.
Approximately 1,800,000 acres, or 30 percent, was poorly drained;
prohibition of hunting for periods up to 15 years has not produced
a population in excess of one bird to each 50 acres on certain of
these large ranches. Probably, therefore, the former abundance of
Attwater'g prairie chicken in Texas approached, but seldom if ever
attained or exceeded, 1,000,000 individuals, even in peak years.

PRESENT DISTRIBUTION AND NUMBERS

Data on the present status of Attwater's prairie chicken (table 2,
p. 8) were obtained as follows:

At least 90 percent of all farms and ranches known or thought
to have been occupied by the birds in 1937 were visited by the writer
in company with local State game wardens. Unless the warden was
thoroughly familiar with conditions on the various areas, conference^
were held with landowners, managers, cowboys, guides, hunters, or
other persons who were in position to know the status of prairie
chickens on particular tracts. After the conferences a general recon-
naissance of the areas wa^ made by automobile or on horseback, and
notes were taken on the topography, vegetation, surface water, and
soil to ascertain the general suitability of the land for prairie chickens,
as was done in similar studies previously made in the Eagle Lake area.

Then in the light of all available information the range of Att-
water's prairie chicken was mapped in each county, and representa-

8 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

tive sample areas were selected for intensive study. The birds were
then "rope counted" * on the sample tracts. Finally, the total popu-
lation of the area was estimated from the data obtained by counting
the sample plots.

Tarle 2. — Population status of

^he Attivater's prairie chicken in Texas

(1937)

Size of
tract

Census area

Territory occupied
per bird

Popula-
tion of
■whole
tract

County

Size

Popula-
tion (rope
count)

Census
area

Whole
tract

Acres

Acres
1,080
530

Number

56

106

Acres
19.2
5.0

Acres

Numhtr

Total or average (1) '

42,000

6,554
5,000

1,610

162

9.9

4,242

3 25

Goliad ^

34

Total or average (2) -

11,554

f

230
873
218

4
14
0

57.5
62.3

1

Total or average (3)

65, 535

1,321

18

__ 73. 4

892

4,000

4.915
IS, 022
49, 152

4,200
14,250

32
315

131.2
45.2

76, 089

18,450

347

53.1

1,433

736

0

Total or average (5)..

12, 288

736

0

550

f ---

542
364
219

1.57
716

482

4

0
1
3
0
0

135.5

Harris --

\-

219.0
52.3

1

Waller

Total or average (6) -.-

103, 878

2,480

8

310.0

335

f

585
248
269
334

29
2
1
2

20.1
124.0
269.0
167.0

1

Brazoria

1

Total or average (7)

54. 067

1,436

34

42.2

1,281

4,000
700
400

24

1
0

166.6
700.0

Total or average (8)

91, 724

5,100

25

204.0

449

Grand total or average

457, 135

31, 133

594

52.4

8,711

I Counties in which prairie chickens occur but in which counts were not made because of scarcity of birds
or similarity of the areas to adjoining counties.
' Numbers in parentheses in total lines refer to areas correspondingly numbered in the map, fig. 1, p. 3.
' Estimates supplied by game wardens.

* For an account of the method used in rope counting see p. 49.

ATTWATER'S PRAIRIE CHICKEN 9

The known range of Attwater's prairie chicken in Texas, as of
September 1937, aggregated only 457,135 acres, as compared with
approximately 6,000,000 acres in former times. The range has there-
fore been reduced more than 93 percent during the past 100 years or so.
The birds are still to be found in the counties of Aransas, Kefugio,
Goliad, Victoria, Calhoim, Dewitt, Jefferson, Waller, Galveston,
Chambers, Liberty, Lavaca, Wharton, Colorado, Matagorda, Austin,
Fort Bend, Harris, Brazoria, and possibly Jackson and Orange,
although they have not been reported by reliable observers in the
last two for several years. They have definitely been extirpated
from Willacy and Montgomery Counties, and their distribution has
become restricted throughout the State, especially in the counties
of Goliad, DeWitt, Lavaca, Calhoun, Matagorda, Galveston, Fort
Bend, Liberty, Jackson, and Orange, if they occur there at all.
Prairie chickens had not been seen in Goliad County for at least
10 years prior to 1937, at which time four birds were reported on
the W. J. O'Conner ranch.

No more than half the grassland range in any county except
Refugio is occupied by prairie chickens. In the counties of Mata-
gorda, Lavaca, Wharton, Calhoun, Liberty, Jackson, and Fort Bend
even less than 10 percent of apparently satisfactory pasture is inhab-
ited. Roughly, the available range for prairie chickens is only about
30 percent occupied and, with the exception of about 20,000 acres
in southeastern Refugio County, all the occupied area has a sparse
population.

The total number of prairie chickens in coastal Texas in the
summer of 1937 was only about 8,700. The estimated 1937 popula-
tion was probably less than 1 percent of the number believed to
have occupied the coastal prairie in peak years before it was devel-
oped by white men. Approximately 4,200 chickens (or almost 50
percent of the known population of the State) inhabit two ranches
in Refugio and Aransas Counties. The estimated population of
4,500 birds for the remainder of Texas is small indeed !

The consensus is that, during the past 10 years, the number of
prairie chickens has decreased in all coastal counties except Refugio
and Brazoria. In Refugio County there has been a rapid increase
in recent years, probably largely because of excellent protection on the
Salt Creek and Martin O'Conner ranches. Since 1935 the birds have
spread from these onto the O'Brien, Powers, Welder, and Heard
ranches near Greta, Refugio County, and probably also into Goliad
County. Because of protection during a 5-year close season in
Brazoria County (1932-36), R. Z. Cowart, State game warden there
believes that in 1937 the number of birds had reached and possibly
slightly exceeded the 1927 population level.

10 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

The prairie chicken population of the counties of Orange, Liberty,
Fort Bend, Matagorda, Jackson, Lavaca, Calhoun, and Goliad, where
populations of less than 100 birds per county obtain, may already be
reduced to the point where recovery will be extremely difficult or
even impossible. So long as any birds remain, however, every effort
should be made to build up their numbers.

HABITS

COURTSHIP AND MATING

Prairie chickens do not pair for breeding, but are promiscuous.
Males occupy selected courtship stations on booming grounds (see
frontispiece) , which are visited by the females. Copulation may take
place elsewhere, however, in case of chance meetings. To attract the
females, the cocks put on elaborate exhibitions, and their courtship
antics are unbelievably weird. Of special interest is the manner
in which the booming call is rendered.

This call of the male resembles somewhat the sound whur-ru-rrr^
with strong accent on the second syllable. Although it generally
lasts about 5 seconds, the call varies in length and tone. In mid-
season the calls are characteristically deep and full-throated; later
they become shorter and higher pitched, possibly because the males
are then less vigorous. The sound of the booming carries for a
mile or more on quiet days. It has a ventriloquial effect and often
seems farther away or closer than it actually is. During the court-
ship season males boom regularly in early morning (sunrise until
about 8 a. m.) and in late afternoon (5:30 p. m. until sunset), but
calls have been heard at all hours of the day and night. Booming
at night is sporadic, however, even during the peak of the courtship
season in March, being most common when the moon is bright and
when there is little wind.

The appearance of the male, while booming, is striking. As a
preliminary to uttering the call he stretches his neck forward par-
allel to the ground. The erected pinnae, or neck tufts, point for-
ward; the spread tail is held vertically or even inclined slightly
over the back. The wings are extended downward and held firmly
against the body and legs, the primaries almost touching the ground.
The whole body appears strained and rigid. A short run forward is
followed by vigorous stamping with the feet, which lasts only a few
moments, but which under favorable conditions is distinctly audible
for 50 feet or more. Inflation of the air sacs, which are actually
but one sac with two lateral portions (pi. 3) is synchronized with
the stamping. The first syllable of the booming is given before
stamping ends, the male quickly jerking his head downward as he
begins the call and keeping it there until the air sac is deflated.

North American Fauna 57. Fish and Wildlife Service

Plate 3

Male Attwater's prairie chicken, showing vocal sacs. (Photo from Texas Game,
Fish, and Oyster Commission.)

ATTWATER'S PRAIRIE CHICKEN 11

The call of Attwater's prairie chicken apparently is given in the
same manner as was that of the extinct heath-hen, its near relative.
Gross (Bent 1932: 272) describes the booming mechanism of the
heath-hen as follows :

The sacs do not produce the notes, as was thought by some of the earlier
ornithologists, but have much to do with modifying the sounds produced by
the syrinx (the vocal mechanism at the juncture of the bronchial tubes). The
sounds are produced by the air forced from the lungs, which vibrate specialized
membranes of the syrinx under control of a complex set of muscles. The
sound waves then issue through the trachea and glottis to the pharynx. In
the production of such notes as the ordinary cackle the mandibles are opened
and the air accompanied by the sound waves issues out of the mouth. In the
tooting [booming] performance the mandibles are tightly closed, the throat
patch is elevated, and the tongue is forced against the roof of the mouth
(palate) by the mylohyoides muscles, which close off the exit through the
internal nares. The tongue is bent in such a way that it causes the glottis at
the base of the tongue to open directly in front of the esophagus. The air
now coming from the respiratory system is forced to fill the modified anterior
end of the esophagus, or gullet, which becomes distended like a balloon. While
the air sac is filling, the sound waves produced by the syrinx beat against
these tense drumlike membranes, which serve as resonators for the sounds
and give them their great carrying power.

The booming call does not complete the vocabulary of male prairie
chickens at courtship time. The rendition of additional calls, all
distinctly henlike, is described as closely as possible on p. 12. On
windy days cackling sounds, like Nos. 1, 2, and 3 in the list, carry
farther than booming. The call note piooih (No. 14) dominates
when hens visit the courtship grounds. Observers may identify
grounds where females are present by this piuoik call.

Males do not confine their courtship activities to vocalizing, and
fights are common. Opponents usually approach each other, utter-
ing peculiar whining notes, with necks outstretched, ear tufts erected,
tails spread, wings drooped, and air sacs deflated. Then, as if pos-
sessed of the same thought, they suddenly hop off the ground, wings
beating rapidly, and clash in midair. These bouts are usually dis-
continued after three or four flurries, and the victors seem satisfied
after pursuing their opponents for short distances. Many feathers
are frequently lost, but fights seldom if ever end fatally. Males
sometimes engage fancied opponents, as clumps of weeds or tufts
of tall grass, and at other times they joust and bluff for periods up
to 30 minutes or more without striking a blow. With necks out-
stretched, heads held a few inches apart, and wings dangling loosely,
they resemble domestic roosters fighting. At intervals males flutter
into the air to heights of 3 to 5 feet, alighting nearly on the spot
whence they arose. Their surplus energy apparently must be expended

303807°— 41 2

12 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

in one way or another, although activity sometimes lags for brief
periods.

CAIX OF MALE PRAIEEE CHICEIONS ON COXJBTSHIP GROUNDS

1. Ca-ca'-carca-ca-ca-ca-ca-caa-caa—. All except last two notes given rapidly,

2. Ca-ca'-caa, ca-ca'-caa Rapid.

3. Ca-ca-ca-ca'-ca Rapid, accent on second to last syllable.

4. Ca-ca-ca-ca-keece All except last syllable given rapidly.

5. Kmeee, ca-ca-ca-oa-ca-ca First syllable dravpn out, remainder given

rapidly.

6. Kwerr-kwerr-pwah First two syllables drawn out.

7. Kwier-kwier-ku-ier-kicier All syllables drawn out.

8. Pwark or pwarrk Medium rapid. Often preceeds cackling

or booming, carries as far as or farther
than booming notes under certain
weather conditions.

9. Caaa-caaa-caaa-caaaa Slow and drawn out. Sounds almost ex-

actly like protests of a domestic setting
hen that is disturbed.

10. Pwiek, pwark Medium rapid.

11. Pwiek, pwiek, pwiek All notes drawn out with emphasis on

the iek.

12. Pwiek, ca-ca-ca-ca-ca-ca<i Pause after first note, cackle given rapidly.

13. Pwk-pwk-pwk-pwk-pwk-pwk Har.sh notes executed rapidly, but in a

subdued tone.

14. Pwoik, pwoik, pwoik, pwoik Executed rapidly and with much vigor.

These notes predominate all other calls
when a female approaches a courtship
ground.

15. Kliee, kliee, kliee; ca-ca-ca-ca— Kliee's drawn out ; ca's given rapidly. This

is a prominent call in early spring.

16. Kwoo, kiooo; kwah, kwah Rapid. Another prominent early season

call.

The performers do considerable feeding when they first arrive on
the courtship ground, and certain of them feed sporadically throughout
their stay. At other times individuals, sometimes an entire group, sit
or stand in their places and look about. Rest periods terminate
abruptly, however, when a male recognizes a real or fancied challenge,
or when a hen appears.

While the male is bold and noisy during the mating season, the female
is demure and shy. Hens visit the courtship grounds irregularly except
early in March. Even in well-populated territory a week sometimes
elapses before the persistent male performers are rewarded by female
company. When on or near a drumming ground, hens usually appear
little interested in the spirited antics of the obviously excited males.
Sometimes, however, they walk among the contestants and mate with
one or several of them. Hens usually remain at the courtship areas
briefly; usually they stay only a few minutes before leaving to feed
elsewhere or fly to the vicinity of their nests.

ATTWATER'S PRAIRIE CHICKEN 13

In Colorado County signs of the approaching breeding season were
noted early in January (1937) and late in December (1938) when cer-
tain males, probably the most vigorous, pecked at and otherwise bullied
their fellows as the flocks left the roost or fed early each morning. A
few weeks later, early in February (1937), and late in January (1938),
males assembled early in the morning on short-grass areas that later
served as courtship grounds, and fought and maneuvered for choice
positions. For a week or more, calls consisted largely of miscellaneous
squawks and cackles, and fights, though frequent, were of short dura-
tion. Males occasionally attempted to drum or boom, but their notes
lacked midseason depth and vigor. Females, still in winter flocks,
seemed indifferent to the proceedings. It was not until February 12, in
1937, and January 26, in 1938, that booming was commonly heard, and
each year, after 2 to 3 weeks, flocks of females generally broke up and
the courtship season was well under way. During both 1937 and 1938
courtship activity was at its peak in March, continuing through April
and ending on May 20, in 1937, and on May 21, in 1938, when the last
booming calls were heard.

Key areas during the courtship season are the booming grounds where
males assemble each morning from daybreak until about 8 a. m. and
each afternoon from 5:30 p. m. until dark and give their courtship
display.

The preferred booming ground is a short-grass flat, an acre or so in
extent, surrounded by an area of medium to heavy grassy cover suitable
for nesting. Of several hundred sites observed during 3 years (1936-
38) only one was on ground elevated enough to be termed a small knoll.
The others were even with or slightly below the adjacent land surface.
Stoddard (Bent 1932: 245), discussing the greater prairie chicken in
Wisconsin, says that "the 'cooing' ground [courtship ground] at the
sandy west end of Sauk Prairie has been used each spring for over 30
years, the birds always using the same knoll whether in rye, stubble,
or grown to grass." Courtship grounds of Attwater's prairie chicken
do not show the same degree of permanence. Cultivation seemingly re-
sults in immediate eviction, whether the crop is rice, corn, cotton, or
something else. Likewise, the birds do not use fallow fields except where
cultivated land is the only other environmental type available, or where
the fallow land has aged to the extent that its surface and vegetation
are nearly identical with that of nearby grassland. Even those court-
ship grounds that are in pastures may or may not be occupied each year
for a series of years. Of 10 such grounds, on which records were ob-
tained from 1936 through 1938, only 5, or 50 percent, were occupied each
year. Their populations were fairly stable (see table 3, p. 14). Of
the others, 2 were occupied in 1936 and 1937 ; 2 were unused except in
1936 ; and 1 was occupied in 1936, in part of 1937, and throughout the
entire season in 1938. There was little variation in the prairie chicken

14 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

population of the pastures in which these courtship grounds were situ-
ated, but in every instance the vegetation on or near the study areas
varied in density through grazing or burning. Cover changes on and
near courtship areas influenced their attractiveness to the birds, possibly
to the point of determining whether they would be occupied and by
how many individuals.

Table 3. — Occurrences in S years of male prairie chickens on 5 courtship founds

in Color-ado County

Observations

Birds observed

Name of pasture

Extremes

Averages

1936

1937

1938

1936

1937

1938

1936

1937

1938

Thomas . -

2
2
2
2
2

11
8
14
13
20

6
6
6
11
12

7- 9
9-11

6- 9
3-10

4-11
5- 8
3- 8
5-13
6-14

8.0
10.0
5.5
7.0
7.0

7.0
6.0
8.0
9.0
6.0

8.0

Do

7.0

Do

5-6 3-9

7.0

6- 8

5-11

10.0

7- 7

0- 8

10.0

Total or average

10

66

42

7.5

7.2

8.4

NESTING

While the males are still engaged in their courtship performances,
the females quietly select and improve the nest sites and attend to
laying, incubating, and hatching the eggs and rearing the young.
Nests (pi. 4) are made on the ground. Of 19 examined in 1937
and 1938, 17 were in long-grass pastures, 1 in a hay meadow, and
1 in a fallow field. All were in dead grass of the previous year's
growth. Fifteen (about 76 percent) were on or near well-drained
mounds or ridges, and 4 were in poorly drained situations. In a
choice of nest sites, cover appeared of more importance than topog-
raphy and the structure of the soil. Twelve (63 percent) of the
study nests were situated within 10 yards of well-marked trails,
possibly because prairie chickens dislike walking through heavy
matted vegetation when approaching or leaving their nests. Cattle
make many trails, thereby improving nesting areas.

Study nests were found always within a radius of half a mile
to a mile from occupied booming grounds. Sometimes the sites were
rather distant from acceptable feeding territory, although flights of
up to a mile seemed to inconvenience the birds very little.

Nests were merely shallow depressions, about 7 inches in diameter,
lined with bits of dead grass, twigs, and a few feathers, presumably
from the females. All were more or less roofed over because of the
lapping or bending over of surrounding vegetation. Entrances faced
in various directions with no preference shown. There was consid-
erable variation in degree of concealment (pi. 5), at least according
to human standards; 5 nests being excellentl}^, 10 well, and 4 poorly,
concealed. Kapid new plant growth in April and May aided ma-

North American Fauna 57, Fish and Wildlife Service

,^ o

o O

■o3 a
bi o

^ o
o3 ;r

5

North American Fauna 57. Fish and Wildlife Service

Plate 5

^r

:r ^-/i*^.

C3 ii

;^ o S
5 ^ 00^

^O

ATTWATER'S PRAIRIE CHICKEN 15

terially in hiding nests ; consequently, some nests that were relatively
exposed when found were well hidden later. Wild indigo {Baptisia
sp.), vetch {Vicia litdoviciana) , phlox ( Phlox drummondii) ^ peren-
nial ragweed {Ambrosia psilostax^hya) ^ dogfennel {Eupatorium sp.),
and panic grass {Panicum sp.) were green plants that aided con-
cealment materially by mid-May. In Colorado County, favorite
nesting materials were dry bluestem grass {Andropogon scoparius),
and paspalum {Paspalum dilatatuTn).

The earliest date for a nest containing eggs was reported by Wad-
dell near Egypt, Wharton County, February 25, 1925; the latest
record is that of a nest in Colorado County in which the clutch
was completed May 29, 1938. In both 1937 and 1938, however, the
peak of the laying season in Colorado County was late March and
early April. Hens always laid in the morning, usually from 7:30
to 9, flying to the vicinity of their nests when ready. After cautiously
looking about or feeding a bit longer, hens walked to the nests and
remained there for from about 20 minutes to an hour. The laying
completed, they regularly walked about 20 feet from the nest, scanned
the landscape, and flew away. Since incomplete clutches were un-
guarded except during about an hour each day, they were especially
vulnerable to natural enemies.

Hens under observation normally laid an Qgg a day until the
clutch of 8 to 15 was complete, but sometimes they failed to lay for
periods of 1 to 3 days. Clutches usually contained 12 eggs, and lay-
ing was generally completed in about 2 weeks. The period of egg
laying was sometimes extended, however, when nests were destroyed.
Three hens, each the only resident on a small unburned plot, re-
nested during 1937, one of them three times.

New nests, however, were placed 5 to 20 yards from old ones, and
were less effectively concealed. Destructive agents had even greater
opportunities to take the later nests, as they did in four out of five
cases. Since booming ended by mid-May, the period for mating
was short. Late broods were invariably smaller than early ones,
probably because late clutches were small, their hatchability low, or
their mortality heavy. A successful season depends largely on the
fate of early nests, so that a primary objective of management should
be to safeguard these attempts.

Twenty-nine eggs of Attwater's prairie chicken measured by Bent
(1932: 264) averaged 42.3 by 31.5 millimeters in size, showing ex-
tremes of 44.9 by 32, 42.4 by 33.5, 38.8 by 28.9, and 39.8 by 28.6
millimeters. Newly laid eggs were dull cream or bluish buff in
color, some of them minutely specked with red. During incubation
the color of the eggs became dull and the shells shiny. Incubation
began at from 1 day before until 4 days after the last egg was laid.

16 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

Except for two feeding and resting periods daily, extending from
about 7 to 8 a. m. and from 5 :30 to 6 :30 p. m., hens incubated con-
stantly. When incubation was advanced, morning feeding was fre-
quently dispensed with. Two clutches pipped approximately 23 and
24 days after setting began, in each instance requiring about 48 hours
longer to hatch. Of 71 eggs in 7 nests, only 3 (about 4 percent)
were infertile, 66 hatching successfully. Seemingly, fertility and
hatchability are high under favorable conditions.

The hatching period was evidently a time of danger. Chicks
peeped incessantly and scrambled in and out of the nest. Nests
emitted strong odors, apparent even to man. Females at hatching
time appeared nervous and shifted their positions frequently. Unless
disturbed, however, they did not leave until the last egg had hatched,
after which they deserted the nests. In 1937 a nest in which all
young were hatched by 11 : 50 a. m. on May 15 was vacated by 3 p. m.
that day, and a brood that was hatching at 8 : 30 a. m. on June 2 was
gone 24 hours later. One hen left before 2 pipped eggs were hatched
and before the natal down on some of the young was dry, probably
because fire ants [Solenopsis) had entered the nest.

GROWTH AND DEVELOPMENT OF YOUNG

When leading chicks from the nest, old bird^ traveled through
the lightest cover or followed trails, probably because heavy matted
vegetation impeded progress and increased the chance of chicks get-
ting lost. Cow trails were favorite travel ways. Chicks ranged in
front, behind, and on both sides of the hen over an area 1 to 6 yards
in radius. Interruptions for sporadic feeding and for frequent
brooding, which was probably more necessary for assembling than
for warming the young, made progress slow. Hens with chicks less
than 10 days old (pi. 6) seemed mainly concerned with watchfulness
and brooding. Occasionally they caught available insects or nipped
off a few green leaves or bud^, but they did little continuous feeding.
When danger threatened, they gave a warning call, best described
as a low kwerr^ hwerr^ krcerr, and slowly gkulked through the grass
with head lowered and wings dangling loosely, almost touching the
sod. Young birds "froze" with their bodies closely pressed to the
ground. Decoy efforts of adult females were never so energetic as
those of bobwhites under similar circumstances. When hens were
flushed, the chicks in hiding (pi. 6) became impatient after 3 to 5
minutes, and peeped and ran about in spite of the fact that the object
of suspicion remained. After the immature birds of 2 to 3 weeks
of age could fly fairly well, females accompanying them did not
decoy, but always flushed freely, the young doing likewise.

Chicks that were less than a week old were brooded quite often,
probably in all for about 50 percent of the daylight period. Ten birds

North American Fauna 57. Fish and Wildlife Service

Plate 6

Above, Chicks about 10 days old; Colorado County, Tex., approximately 6 miles
north of Eagle Lake, May 3, 1938. Beloiv, Chicks hiding; Colorado County,
Tex., approximately 6)4 miles north of Eagle Lake, May 3, 1938. (Photos
from Texas Game, Fish, and Oyster Commission; E. P. Haddon.)

ATTWATER'S PRAIRIE CHICKEN 17

about 2 days old were brooded 42 minutes out of the II/2 hours during
which they were watched on May 4, 1937. Their position during
brooding was interesting. Hens squatted low with legs at an angle
of a.bout 30 degrees with the ground. Wings were drooped and
feathers of the underparts were ruffled. Some chicks scrambled up
and under the wings. When the brooding hens were frightened and
suddenly stood erect, usually only two or three chicks were visible;
the others, however, tumbled from beneath her wings a few moments
later. As the chicks grew older, the time spent brooding diminished;
those 2 weeks old apparently were brooded little except early in the
morning, during inclement weather, and at night.

JsTewly hatched chicks are about the gize of day-old bantams, and
their coloration is nearly identical with that of young bronze turkeys.
The basic, buffy yellow is streaked with gray on head and upper-
parts. Growth and development are rapid. Wing feathers appear
almost at once; week-old chicks have primaries approximately five-
eighths of an inch long. Chicks fly when 2 weeks old. Except for
differences in the length of the tail and legs, they are about the size
of English sparroWiS. When 3 weeks of age, youngsters are almost
as large as starlings and can make sustained flights of 40 yards or
more. At 4 or 5 weeks, young birds approximate the size of mature
bobwhites, and often fly a hundred j^ards before alighting. When
6 or 7 weeks old, the young are about half grown and at 8 or 9 weeks
they are three-fourths the size of adultjs. Youngsters 10 to 12 weeks
old can scarcely be differentiated from the old birds in the field.
Weight evidently does not increase as rapidly as size, however, for
two birds approximately 3 months old were more than a pound lighter
than mature individuals.

As young prairie chickens grow in size, all cannot, of course, main-
tain a brooding position under the sheltering body of the mother.
Usually by the time they are about 3 weeks old some are forced out-
side; there they sleep with bodies pressed closely to that of the hen.
When 4 to 5 weeks of age, two or three chicks sometimes crowd under
their mother, but the remainder roost from a few inches to about 2 feet
away. At 6 to 7 weeks, young birds adopt the roosting formation of
adults. Flocks of Attwater's prairie chickens sleep about a foot or so
apart, the individuals facing in different directions. Roosting spots
vary in size from 1 to 3 square yards or more, depending on the number
of birds in the group. The number of scats left at a roosting site
is not an absolutely accurate index to the number of birds in a flock,
because slight shifting of individuals during the night brings about the
deposition of more than one pile by a bird.

Chicks about 2 weeks old take vigorous dust baths, a habit that is
indulged in regularly throughout life when dry, powdery material is

18 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

available. Prairie chickens generally dust during the midday rest
period that extends from about 10 a. m. to 4 p. m. Old pocket gopher
mounds and cattle wallows are favorite places. Some birds use the
same dust bath more than once.

BROOD SIZE

The size of the brood generally decreases with the age of the young.
Of 48 broods on which accurate counts were kept (table 4, below) 6
from 1 to 3 days old contained 64 young, averaging 10.6 birds each.
Three broods estimated to be 5 to 10 days old contained only 14 chicks
averaging 4.6 each. Four broods 15 to 27 days old had 22 young, or
an average of 5.5. Fifteen families over 4 and under 6 weeks of age
aggregated 80 young and averaged 5.3. Twenty groups over 6 weeks
totaled 80 young and averaged 4 each. The average size (5.3 young)
of 15 families, estimated to be over 4 but under 6 weeks of age, was
exactly half the average size (10.6) of 6 new broods. The average
size (5.3) of 15 families over 4 but under 6 weeks old was but slightly
larger than the average size (4.0) of 20 families older than 6 weeks.
Therefore, it appears that juvenile mortality is heaviest during the
first 4 weeks and comparatively light thereafter.

Table 4. — Size of 'broods and number of chicks counted during May,

July

June, and

County

Date

Family

groups

observed

Chicks

per
group

Chicks per
average
group

May 4
May 15
May 18
May 23
May 28
May 29
May 31

1
1
1
3
3
1
2

10

12

12

2,4,2

7, 9, 11

8

4,8

10.00

Do —

12.00

Do

12.00

Do

2.66

Do

9.00

Do

8.00

Do - - -

6.00

Total or average for May

12

89

7.41

June 2
June 3
June 8
June 10
June 24

4
1
2
2
3

6, 9, 4, 2

10

3,3

4,8

5,4,3

6.33

Do

10.00

Do

3.00

Do

6.00

4.00

Total or aVTS^ge foJ" .Tutia

12

61

5.08

July 1
July 8
July 14
July 17
July 19
July 26

3
2
1

4
6
2

5,5,6
3,3

7

10, 8, 1, 3

10, 2, 2, 4, 5, 2

10,5

5.33

3.00

7.00

6.00

Do

4.16

Colorado - --

8.00

18

92

5. U

Aug. 12
Sept. 2
Sept. 3
Sept. 4

1
3

1
1

3

5,3,3

3

4

3.00

3.66

Do

3.00

Do

4.00

5

18

3.60

48

263

5.48

ATTWATER'S PRAIRIE CHICKEN 19

JUVENILE MORTALITY

Although much remains to be learned about juvenile mortality,
its causes were fairly well established in some instances. Chicks are
sometimes trapped and drowned in rice fields at the time of flooding
(usually about May 10). In 1925, near Egypt, Wharton County,
Waddell and others picked up hatfuls of chicks and carried them
beyond the levees. In 1937 a farmer near Eagle Lake similarly res-
cued a brood. The number of chicks annually saved from this
hazard, however, is probably insignificant in contrast to those lost.
Prairie fires kill young and, as stated on page 42, burning is still
common in certain areas at nesting and brooding time. Unfavorable
weather, especially rains (pp. 32 to 35) and natural enemies (p. 39),
account for the death of some young birds, but no small percentage
of these may be chicks that are lost.

Chicks stray from the brood more often than one might suspect.
During April, May, and June, 1937, no fewer than 13 strays were
seen, all under 4 weeks old. Usually they occurred as singles, but
sometimes in pairs and trios. How the youngsters became lost, of
course, was usually unknown, but several reasons were apparent.
The characteristic loose feeding formation of broods possibly con-
tributed to straying; also, broods usually scattered widely and flew
far when disturbed; and, possibly most significant of all, adults did
not appear to have a highly developed rallying call that doubtless
would be of assistance in reassembling youngsters.

Lost chicks evidently join other gi'oups occasionally, as hens ac-
companied by young of varied sizes were several times noted in 1937.
Once two chicks, about 2 and 3 weeks old, respectively, were seen
with two molting males. Higher population levels might increase
the frequency of adoptions.

FAMILY DISINTEGRATION

Many young Attwater's prairie chickens 6 to 8 weeks old leave the
family groups and take up life on their own, but, as is true with
domestic chickens, all young do not leave the hen at the same time;
disintegration of the family group is gradual. Some young remain
with the hen well into the fall. Unattached young, 6 weeks of age
or older, as distinguished from lost chicks less than a month old,
became noticeable late in June and they were frequently seen after
July. Family disintegration after 6 weeks or thereabouts is normal.
Young prairie chickens at that age seem as capable of foraging and
resisting adverse weather as are the adults.

20 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

ANNUAL INCREASE

The actual survival of young prairie chickens (table 5, below) prob-
ably is always well below the potential yield (12 or so young from
each hen) even when favorable weather conditions obtain during
the critical breeding season.

Table 5

. — Young and adult birds observed

in census after June SO, 1931

Date

Area

Adults

Young

County

Males

Females

Sex un-
known

Total

Families

Strays

Total

Brazoria

Victoria

July 1
July 14
July 17
July 19
July 26
July 27
Aug. 12
Sept. 1

Acres
585
921

1,080
530

1,450
851

1,282

2,000

Number
2
1
1
3
0
2
1
1

Number
4
1
4
8
2
0
1
5

Number
2
6
23
53
4
2
2
8

Number
8
8
28
64
6
4
4
14

Number

5.5,6

7

10, 8, 1, 3

10, 2, 2, 4, 5, 1

11,5

3

3

5,3,3

Number

1,1,1,1,1

1,2

2,1,1,1,1

1, 1, 2, 1, 1, 1

2.3

1

1

1, 2, 1. 3

Number
21
10

Refugio

28

Do

32

Colorado.

Brazoria

Harris

21

4
4

Colorado

18

Total

8,699

11

25

100

136

103

35

138

In a rope count on or after July 1, 1937, 138 young as compared
with 136 adults (about 1:1) were enumerated on an area of more
than 8,699 acres. At this time most of the counted birds were 4
weeks old or more and were beyond the age when mortality is thought
to be most severe. The increase on the counted areas, then, was only
about 100 percent in spite of the fact that food, cover, and weather
conditions were favorable. A 100 percent increase of prairie
chickens in any one year is very good, and the arguments for long
open seasons and large bag limits, based on the potential annual
increase (12 chicks for each hen, or about 600 percent a year), evi-
dently are fallacious.

FLOCKING

Late in summer and early in fall, the prairie chickens displayed no
marked tendency to combine into stable groups. In August and
September of 1936 and 1937 well over half the birds observed in
Colorado County were recorded as singles, pairs, and trios, although
small groups of 4 to 6 were not uncommon. Occasional larger flocks
were recorded, but these appeared to be temporary. In Colorado
County, at about noon on September 1, 1937, a flock of 15 to 25 birds
was noticed in a cotton field. On the following 3 days at the same
hour 11, 15, and 9 birds, respectively, were present in that field;
but they had come in between 9 : 30 and 11 : 30 as singles, pairs, and
in small groups not exceeding 5 birds each. Between 4 p. m. and
dark they left the field as they had come. Again in Colorado
County, at 6 p. m. on September 3, 1937, another group of 16 birds
found in a pasture came together as follows ; At 5 : 40 p. m. a group

AtTWATER'S PRAIRIE CHICKEN 21

of 8 flushed approximately 1 mile from the spot where the large
flock was later noted, and as they flew over the prairie, a pair, a
single, a trio, and another pair joined the original group. All set-
tled and fed together for a time, but the bevy disintegrated by dusk.
Like instances suggest that early fall flocks of a dozen or so birds
are unstable groups brought together largely by chance.

In fall, after the weather turned cool, groups of prairie chickens
became the rule rather than the exception. Early November bevies
generally contained 4 to 12 birds each, but large flocks became in-
creasingly common from about December 1 to the onset of the breed-
ing season. Late in winter (January) Guy Ferguson, State game
warden, Sinton, Tex., observed flocks in Refugio and Aransas Coun-
ties that contained more than a hundred birds. Wardens Waddell
in Colorado, Austin, and Wharton Counties, and McClosky in Vic-
toria County, reported winter aggregations of about the same size.
In 1936, J. O. Linney, foreman, Salt Creek (Hallahan) ranch, Re-
fugio and Aransas Counties, noticed late winter concentrations esti-
mated to contain 250 to 300 individuals. The writer has not observed
such large winter flocks, possibly because he has not made observa-
tions in areas where the birds were suflficiently numerous. January
assemblages of 25 to 35 birds were not uncommon, however, in Colo-
rado County. Despite the fact that large flocks became more fre-
quent from November until the breeding season, small groups of 8
or fewer birds or singles were always to be found. All packs ob-
served in Colorado County late in November, December and January
contained birds of only one sex. Late in January, residents of the
coastal country eagerly listen for the first booming calls, which,
besides promoting rapid disintegration of winter flocks, signal the
departure of winter and the coming of spring.

seasonal movements
Spring

Comprehensive data on prairie chicken movements are lacking, but
the records obtained in 1937 are of interest. Two broods that were
observed two or more times daily from the time they were hatched
until they were 7 and 12 days old, respectively, were, at last observa-
tion, less than half a mile from the nest sites. Another brood, esti-
mated to be 8 days old when first discovered on June 2, was within 150
yards of the same spot at various hours during the next 6 days. A
fourth brood, about 3 days old on April 29, remained within 400 yards
of a certain windmill from April 29 through May 31. A 640-acre
pasture that contained four broods, all under 2 weeks of age when
rope counted on June 2, likewise harbored four broods 10 days later.

22 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

If this is indicative, the prairie chickens under observation spent
their first several weeks of life in close proximity to the places where
hatched. The daily cruising radius of a brood apparently was small,
seemingly less than 300 yards in the case of birds under 4 weeks old
in a favorable environment. Some 30 broods observed between May
1 and June 1 were found in light cover on or near hardpan flats and
recent burns, indicating a distinct preference for those types of habitat.

Summer

An extensive movement involving both young and adult prairie
chickens in Colorado County began about June 1, 1937, when many
of the young were 3 to 5 weeks old, and lasted until about June
30. The sudden scarcity of the birds in places where they had been
common only a few days before was striking. A 1,000-acre pasture
that contained 37 individuals (16 old and 21 young) on June 2, held
only 16 in all on June 10. As the prairie chickens decreased in some
pastures, they increased in others. A 460-acre pasture that was un-
occupied on May 1 contained 14 birds on June 8 and 23 on July 26.
This movement from the spring range was by stages. One brood
that was watched closely made trips of approximately 1 mile, three-
fourths mile, and 1^ miles in 6 days from June 2 through June 8.
After the first major movement, this family remained for 3 days in an
area less than 500 yards in diameter ; their droppings in piles formed a
triangle with sides of 5, 15, and 17 yards, respectively. The move-
ment of a combined brood of 3 hens and 16 to 25 young are recorded
in figure 2, p. 23.
Leopold (1933: 291) reports that-
All observers unanimously and independently report a strong tendency for the
grown young of most species of grouse to seek the vicinity of drinking water
in late summer and fall, but whether they do this out of choice or necessity
is not known.

The early summer movement of young and adult Attwater's
prairie chickens also was to the vicinity of surface water, but it
was to water near which there also was shade. Pastures having an
abundance of surface water but little or no shade-producing cover
had few if any birds after mid-June. Likewise, places in which
dense stands of weeds, shrubs, or tall grass were abundant, but sur-
face WEiter scarce, were sparsely populated. More than 95 percent of
the more than 500 Attwater's prairie chickens observed from June
24 through September 4, 1937, were in heavy cover within a mile,
generally within less than half a mile of surface water.

The beginning of the summer movement is synchronous with the
drying up of the wild indigo {Baptisia^ pi. 7), a plant that fur-
nishes the principal shade on burns and heavily grazed areas from
April through May. Prairie chickens require abundant shade in sum-

North American Fauna 57. Fish and Wildlife Service

PLATE 7

B5068S: B49718A

Above, Wild indigo (Baptisia) in a closely grazed pasture; Austin County, Tex.,
approximately 8 miles southeast of Sealy, April 10, 1938. Below, Shocked
grain and waste in rice fields sometimes attract prairie chickens; Colorado
County, Tex., 3 miles northeast of Eagle Lake, September 5, 1936. (Photos
by V. W. Lehmann.)

ATTWATER'S PRAIRIE CHICKEN

23

mer, for birds that were herded from such cover at midday panted
vigorously, drooped their wings, and showed other signs of discomfort.
They evidently rarely drink from surface water. In 1937, birds
near water were closely observed, but only one was seen to drink
during the entire summer. That was in Colorado County on June
1, 1937, when a chick about 3 weeks old drank a few times from a
puddle formed by water from a leaking windmill. The soft mud
bordering ponds in inhabited prairie chicken range in Brazoria,
Colorado, and Austin Counties was examined thoroughly at various
times, but tracks of this species were never found. Grasshoppers

FiGxmE 2. — Movements of a combined brood, May 1 to July 26, 1937, Colorado
Comity. Birds seen in areas as follows : 1, May 1 to 28 ; 2, June 10 to July
10; 3, July 15 to 26.

and other favorite foods were frequently more abundant in summer
in heavy cover near water, but the food factor was not thought to
be of great importance at the time. The summer movements of
prairie chickens to heavy cover near water are not satisfactorily ex-
plainable on the basis of cover, water, and food, but these habitat
conditions must be provided where stable populations are desired.
After they found a satisfactory summer range, the prairie chickens
moved little until fall, unless their summer territory was depleted
or that nearby was more suitable. The population of a 460-acre
pasture in Colorado County remained at nearly the same level (25

24 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

to 36 birds) from July 26 through September 3, 1937. The birds
that rested in this area at noon each day, however, foraged and
roosted largely in adjoining pastures, but their range was never more
than 114 miles and usually under half a mile.

Rains in 1937 sufficiently heavy to collect in low spots caused tem-
porary spreading of prairie chickens from previously occupied areas.
As the temporary water disappeared, however, the birds again con-
centrated.

Faix and Winter

Populations that had been rather stable in certain pastures in Colo-
rado County during the summer months of 1937 began shifting early
in fall. About 25 birds that were summer residents of the M. Shin-
dler cotton field from July through August were absent after Sep-
tember 4. Two thousand acres of regularly censused pasture where
prairie chickens were common in summer contained only 9 birds when
rope counted on October 22. While the birds decreased generally
in the large pastures, they increased around small farms near Sealy,
Austin County ; Lissie, Wharton County ; and Bernardo and Chester-
ville, Colorado County — ^territory 5 to 10 miles removed from the
pastures in which birds had been most common during the preceding
spring and summer.

Distances traveled daily were evidently great in some instances. A
bird killed by a farmer at 8 a. m. on September 1, 1936, was known
to have traveled at least 3 miles since dawn, because its crop was
filled with rice and the nearest rice field was that distant. Two in-
dividuals, observed for 2 hours on the afternoon of October 22, 1937,
traveled approximately 1% miles southeast of the point where first
seen. When finally flushed, they flew an additional 2 miles or so in
the same direction. A flock of four birds observed from 4 p. m. to
6:15 p. m. on January 4, 1938, traveled more than 1% miles. The
movement was in a circular direction, however, for at nightfall, the
birds were less than half a mile from the point where they were first
observed. Cool weather, fall rains, and a seasonal abundance of
food and cover, especially in the vicinity of farming commrmities,
probably were important in promoting the general fall scattering
and the long daily trips the prairie chickens made in territory that
was sparsely populated at other seasons. The birds reconcentrated
in large pastures, however, as fall passed into winter.

The population of the Everett pasture (640 acres), Colorado
County, increased from November 3, 1937, through January 28, 1938 ;
five censuses during that period showing 46, 58, 56 to 58, 73, and 84
birds, respectively. Excellent food and cover conditions prevailed,
for the area was lightly grazed. This increase in the number of
birds apparently resulted from influxes from adjoining areas. After

ATTWATER'S PRAIRIE CHICKEN 25

November 15, few fluctuations in numbers between different pastures
were reported by fence riders and others. Pastures that contained
the greatest numbers in November and December 1937 also harbored
the largest breeding populations the following spring.

The data at hand show that late in fall, probably by about Novem-
ber 15, the prairie chickens move to pastures where food and cover
conditions are adequate. Having found such an area, they remain
until spring. Probably the best way to attract a good breeding
population, therefore, is to provide suitable food and cover conditions
during the preceding winter.

FOOD

Data on the food of Attwater's prairie chicken were derived mainly
from analyses of 21 stomachs (crops, or gizzards, or both) and more
than 200 droppings (scats). Additional information was obtained
by watching feeding birds at close range through field glasses. Of
the 21 stomachs, 18 were of adult prairie chickens, 2 of chicks ap-
proximately 10 days old, and 1 of a juvenile about 7 weeks old.
Specimens were obtained as follows: 6 in winter (January and
February), 6 in spring (April and May), 5 in summer (June through
August), and 5 in fall (September through November). J. H. Gaut
collected 3 stomachs near East Bernard, Wharton County, in May
1905. Over the period beginning September 1936 and ending Au-
gust 1938, 2 stomachs were obtained in Austin County, 4 in Eefugio
County, and 12 in Colorado County. As the crops and gizzards
of all birds found killed by automobiles, predators, poachers, and
from other causes were saved, it was necessary to collect only 11
specimens to balance the series according to seasons.

Except during the breeding season, adult prairie chickens regu-
larly feed twice daily, early in the morning (dawn to about 8 a. m.),
and late in the afternoon (4 p. m. to dark) . Occasional bits of food
are picked up throughout the day, but the gullets of specimens col-
lected about noon are usually empty or nearly so. The food capacity
of prairie chickens is large. Gullets frequently contain about 20
cubic centimeters, and the gizzard about 30 cubic centimeters, of
material. Since the birds ordinarily feed slowly and deliberately,
apparently selecting their food with great care, it is not surprising
that their diet in favorable areas is varied. Stomachs have been
examined that contained 29 kinds of food and more than 1,300 items ;
stomachs rarely contain less than 13 kinds of food of 500 items.
Mature birds evidently feed mostly on vegetation at all seasons,
for the stomachs of 18 adults (table 6, p. 26) contained 88.28 percent
of plant material and 11.72 percent of insects. Animal matter prob-
ably ranks higher than plants in the diet of young birds, however,

26 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

for the stomachs of 2 chicks and a juvenile (table 7, below) contained
88.5 percent of insects. The ratio of plant to animal food varies
according to season (table 8, below), insects, for instance, apparently
being eaten in greatest quantity in summer.

Table 6. — Composition of the stomach ^ contents of 18 adult prairie chickens

Item

Stomachs

in which

found

Quantity
present

Item

Stomachs

in which

found

Quantity
present

Number

18

12

16

5

6

Percent

88. S8

27. n

55.67

1.30

4.20

Animal matter -

Number
18

17
8
1
1

11

Percent

11.711

Leaves and stems

Seeds and pods

Buds and flowers

Miscellaneous *

Insects:

Adults

10.83

Eggs and larvae. -

Round worms

Prairie chicken feathers...
Grit

0.89
W
(')
(»)

' Crops or gizzards, or both.

» Woody pod septa, root stocks, and the like.

s Trace.

Table 7. — Composition of the stomach contents of three younff prairie chickens

Item

Stomachs

in which

found

Quantity
present

Item

Stomachs

in which

found

Quantity
present

Vegetable matter

Number
2
2
1

Percent
11.5
1.5
10.0

Animal matter

Nuviber
3

3
2

Percent

88.6

Seeds or pods ...

Insects:

Adults

86.0

Eggs or larvae

2.5

Table 8. — Percentage of plant and animal food according to season

Item

Spring

Summer

FaU

Winter

Whole year

Plants -

94.25
5.75

71.0
29.0

85.8
14.2

95.0
6.0

86.51

13.49

Total . .

100. 00

100.0

100.0

100.0

100.00

The data at hand indicate that Attwater's prairie chickens are
preponderantly granivorous, for seeds and seed pods made up slightly
more than 50 percent of all the material in the stomachs of 18 adults.
Much succulent vegetation is eaten, however, including leaves, buds,
flowers, and root stocks. The birds also consume insect eggs, larvae,
and adults, as shown in tables 6 and 7.

Parts of some 50 kinds of plants and more than 65 species of
insects were identified in the food from stomachs or scats, or by
observations in the field. The names of these plants and insects
together with the seasons when they are known to be eaten, are
listed in tables 9 and 10, pp. 27 and 28.

ATTWATER'S PRAIRIE CHICKEN
Table 9. — Plant foods (56) of Attioatefs praiiHe chicken

27

Plant

Marsileaceao: Pepperwort (MarsUea)...

Poaceae:

Paspalum (Paspatum citiatifotium type) _ .

Bull grass {Paspaium boscianum ?)

Paspalum {Paspaium plicatulum)

Panic grass (Panicum scribnerianum)

Panic grass (Panicum spp.)

Crabgrass (Digitaria)

Sandbur (Cenchrus)

Hegari (Sorghum vulgare)

Rice (Oryza sativa)

Corn (Zea mags)

Cyperaceae:

Beakrush (Rynchospora)

Sedge (Carex)

Commeliuaceae: Dayflower (Commelina cris-
pa).

Alliaceae: Wild onion (Nothoscordum bivalve) . _ _

Liliaceae: (Undetermined)

Leucojaceae: Stargrass (Hypoxis)

Convallariaceae: Solomons seal (Polygonatum
commutatum) .

Ixiaceae: Blue-eyed-grass (Sisyrinchium va-
rians).

Polygonaceae: Dock (Rumex near crispus type),

Ranunculaceae: Buttercup (Ranunculus near
hispidus).

Rosaceae: Dewberry (Rubus)..

Malaceae: Chokeberry (Pyrus)

Mimosaceae:

Sensitive briar (Neptunia lutea).-

Mimosa (Mimosa)

Acacia (Acacia)

Cassiaceae: Partridge-pea (Chamaecrista fas-
ciculata) .

Fabaceae:

Wild pea (undet.)

Wild pea (Lathyrus pusillus)

Peanuts (Arachis hypogaea)

Oxalidacoae: Woodsorrel (Oxalis)-.

Euphorbiaceae:

Doveweed (Crofon capitatus).

Doveweed (C. glandulosus)

Doveweed (C. monanthogynui)

Spurge (Euphorbia)

Spurge (Crotonopsis linearis)

Spurge (Chamaesyce)

Vitaceae: Grape (Vitis)

Malvaceae: Mallow (Malva).

Epilobiaceae: Oaura (Oaura)

Ammiaceae (Cynosciadium)

Convolvulaceae:

Bindweed (Convolvulus) _

Evolvulus - - --

Polemoniaceae: Phlox (Phlox drummondi)

Boraginaceae: Gromwell (Lithospermum)

Verbenaceae:

Fog fruit (Phyla nodiflora)

Verbena (Verbena)

Aeanthaceae: Ruellia (Ruellia ciliosa var.
humilis).

Rubiaceae:

Buttonweed (Diodia teres)

Buttonweed (Diodia virginiana)

Bedstraw (Galium)

Ambrosiaceae:

Marsh-elder (Iva ciliata)

Perennial ragweed (Ambrosia psilostachy a).

Carduaceae:

Thistle (Carduus)

Tickweed (Coreopsis)

Cichoriaceae: (Serinea oppositifolia)

Parts eaten

Leaves.

Leaves, seeds.

Seeds.

..._do

do.

do

.do.
.do.
.do.
.do.
.do.

do

Seeds, leaves..
Leaves, seeds.

Leaves, flowers.

Seeds

Leaves, seeds...
Seeds

Seeds, pods.

Seeds ,

Leaves, seeds, pods.

Seeds, fruits

Flowers, fruits.

Leaves, flowers, seeds.

Leaves, seeds

Seeds

Seeds, flowers

Flowers

Leaves...

Fruits...

Leaves, seeds.

Seeds

do

Seeds, leaves.

Seeds

do

do

Seeds, fruits.
Seeds, pods..

Pods..

Leaves.

Seeds

Seeds, pods

Seeds, pods, flowers.
Seeds

Leaves, flowers, fruit...
Leaves.-

Leaves, stems, seeds
buds, pods, flowers.

Seeds

do_

Leaves, seeds.

do.

Seeds. .

do

Flowers

Seeds, pods.

Seasons when
eaten i

Wi

Sp, su, au, wi

Au

Sp, su, au

Sp.

Sp

Au

Wi

Su, au

Au, wi

Sp...

Su, au, wi

Au

Au....

Wi_

Su, wi..

Sp-

Su...

Sp, su

Wi...

Wi,sp-.

Sp-..

Sp

Au, wi, sp

Su, au, wi

Wi

Su, au..

Sp-.

Wi..

Au

Wi, sp, su

Au, wi

Au, wi

Au

Au, wi, sp

Au

Au

Su

Sp

Sp

Wi

Sp

Sp

Sp, su

Su

Sp, wi, au

Wi...

Au, wi..

Su, au, wi

Su

Sp, su, au

Wi..

Au, vri

Sp.

Sp

Sp, su

Source of
data'

St.

St, sc.

St.

St.

St.

St.

St.

St.

Si.

Si, St.

St.

St.
St.
St.

St.
St.

St, si.
St.

St.

St.
St.

St.
St, si.

St, si.
St, sc.
St.
St, si.

St.
St.
Si.
St, sc, si.

St, sc, si.

St.

St, si.

St, sc.

St.

St.

St.

St.

St.

St.

St.

St.
St, sc.

St.

St, si,

St.

St, sc, si.

St, se.
St.
St, sc.

St, si.
St, sc.

St.
Si.
St, sc.

1 Abbreviations of seasons: Sp, spring; Su, summer; Au, autumn; and Wi, winter.

2 Abbreviations of sources: St, stomach examination; Sc, scat examination; and Si, sight record.

303807'

28 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE
Tasle 10. — Some insect foods (68) of Attwater's prairie chicken

Name

Araneida:

Spider (Lycosidae)__-

Spider (undetermined)

Orthoptera;

Grasshopper (Cyrtaeanthacrinae)

Pigmy locust (Aerydinae)

Grasshopper (Syrbula)

Grasshopper (Oedipodinae) --.

Western grasshopper (Melanoplus ciner-

eus).
Green grasshopper (Chortophaga viridi-
fasciata) .

Grasshopper (Oedipodinae)

Grasshopper (Schistocerca americana)

Grasshopper (SaUatoria)

Meadow grasshopper (Conocephalus)

Long-horned grasshopper (Tettigoniidae) .

Hemiptera:

Shield bug (Pentatomidae)

Bug (undetermined Hemiptera)

Stinkbug (Euschisfus)

Homoptera:

Leafhopper (Cicadellidae)--. —

Lantern fly (Fulgoridae). — .

Soft scale (Leucanium)

Coleoptera:

Weevil {Graphorhinus vadosus), -..

Weevil {Lixus)

Weevil (Thecesternus humeralis)-.

Billbug (Sphenophorus minimus)

Billbug {Sphenophorus bartramiae)

Billbug {Sphenophorus germari)

Billbug {Sphenophorus)

Weevil {Boris)

Weevil {Hyperodes)

Rice-water weevil {Lissorhoptrus simplex).

Weevil {Pachyphanes)

Weevil {Anthonomus fulvus)

Snout beetle (Curculionidae)

. Weevil {Apion)

Scarred snout beetle ( Tanymecus lacaena) _
Scarred snout beetle {Eudiagogus pulcher)
Scarred snout beetle {Compsus auricepha-

Leaf beetle {Phaedon viridis)...

Leaf beetle {Cryptocephalus venustui)

Leaf beetle ( Cryptocephalus)

Leaf beetle {Zygogramma disrupta)

Leaf beetle (Oedionychis petaurista)

Leaf beetle {Metacroma ustum)

Leaf beetle {Disonycha)

Leaf beetle (Chrysomelidae)

Leaf beetle {Calligrapha similis)

Leaf beetle {Graphops pubescens)

12-spotted cucumber beetle {Diabrotica
duodecimpunciaia) .

May beetle {PhyUophaga)

May beetle (Scarabaeidae)

Leaf chafer {Anomala ludoviciana)

Dung beetle {Aphodius sp.)

Ground beetle (Triplecfrus)

Ground beetle {Eumolops)

Ground beetle (Carabidae)

Ground beetle (Chlaenius)

Darkling beetle (Tenebrionldae)

Lepidoptera:

Moths, butterflies, and skippers (3 spe-
cies) .
Salt-marsh caterpillar {Estigmene acraea)..

Diptera:

Gall gnat (Cecidomyiidae)

Robber fly {Asilus) -

Hymenoptera:

Gall fly (Gynipidae)

Chalcid fly (Ghalcidae)...

Paper wasp (Polistes) .-

Ant {Odontomachus haemotodes)

Ant {Pheidole sp)

Fire ant {Solenopsis sp.) .-

Form eaten

Adult..
do-

do

Adult, larva.
do

_do-
-do.

-do-

do

do

do

do

Adult, larva, egg.

Adult Sp.

Adult, eggs Sp-

Adult Su.

.do-
.do..
.do-

.do.
-do-
.do.

-do-
.do-
.do.
.do.
.do.
.do-

_do_
.do.
-do.

-do.
.do.
.do.
.do.
.do.

.do.
.do.
.do.

.do.
.do.
.do.

.do-
.do-
.do.
-do.
.do.

.do....

.do

.do.--.

-do

-do

-do

-do

-do--..
-do

Adult, larva,
---do

Larva.
Adult.

Adult, eggs

Adult.-..

Adult, pupa eases.

Adult

....do

....do...

Seasons when
eaten i

Au

Su, au-

Sp, su

Sp, su, au-
Au

Au

Su, au

Su.

Su

Su

Su

Sp, su, au.
Su, au

Wi, sp.
Su, wi-

Wi

Sp----.
Sp, au-
Sp, au-
Su, wi-
Sp, su..
Sp.-...

Sp

Sp-.-..
Sp....-
Sp, su..

Wi

Wi

Au

Sp.....

Wi

Su, au

Su

Su, au

Au

Au

Sp, su, wi.
Sp, su, au.

Wi

Wi

Wi

Sp..-
Su, au.
Au

Au, wi

Wi

Sp, su, au.

Su

Au

Su, au.
Su, au.

Sp.
Au.

Sp.....

Su

Su, wi.

Wi

Su

Su

> See footnote 1, table 9.
' See footnote 2, table 9.

Au, wi, sp St.

Sp, su, wi. St.

Au St, sc.

ATTWATER'S PRAIRIE CHICKEN 29

Native plants are the most important source of food for the
prairie chicken. Rated according to frequency of occurrence in
stomachs and scats, gross bulk, and periods of availability, ruellia
{Ruellia) appears to be by far the most important single food. It
occurred in 13 of the 18 stomachs and made up almost 27 percent
of all the material eaten. Stargrass {Hypoxis) , bedstraw {Galium)^
doveweed ( 6^rc>z^cn) , perennial ragweed {Ambrosia psilo'Stachya) also
were eaten freely through long seasons. Practically all the impor-
tant food plants utilized by the prairie chicken grow naturally in
pastures that are moderately grazed. Corn was the only cultivated
grain found, and the small quantity present was probably waste.
It is known, however, that prairie chickens are fond of certain crops,
especially peanuts, hegari, and ripened rice. The birds frequently
congregate in peanut patches, particularly after the harvest, and
scratch for the waste pods. They also use conveniently situated
hegari fields extensively in summer, but the good shade in such
areas is probably as attractive as the grain. Prairie chickens also
range into rice fields after the crop is cut and shocked, and they
sometimes feed on the grain in the shock as well as on that so freely
wasted on the ground (pi. 7). The rice taken from shocks usually
is not objectionable, although L. D. Roberts, Eagle Lake, Tex.,
reports that he saw approximately 1,500 of the birds feeding in a
single field of about 500 acres in the Egypt section, Wharton County,
in September 1920. By scratching, the prairie chickens loosen the
shocks, thus allowing moisture to seep in, and this causes some com-
plaint. A large increase of prairie chickens might conceivably bring
on control problems in certain areas. The difficulties would prob-
ably not be serious, however, because the birds could easily be
frightened by shotgun fire or by other disturbances, and they quickly
desert areas of potential danger.

Among insect foods of Attwater's prairie chicken, 11 grasshoppers
(6 identified to genus or species) are especially prominent; 32
beetles (identified to genus or species, including 16 weevils) also
are important. The vast majority (50 of 65) of the insects eaten
by prairie chickens are kinds neutral (25) or harmful (25) to
agriculture. Field observations, and reports of cooperators, show
that prairie chickens eat in large quantities the moths of the cotton
leaf worm {Alabcmm argillacea), one of the worst insect pests in
the coastal area. Under ordinary conditions, the food habits of
Attwater's prairie chicken, considering both insect and plant con-
sumption, are such as to make it one of the most valuable birds of
farm and range.

30 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

HABITAT REQUIREMENTS
KIND OF HABITAT BEST SUITED

The coastal prairie grassland is the real home of Attwater's
prairie chicken, particularly in areas characterized by diversity of
vegetation (pi. 8). Woodland, brushland, and cultivated land each
furnish some food and cover at certain times and under certain
conditions, but use of these types is optional with the prairie chicken,
not vital. These secondary habitats are frequented mostly when food
and cover are at the annual peak, as in September and October,
but are little used at times of seasonal scarcity, as in December,
January, February, and early in March. Wooded, cultivated, and
brushy areas, individually or in combination, contribute little or
nothing as courtship grounds and nesting cover. Properly managed
grassland (pi. 9), however, satisfies every known requirement of
Attwater's prairie chicken, and management, therefore, should be
directed toward improvement of these areas.

CHARACTER AND DENSITY OF VEGETATION

Optimum food and cover conditions seemingly are approached
when the prairie vegetation is varied in species, interspersion, and
density. The plant life of well-populated areas includes a variety
of grasses, sedges, rushes, and legumes, and tall weeds or their cover
equivalent in the form of scattered clumps of myrtle or live-oak
brush. The combination and density of the plants in the most
favored places invariably is such as to provide cover in all degrees
and well distributed.

Light cover serves (1) exclusively for the courtship performance,
(2) for feeding at all seasons, and (3) for a resort when dew is
heavy or after rains. Light to medium heavy cover is used (1) for
roosting, especially on gentle slopes, (2) by chicks under 5 weeks old,
and (3) for feeding by adults throughout the year. Cover of a
medium heavy to heavy character (pi. 9) is utilized (1) extensively
for nesting, (2) as a loafing cover except during the hot summer
months, and (3) as feeding grounds and escape cover in emergencies.
Heavy cover (pi. 9) is essential (1) for shade in summer, (2) for
protection against unfavorable weather and predators at other sea-
sons, and (3) as a source of food, especially in fall.

TOPOGRAPHY

Kichness and variety in the vegetation are promoted by even slight
variations in topography and soil (pi. 8). Consequently, the best
natural range for Attwater's prairie chicken comprises country in
which knolls, ridges, or hog wallows, are frequent. Further, knolls

North American Fauna 57. Fish and Wildlife Service

Plate

ik.w$^ ^^

,yi,

B49728; B50016

Above, Diversified cover — excellent prairie chicken range; Colorado County,
Tex., approximatelj' 7 miles northeast of Eagle Lake, September 4, 1936.
Below, Diversity of topography and vegetation; Austin County, Tex., approxi-
mately 6 miles northeast of Bellville, June 13, 1936. (Photos by V. W. Leh-
mann.)

North American Fauna 57. Fish and Wildlife Service

Plate 9

Above, Medium-heavy to heavy cover — excellent food-cover conditious in a
moderately grazed pasture; Colorado County, Tex., approximately 8 miles
north of Eagle Lake, December 21, 1936. Below, Heavy cover, mostly myrtle
brush, near stream- — excellent summer range; Austin County, Tex., approxi-
mately 4 miles east of Bellville, July 14, 1936. (Photos by V. W. Lehmann.)

ATTWATER'S PRAIRIE CHICKEN 31

and ridges are least likely to be inundated in times of flood and they
afford the birds better opportunity of eluding their natural enemies
and man.

WATER

The balanced prairie chicken habitat should offer a generous supply
of surface water throughout the year. Although Attwater's prairie
chickens may not be dependent on free water for survival during
normal years (see p. 23), it has been established that their favorite
summer range is rather well watered. During unusually dry years
such as occurred in Refugio County in 1917, surface water may be
an absolute necessity. Also, through its effects on vegetation and
insect life, water is necessary for the maintenance of optimum cover
and food conditions. The water supply of prairie chicken areas
apparently is about optimum when permanent sources are available
throughout the range at intervals not greater than a mile.

Briefly, then, habitat conditions for Attwater's prairie chickens
seemingly approach the ideal in grassland area when (1) the vegeta-
tion is diversified and native grasses, sedges, legumes, and small and
large weeds, or their equivalent in the form of brush or dwarfed trees,
are present in such stands as to provide all densities of cover; (2)
knolls, ridges, and hog wallows are frequent and the soils vary from
loose sand to tight clay or silt; and (3) permanent sources of surface
water are available not more than a mile apart.

SEASONS OF SCARCITY

In evaluating the suitability of an area for Attwater's prairie chick-
ens it is to be kept in mind that its productivity or carrying capacity
is not determined by conditions during the best season in a good year.
Rather, as Taylor (1934) states, conditions that prevail during the
most critical season of the year and in the most extreme year in a
series of years determine carrying capacity. In the coastal country of
Texas the season of scarity, or the period when food and cover are at
a minimum, normally is from December through early March. The
most critical years are those of heavy rainfall in May.

LIMITING FACTORS

Factors that have contributed to the decrease of prairie chickens
in Texas may he classed roughly as (1) natural, including unfavor-
able weather, predators, and disease; and (2) artificial, including
cultivation, heavy grazing, burning, and overshooting. It might
be more accurate to class limiting factors as those brought about by
man, directly or indirectly. Although it is not generally appre-
ciated, the decrease of prairie chickens in coastal Texas corresponds

32 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

with the spread of civilization. Prior to the coming of white men,
the number of birds probably was well maintained, but there was
a decrease as the country was developed. Attwater's prairie chicken
will become extinct unless man ceases to exploit the soil, water, and
other natural resources of its range.

NATURAL FACTORS

Rainfall Duiung the Nesting Season

Prairie chickens in Texas evidently suffer greatly at times from
extremes of weather. Heavy precipitation during the nesting and
brooding season (March through June) appears to be an especially
serious hazard, as indicated by the studies of Waddell and others in
Colorado, Austin, and Wharton Counties. From 1925 through 1937
Waddell estimated the size of the annual crop of young prairie
chickens on the basis of the number of birds, both young and old
(1) observed on almost daily trips through their range, (2) seen by
reliable resident observers, (3) bagged by hunters, and (4) counted an-
nually on the courtship grounds in spring. From his studies he con-
cluded that crops of young prairie chickens were (1) good in spring
months when rainfall was below average, (2) fair to good when
rainfall was average or only slightly above average, and (3) poor,
very few young being reared, when the nesting season was abnormally
wet.

Waddell's impressions as to the correlation between the amount of
precipitation in spring and the size of the amiual crop of chickens
were tested rather thoroughly in 1936 and again in 1937. In August
1936, after a reconnaissance made with car and dog (see p. 52) over
approximately 25,000 acres of territory in Colorado and Austin
Counties, it was estimated that the annual increase was less than 10
percent. Rainfall there was below average in March, April, and
June 1936, but it exceeded 10 inches, or approximately twice the aver-
age, in May, as shown by the records of the Weather Bureau at
Columbus, situated centrally in that area. In 1937, when records
of this station showed that rainfall was 2 inches or more below
average in April, May, and June, rope counts made of 3,450 acres
both before and after the breeding season revealed a 95-percent in-
crease, supporting Waddell's estimate that the increase was good
in a dry season.

In table 11, p. 33, Waddell's estimates of the favorableness of the
years from 1925 through 1937 for prairie chicken reproduction are
presented together with precipitation records of the Columbus
Weather Bureau Station for March, April, May, and June in those
years.

ATTWATER'S PRAIRIE CHICKEN

33

Table) 11. — Reproductive yield of AtHvater's prairie chicken in relation to spring
rainfall in inches * in the Colorado County area ^

Estimated
yield

March

April

May

June

Year

Precip-
itation

Depart-
ure
from

normal

Precip-
itation

Depart-
ure
from

normal

Precip-
itation

Depart-
ure
from

normal

Precip-
itation

Depart-
ure
from

normal

1925

1926

Good

Fair...

0.33
11.54
3.49
1.42
4.54
2.34
3.84
3.63
2.36
3.91
3.72
1.23
6.01

-2.50

-8.71

.66

-1.41

1.71

-.49

1.01

.80

-.47

1.08

.89

-1.60

3.15

0.99
7.86
4.00
3.76
2.58

.48
1.43
2.19
1.43
4.28
4.58
3.48

.52

-2.74
4.13

.27

.03

-1.15

-3.25

-2.30

-1.54

-2.30

.55

.85

-.25

-3.21

2.87
4.10

1.24
2.00

16.12
3.11
1.98
.66
3.67
1.90
9.21

10.65
.47

-1.51

-.28

-3.14

-2.38

11.74

1.27

-2.40

-3.72

-.71

-2.48

4.83

6.27

-3.91

1.06
3.37

6.43

8.52

.99

.89

.90

3.68

1.40

.22

2.48

.79

1.37

-2.12
.19

1927

1928

1929

Good...

do

Poor

Fair

Good

do

Fair

3.25

5.34

-2.19

1930

1931

-2.29
-2.28

1932

1933

.50
-1.78

1934

Good

Poor.

do

Good

-2.96

1935

1936

1937 - .

-.70
-2.39
-1.81

I Records of U. S. Weather Bureau Station, Columbus, Colorado County.

' Colorado County, north central Wharton County, southwestern Austin County.

Waddell found good crops of young birds in the Eagle Lake area
in 1925, 1927, 1928, 1931, 1932, 1934, and 1937, years when rainfall
in May was 1.5 inches or more below average. Fair crops of young
prairie chickens were thought to have been reared in 1926, 1930,
and 1933, when rainfall in May was approximately average (0.28
below in 1926) to only slightly above (1.27 above in 1930). Poor
crops were matured in 1929, 1935, and 1936, when May precipitation
was appreciably above (approximately twice) the average for that
month. Unusually heavy or light precipitation in March, April, or
June evidently had little influence on the broods of young, for good
crops were recorded in 1927, when rainfall was decidedly above
average in June, and a poor crop is known to have occurred in 1936,
when rainfall was below average in all months of the nesting season
except May. The records at hand suggest, therefore, that the rain-
fall in May is a fairly satisfactory index of the suitability of the
year for the reproduction of Attwater's prairie chicken under natural
conditions. Good crops usually result when the rainfall in May is
1.5 inches or so below average; fair crops are probable when it is
approximately average or only slightly above ; and poor crops appear
almost a certainty when the rainfall for that month is decidedly
above average.

Rainfall in May is of greater significance than that in any other
month, as the 1937 and 1938 nesting studies showed that most of the
chicks hatch in May. Those hatched in April do not yet have a
serviceable covering of feathers by May and, consequently, are almost
as vulnerable to the rains as are birds hatched in that month. Nests
flooded in March and April may be rebuilt, for the booming season
is still in full swing, but nests flooded after May 1 are seldom re-

34 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

established because the mating season is then nearly over. May, then,
is the climax, or peak month, of the breeding season, and rainfall
then is of greater significance than at any other period.

The nature of rains in May may be a factor modifying the use of
precipitation records for that month as indices of the number of
chicks produced, as short, heavy downpours may be more serious than
slow steady rains. This, however, is not known to be true. Fre-
quently recurring light rains may be as serious as heavy ones be-
cause persistent damp conditions result in heavy juvenile mortality
from chilling. Stoddard (1931: 39, 202) shows that wet spring
months are favorable for hatching but not for rearing bobwhites.
Percolation and drainage are slow in the heavy coastal prairie soil,
and surface moisture accumulates from persistent light rains as surely
as from brief heavy ones ; the amount rather than the severity of the
rain seems to rule.

While it is realized that annual precipitation, drainage, cover, and
other environmental conditions in Colorado County are not identical
with those obtaining throughout the coastal prairie chicken country,
a marked similarity does, nevertheless, exist. Rainfall is moderately
heavy, 39 inches annually, at Columbus, Colorado County, and it is
also generous throughout the bird's range. Average annual precipi-
tation varies from 49.35 inches at Beaumont, Jefferson County, to
33.69 inches at Austwell, Refugio County, at about the eastern and
western limits, respectively, of the subspecies. Rainfall during May
at Columbus (average, 4.38 inches) is heavier than in any other
month. May is the wettest month in Jackson, Goliad, Lavaca, and
Harris Counties as well. Heavy or persistent rains transform tre-
mendous areas in Colorado County into veritable lakes ranging from
a few inches to several feet in depth; rains produce similar results
throughout the coastal region. It appears justifiable, therefore, to
assume that rainfall in May is the key to prairie chicken reproduc-
tion throughout coastal Texas (fig. 3).

Of every 5 years in a given locality, apparently 2 are favorable
for nesting, 2 fair to poor, and 1 bad, as determined by rainfall in
May. Conditions are never uniform in the chicken country a^ a
whole because there is variation between counties and even between
parts of the same county. Records of the Weather Bureau for May
1935 show, for example, that rainfall at Galveston, Galveston County,
was favorable (2.71 inches below average) ; at Houston, Harris
County, fair (only 0.20 inch below) ; and that at Columbus, Colorado
County, poor, being approximately twice average (4.83 inches above).
During 1926 in Brazoria County conditions were good at Alvin, fair
at Angleton and Freeport, and poor at Brazoria. In 1932 conditions
were good at Angleton, fair at Freeport, and poor at Alvin. Though
man cannot regulate rainfall to promote prairie chicken welfare

North American Fauna 57, FUh and Wildlife Service

LEGEND

Rainfall 1.50 inches or more below average f^— ) (Good)

Rainfall 1.49 below average to 1 .99 above average fTTT] (Fair)

Rainfall 2 inches or more above average, but less than twice average MlIlX (Poor)

Rainfall twice average or more ^^B (Bad)

Compilations based on records of average annual May rainfall at each individual station as

supplied by the Climatological Division. U. S. Weather Bureau.
Records missing or unsatisfactory ^ | O )

Rainfall conditions in May in the range of Attwater's prairie chicken in Texas, in the 66 years 1871 to 1936.
indicating the probable frequency of good and other reproductive years for the birds

ATTWATER'S PRAIRIE CHICKEN 35

at nesting and brooding time, in many caseg he can regulate pasture
burning and grazing to provide adequate nesting cover in the best-
drained parts of the land. Also, he can consult weather records
before setting open seasons and bag limits, which should not be
uniformly applied but adjusted to local conditions to preserve the
birds.

Floods

Flood conditions are often produced by heavy rains in the over-
grazed and overf armed sections in the upper part of the State.
Heavily burdened streams carrying flood crests from the upper coun-
try sometimes spill over their low banks and spread their silt-laden
waters over thousands of acres of prairie chicken range. This oc-
curred in the Rock Island-Garwood section (Colorado County) in
June 1936, when the prairie chicken population of that section was
extirpated. Floods evidently are a constant menace to birds near
major streams.

Dbought

Extreme drought seriously affects prairie chickens, especially dur-
ing the hot summer months. G. P. Ferguson, State game warden,
and fence riders on the M. O'Conner ranch, Refugio County, found
many dead birds in the especially dry summer of 1917 and saw others
too weak to fly. Drought reduces food supplies for both present and
future use. Large cracks that form in black soil in dry weather pos-
sibly trap some young birds, according to the observations of Gross
(Bent 1932: 253). Birds weakened by excessive heat, and possibly
also by a shortage of food, are especially vulnerable to disease, pred-
ators, adverse weather, and other hazards.

HUBBICANES

Tropical hurricanes sometimes produce flood conditions in prairie
chicken country 20 miles or more from the Gulf. In 1917 a storm
backed salt water over the greater part of the Pipkin ranch in the
Big Hill area in Jefferson County and drowned livestock by the
hundreds. That it evidently destroyed many prairie chickens as
well was indicated by their exceeding scarcity for 15 years afterwards.

Hatt.

Heavy hail storms destroy many Attwater's prairie chickens, es-
pecially in areas where heavy protective cover is lacking. After a
storm in May 1934, J. O. Linney, Guy Ferguson, and fence riders on
the Salt Creek ranch, Refugio County, saw about 150 dead or crippled
chickens.

36 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

Local Stobms

Local storms, especially those that commonly occur in fall, kill
turkeys and other domestic fowls and prairie chickens and other game
birds, as reported by Marcus Shindler, Ed Koy, and other resident
land-owners in the northeastern part of Colorado County.

Disease

Gross (1930a: 39), and Stoddard, Curtis, Lews, Terrel, and others
(Leopold 1931: 182-183), recorded incidents strongly suggesting that
disease and parasites probably were important controlling factors on
the abundance of the greater prairie chicken of the Northern States.
Records at hand do not show that, in the past, disease has been a factor
of importance limiting the numbers of Attwater's prairie chicken in
Texas. The observations, mentioned above, made by G. P. Ferguson
on the M. O'Conner ranch furnished the only known evidence even
faintly suggesting an outbreak of disease. In that instance, however,
it is probable that mortality, if really due to disease or parasites, was
an indirect result of prolonged drought. No evidence of disease or
heavy parasitism was found in autopsies made on 13 prairie chickens,
and no evidence of any unhealthful condition was observed among
hundreds of birds in the field. Prairie chickens are doubtless sus-
ceptible, however, to ailments of domestic poultry. An outbreak of
blackhead disease, probably contracted from domestic turkeys, is con-
sidered by Gross (Bent 1932: 268) as a major factor in the extermina-
tion of the heath-hen. Turkeys and other poultry, therefore, probably
are unhealthful influences on a prairie chicken range.

Speead of V^oody Vegetation

The encroachment of mesquite, live oak, various acacias, and other
kinds of brush onto open prairie land has been an extremely impor-
tant factor in reducing the range and doubtless the numbers of Att-
water's prairie chickens in Refugio and other counties to the south
and west. Within the memory of living men extensive prairies have
been transformed into brush jungles. Specific factors that have in-
fluenced the rapid vegetational changes in the southwestern brush
country are imperfectly understood. Factors probably of importance
in enabling woody plants to replace the native grassland flora have
been overgrazing, especially during drought j^ears; the mechanical
planting of tree seeds by cattle and horses, because livestock eat large
quantities of mesquite and other beans, the seeds of which pass through
the digestive tract and are distributed or planted by the droppings;
the elimination of burning, previously mentioned by Bray (1901:
288-290) and Tharp (1926: 71) ; and the lowering of the water table.
Be that as it may, hundreds of thousands of acres of what was once

ATTWATERS PRAIRIE CHICKEN

37

tall-grass prairie are now brushlaiid, and prairie chickens are gone
from these areas.

Pkedation

Natural enemies consume some of the eggs, young, and adults of
Attwater's prairie chicken. The extent of predation on nests and the
identity of other factors responsible for nest loss in the Eagle Lake
area are eiven in tables 12 and 13.

Table 12. — Fate of nests, Eagle Lake, Colorado, 1937

Nest
No.

Pasture

Date
found

Date
destroyed

or
hatched

Probable cause of destruction

1

Everett

Apr. 7

Apr. 8
Apr. 12
__..do....
Apr. 13
Apr. 21
____do..__
_...do.-_.
Apr. 27
Apr. 29
May 1

May 11
June 1

Apr. 22

Apr. 4 3
Apr. 7
Apr. 13
May 17
Apr. 15
Apr. 21
Apr. 18
May 4
May 15
May 2

Apr. 6
June 2

or
June 3

21

3 1

do

Wintermann_

by a farmer.
Red wolf— female bird killed on nest.

4..

5

do

Sklar-M arcella

Wiutermann

Skunk.

Hatched successfully.

61

Skunk.

7 1

81

9

do

do

Duncan

Do.
Deserted, cause unknown.
Hatched successfully.

10

Everett .

Do.

11

Willis

Man — nest deserted after farmer plowed territory

nearby and revisited nest frequently.
Opossum.

12 2

13

Sklar-Marcella

... do

1 Nest destroyed when found.

' Estimated in case of nests destroyed when found.

> Indicated by circumstantial evidence at the nest.

Table 13. — Fate of nests. Eagle Lake, Colorado County, Tex. (1938)

Nest No.

Pasture

Date
found

Date de-
stroyed

or
hatched

May 3
Apr. 23
Apr. 29
May 3
May 11
June 21

Probable cause of
destruction i

14

Sen

Apr. 13
Apr. 18
.. do ..

Heavy rain.

15

Thomas . .

House cat.

16

do

Hatched successfully.

17

... do

Apr. 29
Apr. 20
June 3

Heavy rain.

18

Everett

Opossum.

19

. do

Hatched successfully.

I Indicated by circumstantial evidence at nest.

Of 19 prairie chicken nests studied in 1937 and 1938, 6 (31.5
percent) were successful, and 13 (68.4 percent) were lost. In 1937
8 of 13 nests studied were destroyed before the clutches were com-
plete, showing that the laying period may be the one of heaviest
nest loss. This might be expected, as the eggs are covered only about
an hour or so each day during that time. This loss is somewhat
compensated, however, by renesting (see p. 15). Opossums and
skunks destroyed 6 nests — more than any other agency. Of the 6,

38 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

however, 4 were in a pasture where unregulated fire had restricted
good nesting cover to 2 small unburned areas about 3 and 10 acres
in size. Fur animals as well as nesting chickens were unnaturally
concentrated in these unburned plots; dens containing young were
100 yards or less from each of the nests destroyed. Excessive pas-
ture burning appeared to be the primary cause of the heavy nest
loss ; predation by fur animals being merely an effect, the agency of
destruction that was inevitable after the burning.

Field evidence showed that a red wolf killed a female prairie
chicken and destroyed her nest; a feral house cat devoured the
eggs from another nest (pi. 10). It is surprising that dogs did
not figure as predators on the nests and that house cats did not
take an even greater number. Wandering dogs, usually in groups of
three to five, were not imcommon on Colorado County prairies; L.
A. Burchfield, a trapper who worked for the former Bureau of
Biological Survey in Colorado County in 1937, and Waddell found
that dogs did much of the damage for which the few red wolves,
now largely extirpated in the area, were blamed. Heavy predation
on a flock of domestic turkeys, supposedly by wolves, stopped imme-
diately when a hound, which frequently hunted on its own initiative,
was killed after having been caught in a trap set for the alleged
wolves. Feral house cats on Colorado County prairies probably
outnumber skunks, opossums, minks, or any other fur animals. Cot-
ton rats and other rodents were common near several nests but took
no eggs. Neither did racers, chicken snakes, king snakes, or other
reptiles frequently noted after May 1 in both 1937 and 1938.

Three nests were abandoned, desertion of two of these, possibly
all three, being caused by man. Nesting* prairie chickens seem
especially sensitive to interference, and they should not be dis-
turbed by persons making repeated visits. Of six nests under obser-
vation in 1938, floods destroyed two, and accumulated water from
heavy rains came within ly^ feet of a third (nest 16). The following
excerpt from the writer's field notes of May 3, 1938, emphasizes the
importance of floods :

The prairie has been transformed into a miniature ocean clotted by tiny
islands that previously had been the tops of knolls and ridges. On these
islands sit wet and bedraggled prairie chickens and other birds that seem as
confused and astounded as I by the sudden change in their environment. About
a 5-inch depth of vrater covers the sites of nests 14 and 17, and former
nest 15. Nest 16 has escaped by a hair's breadth, but the lining is very soggy.
Problems due to hawks, skunks, and other predators seem so petty v^^hen exces-
sive rain destroys virtually everything at a single stroke.

YOUNG

Although predators doubtless exert great pressure on the popula-
tion of young prairie chickens in some areas, especially because the

North American Fauna 57, Fish and Wildlife Service

Plate 10

ICV

■■"4ili

Above, Shells of eggs at prairie chiclcen nest destroyed by house cat; Colorado
County, Tex., approximately 5 miles north of Eagle Lake, April 23, 1938.
(Photo by Texas Game, Fish, and Oyster Commission; E. P. Haddon.) Below,
Freshly killed prairie chicken recovered from a ferruginous rough-legged hawk;
Colorado County, Tex., approximately 6 miles north of Eagle Lake, April 7,
1937. (Photo by V. W. Lehmann.)

North American Fauna 57, Fish and Wildlife Service

PLATE 1 1

B49777A; E60058

Above, Xative bluestem prairie— well populated b.y prairie chickens: Colorado
County, Tex., 6 miles northeast of Eagle Lake, December 21, 1936. Below,
Prairie after plowing for rice — deserted by prairie chickens; Colorado County,
Tex., ajjproximately 5 miles north of Eagle Lake, March 7, 1938. (Photos
by V. W. Lehmann.)

ATTWATER'S PRAIRIE CHICKEN 39

loose formation of the brood (see p. 19) and straying apparently
induce mortality from this cause, definite information was difficult
to obtain.

A female Krider's hawk caught a third-grown prairie chicken on
May 24, 1937. From then on until June 9, when the male was col-
lected, this hawk, assisted by her mate, hunted a section containing
six broods. Most migrant raptors (roughlegs, redtails, and marsh
and duck hawks) had left Colorado County prior to May 1, 1937, and
April 15, 1938, before many young had hatched, and the resident
species (red-shouldered. Cooper's, Sennett's white-tailed, and Krider's
hawks) confined their activities largely to wooded areas. Because
cover is dense in summer, and hawks are then uncommon, probably
few young prairie chickens are taken in normal years.

House cats with freshly killed young prairie chickens were noted
twice in 1937 and were seen stalking broods on three other occasions.
Because of their numbers and predilections, house cats are thought
to be exceedingly destructive.

ADULTS

Prairie chickens on the courtship grounds seemed more intent
on mating than on self-preservation ; consequently, losses from preda-
tion were probably heaviest at mating time. In Colorado County,
during most of the 1937 and 1938 courtship seasons the abundant
hawks harassed the prairie chickens persistently, sometimes with
success. On April 8, 1937, 3 duck hawks, 7 marsh hawks, 2 rough-
legs, 3 Krider's hawks, and 2 bald eagles kept the chicken population
(about 45 birds) of the Everett pasture (640 acres) constantly mov-
ing. A freshly killed male prairie chicken (pi. 10) was taken from
a ferruginous roughleg in that area on April 17. Marsh hawks,
which Stoddard and others have found to be sometimes more bene-
ficial than harmful to quail and other game, were especially an-
noying to courting birds, no other factor interfering with their
activities to so great an extent. Wlien a marsh hawk darted at
one occupant of the booming ground, others generally cowered. The
hawks pursued their intended victims for short distances, but soon
returned and flushed others, or after dispersing the grouse, fre-
quently alighted on the courtship grounds to await their return and
resume the flushing tactics. On April 8, 1937, 4 marsh hawks con-
centrated on a single courtship ground and harassed the 6 male
occupants from 5 to 7 :30 p. m. Although no birds were killed, one
lost many feathers when two hawks dived at it simultaneously.

By flushing prairie chickens, marsh hawks render them vulnerable
to more efficient winged enemies, as duck hawks, goshawks, and the
like. Waddell has seen duck hawks catch adult chickens on at
least two occasions.

40 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

Mammals also take some of the birds on courtship areas. The
stomach of a male house cat collected April 12, 1937, near a booming
ground in the Everett pasture, Colorado County, contained the head,
feet, and part of the breast of a freshly killed male prairie chicken.
The remainder was found about 50 feet away.

A red wolf was suspected of killing a female prairie chicken on
the nest, and either red wolves or dogs took three others in Colorado
County in April 1937.

Only six instances of adult mortality were discovered that year,
although intensive search for remains was made on foot and in cars
over approximately 2,500 acres. In 1938, when none of the pastures
were burned, no dead birds were found.

Review of Natueai. Factors

Natural factors limit the abundance of prairie chickens by destroying
eggs, young, and adults and by reducing favorable territory. During
the breeding season floods, storms, hail, drought, and excessive or per-
sistent rains are known to be locally serious, the rains in May being most
damaging. Drought has been associated with the only reported out-
break of disease that occurred in the Refugio area in 1917. The en-
croachment of brush on prairie land has transformed thousands of acres
of what was once good prairie chicken range (pi. 11) into an unfavorable
habitat. Although some predators harass the birds throughout the
year, their effects are probably most serious at mating and nesting
time. Natural mortality from climate and predators is severe in
inferior or isolated cover.

The serious effects of natural factors are in every case either brought
about or intensified by man's generally unwise treatment of natural
factors. All except feral house cats and predatory dogs were operat-
ing against the prairie chickens, apparently without disastrous results,
before the environment was radically modified by man. Since the
unfavorable influences of natural agencies are due chiefly to man, it
is encouraging to know that it is within his power and often decidedly
advantageous to him so to modify his actions as to improve existing
conditions and promote the welfare of the prairie chickens as well as
his own.

ARTIFICIAL FACTORS

Agrictjlttiee

Much of the best prairie chicken range has been recently appro-
priated for agricultural uses. More than 2,000,000 acres (table 14)
were cultivated in 1936. In addition, thousands of acres of sod are
plowed annually, with the extension of agriculture, especially rice
farming. The acreage yearly planted to rice in coastal Texas in-

ATTWATER'S PRAIRIE CHICKEN

41

creased from approximately 174,500 acres in 1922 to more than 196,500
acres in 1937.^ This latter acreage represents only a small part of the
area actually depleted: rice farming has ruined more than 84,000
acres for the birds in Colorado County (pi. 11) alone, and probably
in excess of a million acres in the State as a whole. Rice is hard on
the land and most areas devoted to it can be profitably cultivated
during only about 1 year in 4, after which they must be left fallow for
about 3 years to "sweeten." Weedy rice fields ostensibly provide satis-
factory grouse range ; actually, however, they lack suitable courtship
grounds and safe nesting cover, and, furthermore, the levees collect
water that floods nests. Prairie chickens in fallow rice land ap-
parently are doomed even though they are hunted lightly or not at
all. According to Waddell, there were 10,000 of the birds on 30,000
acres of the Egypt section, Wharton Comity, in 1924. Rice farming
began there in 1925, and by 1937 all the 30,000 acres were either in
cultivation or fallow. Hunting pressure was reduced annually after
1925, and few, if any, birds were killed after 1935. In 1938, however,
less than 150 prairie chickens remained. Prairie chicken decrease was
also positively correlated with the expansion of rice farming in eastern
Chambers and central Matagorda Counties. As additional acres of
prairie are plowed, further decreases are certain to follow.

Table 14. — Harvested and other crop land (1936) in counties partially or en-
tirely within the probable former range of Attwater's prairie chicken in
Texas ^

County

Harvested
crop land

Other
crop land

Total
crop land

County

Harvested
crop land

other
crop land

Total
crop land

Aransas

Acres

2,484
105, 396

92, 247
98,045
34, 425

130, 684
16, 772

93, 562
153, 307
176, 495

19, 848
64, 374
100, 263
82, 609
44,205

Acres
1,772
11,313
20, 685
17, 787
12, 580
32, 483
5,727
11,038
30, 843
41, 793
3,396
12, 711
25, 031
15, 114
12, 386

Acres

4, 256
116, 709
112, 932
115, 832
47, 005
163, 167
22, 499
104, 600
184, 150
218, 288
23,244
77, 085
125, 294
97,723
56, 591

Kenedy .

Acres

204

28, 639

158, 604

47,704

59, 714
228, 609

8,245
40, 147
165, 691
100, 300
47, 986
187, 555

60, 981

Acres

Acres
204

7,211
10, 652
12, 836

21, 952
62, 090

1,334
16, 183
35, 769
17,828
10, 040

22, 909
10, 001

35 850

Bee

Lavaca

169 256

Brazoria

Liberty

60 540

Matagorda. -

Nueces

81 666

Cameron - -

290, 699
9 579

Chambers

Orange

Refugio

56, 330
201 460

DeWitt - -

San Patricio

Victoria

Fort Bend

118, ] 28

Galveston

58, 026

Goliad -

Wharton

Willacy

210, 464
70, 982

Total

2,349,095

483,464

2, 832, 559

1 Data from Texas Almanac and State Industrial Guide, pp. 231-236, The Dallas News, 1936.

Pasture Burning

Unregulated prairie fires intentionally set or of accidental origin
have been, and still are, common in coastal Texas in every month of

•^ Figures supplied by David Wintermann, Relow Land Company, Eagle Lake, Tex., from
data compiled by the Rice Milling Association.

42 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

the year. In Colorado and Austin Counties, for example, approxi-
mately 80 percent of the best prairie chicken country was burned over
in 1936. A fire of accidental inception ran over about 3,000 acres of
the 6,700 acre J. C. Anderson Estate ranch, Harris County, in May
1937. The cover on about 3,000 acres of the best prairie chicken
country in Matagorda County was intentionally burned in July 1937.
Pasture burning is an annual event on the Pipkin ranch, Jefferson
County, fires being started during first new moon after February 15.
Areas that the first fires do not cover cleanly are subsequently treated,
and burning generally continues well through the nesting season.
Most ranchmen, however, complete pasture burning by March.

The pastures are burned to remove old grass and encourage tender
new growth more palatable to livestock. Fires usually are set when
the grass is dry and the wind brisk, in order to finish the job quickly.
The resulting fast, hot fires entirely denude areas except in low spots
and deplete pastures of food, escape cover, and nesting sites. Prairie
chickens and their natural enemies are crowded into unburned areas
(pi. 12) and predation is undoubtedly intensified. During the breed-
ing season fires destroy nests and probably many young birds as well ;
no fewer than nine nests with charred eggs were found by Waddell in
a 640-acre pasture burned in May 1936. Plant life recovers slowly
in the absence of abundant rain; consequently, fires accentuate the
results of drought. Altogether, fire is one of the most important
factors limiting prairie chicken numbers in pastures. When burning
is carried on as outlined under Management (pp. 53 to 54), however,
the evils are greatly reduced or entirely eliminated, and benefits
accrue to forage and soil as well.

OVEKQRAZINO

With the possible exception of Orange and Jefferson Counties, over-
grazing is severe in most of coastal Texas from late in fall through
early spring. In addition to reducing cover and food for prairie
chickens (pi. 12), overgrazing probably also increases the vulnera-
bility of the birds both to natural enemies and to man. In Colorado
County from 1936 through 1938, for example, it was noted that marsh
hawks and other raptors harried chickens more persistently in lightly
vegetated pastures than in areas where heavy grassy cover was pres-
ent. Waddell observed that hunters regularly kill a higher percent-
age of known populations in areas where cover is light than where it
is heavy. In Colorado County it has been found that the winter
prairie chicken population of a pasture can be forecast with consider-
able accuracy by observing the extent to which the area is grazed.
Large winter populations are rare in pastures where cover is short.

North American Fauna 57. Fish and Wildlife Service

PLATE 12

B49778: B60059

Above, Excellent unburned cover at right of road; inferior burned cover at left;
Colorado County, Tex., approximately 7 miles north of Eagle Lake, December
22, 1936. Below, Scanty cover where there has been overgrazing; the shrub is
Cherokee rose (Rosa laevigata); Colorado County, Tex., 6 miles north of Eagle
Lake, March 7, 1938. (Photos by V. W. Lehmann.)

ATTWATER'S PRAIRIE CHICKEN 43

Aside from resulting unfavorably to prairie chickens, overgrazing
in the gulf coastal country has been and is resulting in (1) the spread
of undesirable brush and weeds, (2) the increase of needlegrass and
other largely unpalatable grasses, and (3) serious erosion. An abun-
dance of prairie chickens cannot be maintained on overgrazed tracts ;
it is equally impossible to maintain forage and soil on such areas.

Orti Deveix)pment

Oil development, which began with the discovery of the Spindle
Top field in Jefferson County in 1901, has extended to every county
in the coastal section. All the Attwater's prairie chicken area is
classed as potential oil land, and almost every acre has been surveyed
not once, but several times by oil crews. Veritable forests of oil der-
ricks now stand in areas that once provided some of the finest prairie
chicken range. In these areas, as in Fort Bend County, prairie
chickens are almost, if not completely gone.

Drainage

Drainage canals, as in Brazoria and certain other coastal counties,
have in some instances improved the territoiy within a mile or so of
their margins by providing a permanent water supply where it was
otherwise lacking during the summer months. On the other hand,
drainage canals have doubtless decreased the general wildlife pro-
ductivity of the counties in which they are situated by speeding up
the run-off and thus lowering the water table. Until recently many
prairie ponds retained water throughout the year, produced crappie,
bream, and other edible fish, held safe nesting cover for black
mallards and other water birds, grew an abundant supply of food for
wintering waterfowl, and served as concentration points for prairie
chickens during the heat of summer. Now they go dry during the
slightest drought and produce virtually nothing.

Pastube Mowing

Regular mowing of grassy areas, mainly for hay or increased
forage production, has promoted a nearly pure stand of grass in some
of the areas treated and has reduced shade and food, and the general
attractiveness of the areas for prairie chickens and certain other
valuable wildlife. In Colorado County, areas that have been mowed
regularly for long periods are virtually game deserts ; prairie chickens
use them little even at nesting time. Pasture mowing in coastal
Texas appears to be extending rapidly, and further reduction in
wildlife resources may be expected from this cause unless definite

303807°— 41 4

44 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

modifications are effected. Fortunately, the detrimental effects of
pasture mowing to upland game birds may be mitigated, at least to
some extent, by following management recommendations listed on
page 56.

Mechanical Accidents

Prairie chickens sometimes fly into telephone wires, fences, and
houses, or are struck by automobiles. In six instances in 1937 birds
were noted as accidentally killed in the Eagle Lake section — as many
as were recorded for predation. Mortality from accidents may be
far more serious than is generally appreciated.

Hunting

Last, but not least, hunters certainly have contributed to prairie
chicken decrease. Hunting has never been well regulated, and laws
governing the taking of the birds have always been inadequate. Ac-
cording to the Texas Game, Fish, and Oyster Commission (Ann.
Kept., 1929: 86-91) prairie chickens had no legal protection until
1883, when a 5-month closed season was declared during the breeding
period. In 1885, 148 counties claimed partial or total exemption to this
and other regulations, and it was not until 1903 that the legislature
passed a bill designating the months of November, December, and Jan-
uary as the open season and setting a daily bag limit of 25. The legis-
lation of 1903 was indeed a forward step, but there was no conserva-
tion body to enforce the measure, local officers being depended on
to carry out its provisions. The Game, Fish, and Oyster Commis-
sion was not created mitil 1910, and for many years it was without
adequate funds and personnel. As late as 1919 there were only
6 salaried wardens in Texas endeavoring to carry out, as best they
might, almost wholly inadequate regulations. The law restricting
the open season on prairie chickens to 4 days, September 1 to Sep-
tember 4, inclusive, and the bag limit to 10 a day or 10 a season,
was not passed until 1929. In 1937 there were only 9 full-time
wardens in all the Attwater's prairie chicken country, and they were
charged with patrolling more than 8 million acres !

Development of the coastal territory, as farming, grazing, and
the exploiting of oil, crowded prairie chickens into ever smaller
areas, where they were more easily found and killed. The Hug-
the-Coast Highway (State Highway No. 35) and various other roads
increased patrol problems; the intercoastal canal in Galveston, Cham-
bers, and Jefferson Counties made formerly remote areas easily
accessible to poachers. The number of hunters increased as trans-
portation facilities and weapons were improved. The open season in
September, normally a dry period (see pp. 57 to 58), did much to

ATTWATER'S PRAIRIE CHICKEN 46

popularize hunting from automobiles. With these and other advan-
tages, hmiters evidently harvested too closely. In the Bernard River
country (western Austin County and northeastern Colorado County),
for example, it is known that in 1936 hunters killed 66 chickens, or
22 percent of the estimated total population (300). This was accom-
plished despite the fact that the gunners were closely supervised and
did not hunt in the most densely populated pastures. It is probable
that excessive kills have occurred in other areas for many years;
the 1937 census showed that Attwater's prairie chickens were nearly
or completely extirpated except on or adjoining lands where they had
been hunted little, if at all, for at least 10 years. In Harris, Galves-
ton, Waller, and possibly in parts of other counties, hunting has
probably been the agency most largely responsible for prairie chicken
decline.

MANAGEMENT

Leopold (1931 : 3) has defined game management as the art of
making land produce annual crops of wild game for recreational use.
In coastal Texas, the management of prairie chickens must consist
largely of the preservation of suitable grassland areas. Increased
protection, habitat improvement, adequate predator control, and
proper regulation of the harvest, however, will greatly encourage
recovery.

PROTECTION

An act (H. B. 30) passed by the State legislature, effective Sep-
tember 24, 1937, forbade the killing of prairie chickens in Texas for
a period of 5 years. This measure removes much of the pressure
previously exerted on the birds during the regular open hunting sea-
son, for true sportsmen will observe the decree. Landowners, game
wardens, and other interested individuals, however, will remember
that close seasons may tend to stimulate rather than retard the opera-
tions of game bootleggers. According to the consensus of State
game wardens in the coastal territory, violators are especially active
(1) during the birds' spring courtship season when the conspicuous
males, their instincts of self preservation dulled by the the mating
urge, are easy targets for .22-caliber rifles; (2) late in July and
August, when the tame young birds are of "frying" size; and (3)
during the duck season, when whirring flocks of Attwater's prairie
chickens evidently tempt gunners who have insuificient self control.
Coastal game wardens report that, in years past, probably as many
prairie chickens were illegally killed during the duck season as were
taken legally during the then open season in September. The restora-
tion of the species demands close protection for the remaining birds
at all times.

46 NORTH AMERICAN fAUNA 5 7, FISH AND WILDLIFE SERVICE

Altliough. game wardens in the prairie chicken country are sin-
cere and energetic, the territory is too vast for adequate protection
under the facilities available. Sportsmen may render valuable aid
by helping conservation officers apprehend irresponsible hunters, but
landowners or their resident agents must handle the job if prairie
chickens are to receive anything like adequate protection. Land-
owners, individually or in groups, would do well to incorporate their
holdings to form game-management areas, as advocated by the Ex-
tension Service, Texas Agricultural and Mechanical College and the
Texas, Game, Fish, and Oyster Commission. Under that plan, in-
creased protection is provided through appropriate posting and
provision by the landowners of qualified, resident, free-service State
game wardens. Also technical service is given in solving predator
problems, improving wildlife food and cover, properly regulating
harvests, and otherwise maintaining wildlife crops. Nonresident
owners should check up on the manner in which their foremen
or lessees protect wildlife during their absence. It is regrettably
true that some supposedly protected ranches are in reality hunting
clubs for irresponsible agents and their friends when the landowners
are away. Long-time lessees who wish to manage prairie chickens or
other game should insist that their contract include control over the
wildlife resources of the property as well as over grazing or other
values. These lessees may thus avoid embarrassment from unwel-
come hunter guests directed to the area by the absentee owners.

Landowners who contemplate leasing their holdings for oil develop-
ment might well follow the precedent set by a ranch owner in Refugio
County. Each of his contracts carries the provision that the lease
shall terminate immediately after any representative of the contract-
ing company is caught on the property with a gun of any kind.
Such a clause properly shifts the burden of supervising irresponsible
oil workers from the landowner to the oil company.

Increased protection of the few remaining Attwater's prairie
chickens is necessary for success in management. Protection alone,
however, is largely ineffective in areas where proper food and cover
conditions are lacking.

HABITAT IMPROVEMENT

At present there are few areas in Texas where excellent conditions
for prairie chickens prevail, and populations fluctuate markedly
(table 15). Increase or decrease in study areas was thought fre-
quently to coincide with fluctuations in the supplies of food, cover,
or surface water. In many areas marked seasonal movements may
be averted and larger and more stable populations maintained by
removing deficiencies in habitat.

ATTWATER'S PRAIRIE CHICKEN

47

Table 15.

-Fluctuations in numhers of prairie chickens, Thomas and Koy pas-
tures, Colorado County, 1937

Size

Prairie chickens found

Pasture

Feb.
221

Apr.
13

May
2

June
2, 8, 10

July
26

Sept.
1

Oct.

22

Nov.

1

Thomas

Acres
817
460

Num-
ber
31
0

Num-
ber
30
0

Num-
ber
28
0

Num-
ber
37 16
14

Num-
ber
0
23

Num-
ber
17
32

Num-
ber
0
0

Num-
ber
1

Koy

0

> Count of birds at or near the courtship grounds; no allowance made for any birds that may have been
missed. All other counts were by the rope method. (See p. 49.)

Evaluating Conditions

In some instances it is relatively easy to point out one or more ways
in which areas are inferior. Safe nesting cover is deficient in burned
pastures that are devoid of old vegetation except in low damp places.
Shade is insufficient on lands kept free of tall weeds or shrubs by
mowing or grazing. Winter food, or cover, or both are usually lack-
ing in areas having few native food-cover plants, as ragweed,
goatweed (Croton), marsh-elder, or ruellia. Sometimes, however,
habitat deficiencies are obscure and general observations of an infre-
quent nature do not identify them. Accurate inventories are of
assistance in determining (1) whether habitat improvement is needed,
(2) what should be done, (3) results of work done, and (4) the sur-
plus available for hunting. Management programs should be formu-
lated on the basis of data obtained during inventories conducted
thrice annually, in spring, summer, and winter. Inventory methods,
recording and interpreting data, and management practices are dis-
cussed in the following paragraphs.

Census Methods
spring count on the courtship gb0und8

The first census method that has been tried and found useful is
the spring count of birds on the courtship grounds. Necessary are
an automobile, preferably of light build and high clearance, a driver
who is well acquainted with the area, and someone to act as observer,
note keeper, and gate opener.

A count is made on each courtship ground in the area, recording
the number and sex of birds assembled there and the number and
sex of birds seen between these grounds. The number of hens is
recorded as a supplementary check. The number of males, increased
80 to 110 percent to allow for females that will be missed is accepted
as the total population of the census area. For best results, the

48 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

spring count is made in March from daybreak to about 7 : 30 a. m.
A population estimate based on the maximum count obtained by
reworking the same area tliree successive mornings, or often enough
to offset variations due to unfavorable weather, is likely to be suffi-
ciently accurate. Where recounts are impracticable, the single enu-
meration should be made on a clear, quiet morning after a brisk
norther.

Courtship grounds may be located by sight or sound of the birds
assembled there, but it is preferable to "drive out" the census area in
belts 150 yards or less wide. Drumming grounds should not be ap-
proached more closely than is necessary, because flushing the birds
leads to inaccurate counts. As birds frequently squat, or freeze, at
the approach of a car, it is desirable to wait at each occupied booming
ground and refrain from counting until after vigorous courtship ac-
tivity has been resumed. It is good policy to encircle a counted area
completely before proceeding to a new site, for the fresh car tracks
often assist in avoiding duplication.

The accuracy of spring counts on the courtship grounds was tested
in the following ways: (1) A section (640 acres) was covered on 10
successive mornings; (2) a 1,000-acre pasture was searched with the
aid of 15 bird dogs; (3) a 1,000-acre pasture was recounted by 5 men
using 2 cars, 3 horses, and 2 dogs; (4) a section worked 3 successive
days was rope counted. Spring counts have been made over ap-
proximately 150,000 acres in Colorado, Wharton, and Austin Counties.
These studies show that the enumeration of birds on the courtship
grounds is the most rapid and economical of all known census tech-
niques. More than 2,000 acres a morning have often been covered in
areas having populations of about 1 bird per 45 acres. There are no
indications that the method affects courtship activities adversely or
that it greatly endangers early nesting. Also, the spring count of
males is useful in yielding data on prairie chicken abundance in com-
parable areas worked at nearly the same time and under nearly the
same conditions.

The spring count of birds on the courtship grounds, however, is
not without its defects. Its accuracy is influenced by weather and
other conditions at a time convenient for counting. Opportunity is
limited to a few hours a day (from about 6 to 7 : 30 a. m.) over a short
period (in March). The spring count does not reveal the number of
females present, consequently, it does not produce reliable quantita-
tive data on sex ratio and total population. Some observers experi-
ence difficulty in distinguishing males from females, especially in the
poor light of early morning. To them the analysis of sex differences,
presented on p. 49, may be helpful.

North American Fauna 57, Fish and Wildlife Service

PLATE 13

r»-' '*»-"i»% *

B49329; B48929

Above, Rope counting of prairie chickens on Matagorda Island, Tex., October 30,
1937. Below, Rope counting in myrtle brush; Liberty County, Tex., approxi-
mately 8 miles southeast of Devers, June 27, 1937. (Photos by W. P. Taylor.)

ATTWATER'S PRAIRIE CHICKEN

49

FIELD BASIS FOR DIFFERENTIATION OF THE SEXES IN SPRING

Males
General color Light gray

Color of breast Not perceptibly different

from that of back.

Barring on back and Heavy, black, well defined- .
breast.

Color of head Orange - colored comblike

structure present above
each eye.

Feathers of crest Seldom erect

Neck Appears thick and heavy,

with large, brightly col-
ored (orange) air sac
apparent on each side of
neck under prominent
(2.25 to 2.90 inches) neck
tufts.

Size Large, heavy (about 2}^ lb.).

Action on courtship Bold, struts, fights, and
grounds. booms in open cover.

Flocking Usually in groups of 8 to 12

when booming, feeding,
or resting at midday.

Flushing Laborious take-off; cackles

when rising from ground.

Females

Brownish gray.

Much lighter than that
of back; appears al-
most white in flight.

Light, brownish black,
poorly defined.

Orange -colored combs
absent.

Frequently erect.

Appears thin and long;
air sac and neck tufts
rudimentary {%a inch
long).

Small, light (about 1^

lb.).
Shy, does not strut, fight,

or boom.
Usually alone.

Easy take-off; usually
does not cackle.

THE ROPE COUNT

A second method of counting, one that has been tested with most
encouraging results in the coastal prairie chicken country, is the rope
count (pi. 13). Essential equipment includes two automobiles,
preferably of light build and high clearance, an inch rope or a
quarter-inch flexible steel cable 60 to 120 yards in length, and two
strong swivels. An extra supply of water for radiators is needed
in hot weather as cars heat up under the heavy going. In addition
to drivers for the two cars, a third person should be taken along, if
possible, to act as note keeper and general handy man.

When the census area is reached, one swivel is attached to the
right end of the rear bumper of the car in which the note keeper
is to ride, and the second swivel is attached to the left end of the
rear bumper of the other machine. Each end of the rope or cable
is then securely tied to a swivel. Care must be exercised to see that
the rope or cable, in turning, will tighten its twist and not loosen or
unravel. One machine takes position parallel to a fence or other
definite landmark while the other goes far enough way to stretch the
rope so that only a slight bend remains. After both cars are in
position, they drive over parallel courses at a uniform speed of 5

50 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

to 15 miles an hour, which may be increased in light cover but
which should be reduced in heavy vegetation to keep the rope, for
the greater part of its length, about 5 inches above ground. Birds
are tabulated as they are flushed. Those flying into uncovered ter-
ritory are deducted when that area is dragged. Wlien the end of
a strip is reached, the car in which the note keeper rides turns and

ICAftl

Belt 5. End Here -

*-1

L

Belt4.

Belt 3.

Bsin.

t

Ei 1 1

Rope^i ^art Ksr< -» ^'" '•

FiGUEE 4. — ^Diagram of the rope count. Arrows show course of each car.

retraces its course while the other car makes a wide swing to the
outside margin of another belt (fig. 4). This is repeated until every
part of the census area has been covered.

The principle of rope counting is not new. Askins (1931: 8) re-
ports that market hunters and others in Kansas "* * * hitched
a wire between two wagons and with these driving across the prairie
300 yards apart, the gunners walked behind the wire taking the
grouse as they arose until the wagon was filled." Butchers of wild-

ATTWATER'S PRAIRIE CHICKEN 51

life (they have no claim to the title of sportsmen) have used modi-
fications of this system in coastal Texas for years.

The rope census method has been checked and rechecked for ac-
curacy against the spring counts of birds on the drumming grounds,
car counts, and counts with bird dogs. The rope count, by far the
most efficient of all, has been tried over approximately 45,000 acres
of grouse range, including every major prairie type in which Att-
water's prairie chickens occur in Texas. Heavy myrtle brush, as
in Jefferson, Liberty, and other counties east of the Brazos River,
light to heavy Paspalum-Andropogon grassland as in Brazoria and
Colorado Counties, live oak shinnery in the northern part of Vic-
toria County, rough weedy hog wallow blackland of Refugio County,
and even the extremely rough salt-grass area of the same section,
have all been negotiated successfully. These tests have demonstrated
clearly the practicability of the rope count under all coastal prairie
conditions. Another advantage is speed; a party can easily cover
2,000 acres a day ; Waddell and the writer having counted the birds
on an area of this size in one morning.

Ropes last for a considerable period, one that has been dragged
over 20,000 acres still being used. This method of counting is not
closely limited to a short season, or to a particular part of the day,
as is the spring count on the courtship grounds. The accuracy of
the rope count is not dependent on special weather conditions or on
other variables over which man has no control, its major advantage
lying in the fact that, when properly used, it gives an accurate
quantitative count in the census area. For that reason, it is most
useful in prairie chicken management.

The rope count is dangerous for the layman to use during the
nesting season and when young birds are small (during the latter
part of March through June), for unless extreme care is taken, nests
may be broken up and young birds injured, scattered, or even killed
by a fast-moving, 1-inch rope. A i^-inch rope, 50 yards long, how-
ever, has been used with success in locating nests and broods without
detriment to the birds. Coffee-bean plants, yaupon bushes, trees,
and old fence posts are obstacles to rope counting, but they can be
avoided without great loss of time, and, with care, few breaks in
the rope or cable result. Of course, cultivated fields cannot be
traversed without injury to standing crops, and counting is difficult
and sometimes impossible in fallow rice fields where levees are high.
Inexperienced persons sometimes have trouble in keeping the proper
amount of slack in the rope and in following the car tracks which
are depended upon to mark the inside margin of every new belt.
These minor difficulties, however, are rapidly overcome by practice.

52 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

CAR-DOG COUNT

A car-dog count is made by driving out an area in belts 40 to
150 yards wide and releasing dogs to work the area where birds are
known or strongly suspected to be. Necessary equipment consists of
one car and one well-trained dog, but two cars, with a man and two
dogs in each, speed up the work more than 100 percent.

Since 1936 this method has been employed in working over 25,000
acres. It is economical and is highly enjoyable to dog lovers, but
it is slow. The work of different dogs and that of the same dog at
different times and under different conditions varies greatly. Dup-
lications in counts or recounts are virtually unavoidable on well-
populated range, and misses are frequent. As reliable data are diffi-
cult to obtain by this method, it should be used only when other census
techniques are impossible.

Using the Census

Censuses produce the most dependable information when applied
over an entire management area. Where this is impracticable and
sampling is resorted to, care must be taken to insure that the selected
area is typical with respect to vegetation, topography, water, and the
like, and is sufficiently large. In a pasture consisting of 60 percent
flat grassland with little or no brush and 40 percent sandhills covered
with live oak shinnery it would be incorrect to sample only the sand-
hill territory and apply the findings to the flat grassland as well.
Sampling should be divided proportionately between distinct environ-
mental types. Sample areas should be at least a section, or 640 acres,
in size, and they should preferably cover 2,000 acres. Thoroughness
should never be sacrificed for extent of coverage, however, as accurate
censuses made thrice annually on a well-chosen section over a period
of several years will yield infinitely more usable data than will hap-
hazard counts sporadically undertaken over more territory than can
be conveniently handled.

During the spring census, investigators should list (1) males, (2)
females, (3) occupied courtship grounds and the number of males at
each, (4) unoccupied courtship grounds, and (5) jack rabbits, in areas
where they are a common resident species, as in the country west of the
Brazos Kiver. Census sheets should show also (1) name of the pasture,
(2) name of owner, (3) size of sample area, (4) exact location, (5)
date of census, (6) counting method used, (7) weather, and (8) names
of the investigators. Additional notes taken should describe (1) the
type of country censused, whether flat grassland with few weeds,
rolling country with scattered myrtle bushes, and so on ; (2) grazing
pressure, whether light, medium, or heavy; and (3) recent burns on
high or low ground, showing the percentage of area burned, whether

ATTWATER'S PRAIRIE CHICKEN 53

burn is general or spotty, and the location of unburned cover. Care-
ful compilation of data greatly facilities accurate interpretation.

If prairie chickens are numerous in the sample area, 1 to every
10 acres or less, cover conditions should not be changed. Instead,
steps should be taken to insure that grazing and burning practices
will duplicate the conditions in future years. If the birds are scarce,
less than 1 to every 10 acres, and if they have not been overshot, habitat
deficiencies should be sought and corrected.

SPRING

Probable habitat deficiencies limiting prairie chicken numbers in
spring are scanty, poorly distributed, or overdense nesting cover
and a shortage of suitable courtship grounds. Common causes
of deficient cover are general burning and overgrazing that result in
scarcity or complete absence over more than 60 percent of the area
of old vegetation, left from previous years at an average height of
at least 5 inches, and poorly distributed cover confined to a particular
part of a pasture or to low, poorly drained situations. Undergrazing
is the usual cause of overdense cover, the thick matted vegetation that
chickens regularly avoid. Where jack rabbits (Taylor, Vorhies, and
Lister, 1936) are a common resident species, they are usually either
very nmnerous, 1 to every 10 acres or less, or entirely absent where
cover is too scanty or too poorly distributed to be suitable for nesting
prairie chickens. Jack rabbits are frequently scarce, 1 to every 80
acres or more, however, in cover that is overdense. A markedly un-
balanced sex ratio, with more than twice as many male prairie
chickens as hens, also has been noted in pastures where nesting cover
was deficient.

In areas in which the cover is scarce because of general burning,
conditions are improved by leaving 40 percent or more of the grassy
cover unburned each year. Unburned cover should be well distributed
over the pasture, the greater part being on the highest, best-drained
ground, in patches of 5 to 40 acres. Favorable conditions are en-
couraged if burning is carried on when there is little or no wind and
the vegetation is slightly damp. A quiet day following a light shower,
or a still night after the dew has begun to fall, is preferable. A test
fire should be set in a protected corner of the pasture. If it burns
slowly, consuming only the most combustible material, and dies down
in 6 to 15 minutes, a series of fires then may be set throughout the
pasture. The number should be strictly regulated by the acreage to
be burned and the manpower available to curb the fires in case of such
unforeseen difficulties as a fresh breeze that may put fires out of con-
trol. The best insurance against trouble from that cause is a plowed
fire lane, 5 to 10 feet wide, completely encircling the pasture and

54 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

dividing it into blocks of approximately 100 acres each. All pasture
burning should be completed by December or early in January, well
in advance of the nesting season.

Moderation in grazing is important. Cattle should not be left to
graze on a tract until unpalatable bunch grasses, as smut grass, salt
grass, big bluestem, and tlie like, are all that remain. Ideal condi-
tions are approached when the number of animal units is strictly regu-
lated according to the quantity and quality of the available forage.
This practice avoids a condition of cover that is scanty in dry years
and overdense in wet years, and, besides being favorable to prairie
chickens, it conserves the soil and the range.

A shortage of courtship grounds, short-grass areas from one-half to
10 acres in extent surrounded by light to medium-heavy grassy cover, is
frequently indicated by an unbalanced sex ratio with more than twice
as many females as males or by a preponderant male population of 8
to 15 or more birds on each booming ground. Common causes of in-
adequate courtship facilities are (1) a lack of hardpan flats; (2)
general burning, which denudes vegetation over a wide area and causes
prairie chickens to leave ; and (3) undergrazing, resulting in tall cover
even on hardpan areas. These deficiencies, however, are remedied by
spot burning and moderate grazing.

SUMMER

Probable deficiencies that limit the number of birds in summer are
an insufficient supply of water in dry years and inadequate shade. A
count made from July 1 through August 10, preferably over the same
area covered in spring, reveals the number of young produced, and
thus serves as a check on the success or failure of the breeding season.

The data recorded for the summer count should be the same as for
the spring count and, also, investigators should note (1) the number
of young, (2) the number, character, and location of water supplies,
and (3) distribution of birds with respect to water and weedy cover.

A larger population of adult prairie chickens than was found in
spring shows either that the spring census was inaccurate, or that
other birds have moved in. In the latter event no habitat manipula-
tion should be attempted unless the resident population plus the in-
flux averages less than 1 bird for every 10 acres, and a larger popula-
tion is desired. If the adult population has decreased since spring,
however, and it is established that poaching has not occurred, the
census data should be examined for information suggesting causes
of the decline.

Indications of deficient water are the absence of watering places a
mile or less apart, and the concentration of birds and jack rabbits in
parts of the area where water is available. Indications of deficient

ATTWATER'S PRAIRIE CHICKEN 55

shade are the scarcity of prairie chickens and jack rabbits where sur-
face water obtains. Ordinary causes of insufficient water are drainage
and unusually dry seasons. Water supplies can be improved by clean-
ing and deepening natural ponds or constructing new ones, and allow-
ing windmill pumps to spill over and maintain puddles nearby.

Common causes of inferior shade are heavy grazing and mowing.
Shade deficiencies usually can be corrected naturally by moderate
grazing and leaving unmowed plots 1 to 10 acres in extent in flats
near ponds. Other methods of obtaining and maintaining tall shading
cover are by (1) fencing tracts from % to 10 acres in size within 100
yards of water holes and leaving the fenced areas ungrazed ; (2) plant-
ing tamarisk, chinaberry, black locust, elm, sycamore, cottonwood, or
other adapted trees near water supplies; (3) constructing two or more
brush racks 5 by 6 by 2 feet high on knolls on high ground near
ponds ; and (4) strip plowing near ponds as outlined on page 56.

An increase of 100 percent in the number of prairie chickens in any
year is excellent. An increase of 50 percent or less may indicate
either a poor breeding season or abnormally high predation. Rainfall
records for May show whether breeding conditions are poor; in the
event heavy rainfall is not the causative agent, predators' may be
responsible. In the latter instance, the number of predatory dogs
and house cats should be reduced by shooting or trapping. If a thor-
ough job is done and yet the increase is small, the aid of State or
Federal wildlife technicians should be solicited.

WINTEB

In winter, a grouse habitat may be deficient in food, cover, or both.
This may best be determined by study of information obtained during
a December or early January reconnaissance of territory that was
covered in summer. Except for the data on the number of young
birds and on the water supply, information recorded in winter should
be the same as that in summer, and it should show whether birds are
generally distributed or heavily concentrated in small areas. The rope
count is the preferred method of winter census, dog counts being made
only when rope counting is impracticable.

Assuming that poaching is not a factor, a winter population larger
than that of the summer, shows habitat conditions on a census area
already more favorable than those in pastures nearby and suggests that
management be directed at maintenance, rather than at alteration of
environment. A winter population smaller than that of the summer
suggests food or cover deficiencies. Other indicators of such inad-
equacies, generally occurring together in heavily grazed areas, are a
prevalence of largely unpalatable plants, as goat weed, marsh-elder,
dogfennel, perennial ragweed, smutgrass {SporobokiS poiretii), and

56 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

cordgrass; a concentration of prairie chickens in small parts of the
area; and an abundance of jack rabbits (averaging 1 to every 10 acres
or less) , or their complete absence. Indications that food only is lack-
ing are a uniform growth of heavily matted grass, a sparse growth
of weeds, and a scarcity of jack rabbits (averaging 1 to every 80 acres
or more). Moderate grazing encourages favorable food and cover
conditions.

To make up for local shortcomings, prairie chicken managers may
establish (1) stockproof -fenced areas of i/^ to 10 acres in as large num-
bers as practicable; (2) plowed strips 20 to 50 feet wide dividing
sparsely inhabited parts of pastures into blocks of 50 to 200 acres; and
(3) unmowed patches of cover of 2 to 10 acres or more, situated not
more than 300 yards apart. Fenced areas should not be grazed, and,
if possible, half of each fenced area should be planted annually to
Schrock, German millet, dwarf milo, hegari (pi, 14) , or red-top cane.
Brush racks built in the corners of fenced areas attract quails as well
as prairie chickens. In average years strip plowing may be done
with satisfactory results from December through April, but February
is considered most favorable. The best effects on experimental areas
in Wliarton County and at College Station have been obtained on
strips that were plowed shortly before or after a rain and harrowed
immediately after the preliminary breaking. Unmowed patches of
cover should be left on knolls or ridges, in flats around ponds, or in
other places where sizable stands of weeds occur.

General Recommenditions for Habitat Control

Landowners who do not undertake intensive management of prairie
chickens based on counts made three times a year may adopt any
or all of the following general recommendations with the assurance
that some improvement will result:

Pastures should be grazed moderately by livestock.

Pasture burning should be completed before February 1 ; in excess of 40
percent of the pasture should be left unburned, with the remaining
cover v/ell distributed in patches of 5 to 40 acres on the best drained
areas.

Mowing should not be done before July 1 ; unmowed patches of 2 to 5
acres or more, not more than 300 yards apart should be left on flats,
knolls, or in other places where there is a good stand of weeds.

In summer windmill pumps should be allowed to form puddles.

Predatory house cats and dogs should be rigidly controlled.

The present 5-year close season (effective September 1937) should be
enforced. If and when the season is reopened, not more than 35 percent
of the known population should be shot when rainfall in May is normal
or less. In years when rainfall in May is approximately twice normal,
no birds should be kUled.

North American Fauna 57. Fish and Wildlife Service

Plate 14

ATTWATER'S PRAIRIE CHICKEN 57

PREDATOR CONTROL

Exhaustive studies by McAtee (1931, 1932, 1935, 1936), McAtee and
Stoddard (1930), Fisher (1893)), May (1935), and others have dem-
onstrated rather conclusively that the food of most flesh-eating birds
and manmials is determined mainly by the availability of prey.
Errington (1935) adds that predation is largely confined to in-
secure or surplus populations; in areas where there is adequate
food and cover, hawks, owls, skunks, opossums, and the like feed
principally on the more common rats, mice, snakes, frogs, and insects,
rather than on the less numerous, swift, and elusive game birds.
Flesh eaters often are neutral, and may be actually beneficial, in
relation to sport and agriculture. Specific data presented on the
relationships of Attwater's prairie chicken and its natural enemies
(pp. 37 to 40) are by no means so comprehensive as desired. The
information at hand, however, apparentl}'^ justifies the following
general recommendations concerning predator control on prairie
chicken range :

Feral house cats and predatory dogs may well be controlled.

Hawks, owls, and fur animals should not be killed Indiscriminately, their
control being limited to known offenders. Inhumane pole traps should
not be used. (All raptorial birds except Cooper's, sharp-shinned, and duck
hawks, goshawks, and great horned owls are protected in Texas.)

Fur animals should be taken only during open seasons when furs are prime,
and the harvest should be regulated to promote sustained yields.

HARVESTING THE SURPLUS

Former laws governing the shooting of Attwater's prairie chickens
left much to be desired. Regulations in effect from 1925 through
1937, providing an open season from September 1 through September
4 and a bag limit of 10 birds a day or 10 a season, actually stimulated
butchery and injured sport. Hunting was allowed when birds were
easily found, many being concentrated near patches of heavy cover
near surface water. Unwary young of the year were easily shot be-
cause they flushed near the gunner, flew straight and slowly for short
distances, and ran but little after alighting. Adults performed sim-
ilarly, probably because the weather was warm, the cover dense, and
because they were in molt. September heat prohibited efficient work
by bird dogs, so crippling losses were doubtless high. As it was
also uncomfortably warm for men to walk, hunting by cars, and
shooting from them, in violation of State law, became the rule in
Colorado and Austin Counties and probably elsewhere in coastal
Texas.

In the future, the power of making regulations might well be
delegated, under proper safeguards, to the State Game, Fish, and
Oyster Commission, which has the benefit of information and counsel

58 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

from its own trained game protectors and from experienced sports-
men, as well as from wildlife specialists connected with the State and
Federal Governments. This power ghould allow for prompt modi-
fication of regulations in response to emergencies of climate, or other
conditions affecting the welfare of the birds. The proper open sea-
son on Attwater's prairie chickens, assuming that the numbers of
the birds can be built up to withstand the drain of shooting, cannot
always be determined in advance, and regulations should be formu-
lated in accordance with local conditions as they develop. The State
legislature, meeting only at 2-year intervals, must rely on the State
Game Commission for appropriate regulation of the take of game;
and only under that arrangement can the people properly hold the
Commission fully responsible for game protection.

If hunting is again allowed, seasons should not in any case open
prior to November 15. Weather late in November is usually suffi-
ciently cool for the comfort of men and dogs, and, normally, the
prairies are too wet to allow hunting from cars. The prairie chickens,
already congregating in winter packs, are widely distributed and
strong flying; consequently, they are hard to find and even more
difficult to hit. Probably because the young of the year are strong
and more worldly wise, and because the weather is cool and the
ground cover reduced, late fall birds regularly flush widely, twist
crazily, fly swiftly and far, usually for a mile or more, and run after
alighting. In other words the Attwater's prairie chicken in Novem-
ber is a game bird of the highest order; hunting it thoroughly tests
the most skillful hunter and the best bird dog. That is as it should
be in true sport.

RESTOCKING

At present the possibility that prairie chickens may be restored
by artificial planting is remote, as wild birds are not available
for trapping and moving, and artificial propagation has shown little
promise. Furthermore, there is no assurance that priarie chickens,
if available, would survive if moved. In Texas and Oklahoma, at-
tempts to transplant lesser prairie chickens have been unsuccessful.
Bent (1932: 263) records the failure of numerous attempts to trans-
plant the greater prairie chicken in northern States. A number of
these birds introduced in the vicinity of the Sault Sainte Marie and
McMillan in northern Michigan persisted for a few years, but F. F.
Tubbs, Michigan Department of Conservation, writes that they have
disappeared. It is true that no intensive efforts have been made to
transplant Attwater's prairie chickens in southern Texas, but there
is no reason to believe that they would survive the process better
than have their relatives.

ATTWATER'S PRAIRIE CHICKEN 59

Natural restocking, however, takeg place rapidly and efficiently
when prairie chickens are properly protected and allowed to increase.
Since 1935 practically all suitable territory in Refugio County has
been restocked by natural spread from the Salt Creek Ranch and
the properties of Martin O'Conner. When the birds are permitted to
increase elsewhere, similar results may be expected.

SUMMARY

Attwater's prairie chicken, a characteristic bird of the coastal
prairie, is one of three kinds that once occurred in Texas. A few
lesser prairie chickens, smaller and paler in color than Attwater's
subspecies, still persist in parts of the Texas Panhandle. The greater
prairie chicken, however, has been entirely extirpated from its former
habitat in the central and northern parts of the State.

An intensive census made in the summer of 1937 revealed that
only about 8,000 to 9,000 Attwater's prairie chickens then remained
in Texas, approximately half of them being in Refugio County.
The birds now inhabit only about 450,000 acres, compared with the
more than 6,000,000 formerly occupied. The numbers of the coastal
prairie chicken have declined 99 percent, and its range has decreasd
more than 93 percent during the past century.

The mating season begins late in January or early in February,
when the males asgemble on short-grass areas early in the morning
and late in the afternoon and boom and otherwise display the mating
urge. Females are attracted to the courtship areas by this activity,
and mating usually takes place there. Prairie chickens are pro-
miscuous. The booming is at a climax in March and ends late in May.

Nests containing eggs have been found from February 25 through
June 17. The peak of the laying period, however, is late in March
and in April. Females build their nests in dry vegetation of the
previous year preferred nesting sites being in good cover in well-drained
areas and within 5 yards of an opening.

The normal rate of laying is 1 egg a day until the average clutch
of 12 is completed, but intervals of 1, 2, and even 3 days are not
infrequent. Subsequent attempts to nest may be made if earlier
nestings are terminated while booming is still in progress. Second
and third nests apparently are made in close proximity to those previ-
ously destroyed, which probably jeopardizes their chances for suc-
cessful termination. The incubation period is 23 to 24 days and
hatching occupies about 2 days more. The peak of the hatching sea-
son is in May. Fertility of the eggs evidently is high. Nest losses
in 1937, however, were 70 percent of 13 nests studied, and those in
1938 amounted to 67 percent of 6 nests.

30.^07° — il 5

60 NORTH AMERICAN FAUNA 5 7, FISH AND WILDLIFE SERVICE

vin 1937, of broods on which accurate counts were obtained, 48
averaged 5.48 birds each. Mortality of young prairie chickens is
highest (about 50 percent) in the first 4 weeks after hatching and
comparatively low (about 12 percent) thereafter, a large share of the
early mortality being in lost chicks. Other known causes of juvenile
mortality include heavy or persistent rains during the brooding,
drowning in rice fields, and depredations by natural enemies. Fam-
ily disintegration, although gradual, begins when the chicks are 6
to 8 weeks old. It is completed after cold northers late in October
and in November induce fall flocking, or segregation of the birds
into flocks according to sex.

Unstable and temporary groups of 5 to 15 birds are common from
September through early October, but singles, pairs, and trios pre-
dominate at that season. As fall passes into winter, flocks become
larger, and in December and January groups containing 35 to 300
individuals have been observed. Singles, pairs, and trios, however,
may be found throughout the year.

Young prairie chickens evidently spend their first 3 weeks within
half a mile of the spot where hatched. Late in May and in June,
both young and adults move to territory where cover providing good
shade is found within half a mile of surface water, there to remain
usually until September. When fall rains and cool weather come
and the fall weeds mature, the birds scatter widely, often becoming
common where scarce or entirely absent at other seasons. Concen-
tration into areas where there is moderate-to-heavy cover and ade-
quate food is evident by November, and populations in favorable
areas fluctuate little from then through spring.

The food of adult prairie chickens is about 85 percent vegetable
matter and 15 percent animal. With young birds the ratio of vege-
table to animal is approximately reversed. Favorite sources of plant
food are ruellia, perennial ragweed, blackberry, doveweed, and sensi-
tive briar. Leading animal foods are grasshoppers and beetles.
Greens (leaves, flowers, buds) are lowest in the diet in November
and December; seeds are taken in the smallest proportions in Jan-
uary, February, and March. Insects are least frequently captured
in November, December, and January.

Important factors limiting prairie chicken abundance include ex-
cessive or persistent rainfall during the nesting season, heavy graz-
ing, excessive pasture burning, agricultural operations, and over-
shooting. Other destructive factors, not generally serious but some-
times locally disastrous, include oil development, drainage, floods,
drought, hurricanes, hail, the spread of woody vegetation into prairie
land, predation, pasture mowing, and possibly disease.

ATTWATERS PRAIRIE CHICKEN 61

Available records from 1925 through 1937 show a positive correla-
tion on unmanaged land between the production of young prairie
chickens and rainfall in May. Good crops of young chickens arie
brought off in years when the rainfall in May is II/2 inches or more
below normal. Fair broods are produced when precipitation in May
is nearly, or only slightly above, normal, while poor crops are probable
when rainfall in May is about twice normal. If the findings in
Colorado County apply to other parts of the coastal country, 2 years
in 5, on the average, are favorable to prairie chicken reproduction^
2 are fair, and 1 is poor. Conditions affecting reproduction are never
the same for the entire range, for a county, or even for different parts
of the same county, because of the scattered character of local rains.
Attwater's prairie chicken is a highly fluctuating subspecies, its
scarcity or abundance depending to a large extent on the precipita-
tion in May.

The annual kill of these birds cannot be intelligently regulated by
such general open seasons and general bag limits as have applied
in Texas in the past, but should be set, when permissible at all, by
regulation by the State Conservation Department on the basis of the
latest detailed information obtainable.

Optimum prairie chicken range apparently consists of well-drained
grassland supporting some weeds or shrubs as well as grasses, the
cover varying in density from light to heavy; and with supplies of
surface water available in summer. In short, diversification within
the grassland type is essential.

Management usually will involve protection against excessive kill-
ing, improvement of food and cover, moderate control of predators,
and wise regulation of the harvest. Responsibility for management
must be assumed by the landowner. Food and cover deficiencies
can best be recognized and their improvement and maintenance as-^
sured by careful counts of the birds on part, or all, of the managed
area at three critical periods in March, July, and December.

To obtain and maintain favorable food and cover, the following
general practices are recommended: (1) Moderate grazing of pas-
tures; (2) completing all necessary pasture burning before February
1 and leaving unburned not less than 40 percent of the best drained
ground ; (3) mowing pastures after July 1 and preserving the native
cover on knolls, around ponds, and in flats; (4) allowing wind-
mill tanks to spill over in summer to increase the supply of surface
water; (5) controlling the numbers of feral house cats and predatory
dogs; and (6) allowing the shooting of not more than 35 percent
of the known prairie chicken population in any year when rainfall
in May is normal or below and prohibiting killing when rainfall in
May is approximately twice normal or above. Hunting seasons
should not open before November 15.

62 NORTH AMERICAN FAUNA 57, FISH AND WILDLIFE SERVICE

Under normal conditions hawks, owls, and fur animals do not
appear to be serious predators on Attwater's prairie chickens, and
should not be killed indiscriminately.

Persons should not request the Government to furnish prairie
chickens for restocking, because there is no surplus for the purpose
and no evidence that the birds can be successfully transplanted.

In the absence of ample reservations for the species all other
favorable factors together cannot be counted on to save the bird
from extinction. Before too late a large tract or tracts of suitable
range should be established as a prairie chicken refuge by the Federal
or State Government.

LITERATURE CITED

AsKiNs, Charles C.

1931. Game bird shooting. 312 pp., illus. New York.
Bailhtt, Plobence Mebriam.

1927. Handbook of birds of the western United States. Ed. 10, rev., 590 pp.
illus. Boston and New York.
Bailet, Vernon.

1905. Biological survey of Texas. U. S. Biol. Survey North Amer. Fauna
25, 222 pp., iUus.
Bbndire, Charles Emil.

1892. Life histories of North American birds with special reference to

their breeding habits and eggs, with twelve lithographic plates.
U. S. Nat. Mus. Special Bull. 1, 446 pp.
1894. Tympanvchus americanus attwateri Bendire. Attwater's or southern
prairie hen. Auk 11 : 130-132.
Bent, Arthur Cleveland.

1982. Life histories of North American gallinaceous birds. U. S. Nat. Mus.
Bull. 162, 490 pp., illus.
Brat, William L.

1901. The ecological relations of the vegetation of western Texas. Bot.
Gaz. 32:99-123, 195-217, 262-291, illus.
Erkinoton, Paxil Lesthe.

1935. Overpopulation and predation: A research field of singular promise.
Condor 37: 230-232.
Fisher, Albert Kenrick.

1893. The hawks and owls of the United States in their relation to agri-

culture.. U. S. Dept. Agr. Biol. Survey (Orn. and Mamin.) Bull.
3, 210 pp., illus.
Gkoss, Alfred Otto.

1930a. The Wisconsin prairie chicken investigation. Amer. Game 19 : 39-40,

50.
1930b. Progress report of the Wisconsin prairie chicken investigation. Wis-
consin Conserv. Comn. 112 pp., illus. Madison, Wis.
Leopold, Aldo.

1931. Report on a game survey of the North Central States. 299 pp., illus.

Madison, Wis.
1933. Game management. 481 pp., illus. New York and London.
May, John Bichabd.

1935. The hawks of North America. 140 pp., illus. Nat. Audubon Soc,
New York.

ATTWATER'S PRAIRIE CHICKEN 63

McAtee, Waido Lee.

1931. A little essay on vermin. Bird-Lore 33 : 381-384.

1932. Confusions of an economic ornithologist. Bird-Lore 34 : 315-321.

1935. Food habits of common hawks. U. S. Dept. Agri. Circ. 370, 36 pp.,

illus.

1936. The Malthusian principle in nature. Sci, Monthly 42 : 444-456.
and Stoddabd, Heobebt Lee.

1930. American raptores and the study of their economic status. C5ondor

32: 15-19.
Mebkill, James Gushing.

1879. Notes on the ornithology of southern Texas, being a list of birds
observed in the vicinity of Fort Brown, Texas, from February, 1876,
to June, 1878. U. S. Nat. Mus. Proc. 1 : 118-173, 1878.
Obebholseb, Habey Church.

1938, The bird life of Louisiana. 834 pp., illus. New Orleans.
Simmons, George Finlay.

1925. Birds of the Austin region. 387 pp., illus. Univ. Texas, Austin.
Streckeb, John Kern.

1927. Notes on the ornithology of McLennan County, Texas. Baylor Univ.
Mus. Spec. Bull. 1, 65 pp.
Stoddard, Herbert Lee.

1931. The bobwhite quail: its habits, preservation, and increase. 559 pp.,

illus. New York.
Taylor, Walter Penn.

1934. Significance of extreme or intermittent conditions in distribution of

species and management of natural resources, with a restatement
of Liebig's law of minimum. Ecology 15: 374-379.
Vorhies, Charles Taylor; and Lister, Paul B.

1935. The relation of jack rabbits to grazing in southern Arizona. Jour.

Forestry 33 : 490-498, illus.
Texas Game, Fish, and Oyster Commission.

1929. Grouse or prairie chicken. Ann. Rpt. 1929 : 86-91, illus. Austin, Tex.
Tharp, Benjamin Carroll.

1926. Structure of Texas vegetation east of the 98th meridian. Univ. Texas

Bull. 2606, 100 pp., illus. Austin, Tex.

f'r^

INDEX

Abundance, 6, 7.
Agricultural factors, 40.
americanus, Tympanuchus, 4.
attwateri, Tympanuchus cupido, 1.

Brood size, 18.

Call, 12.

Cat, house, 38, 39, 40.
Census methods, 47.
Coloration, 4, 5, 6.
Courtship, 10.
cupido, Tympanuchus, 1.

Development of young, 16.
Disease, 36.

Disintegration, family, 19.
Distribution, 2, (map) 3, 7.
Drainage, effect of, 43.

Eagle, bald, 39.

Flocking, 20.
Floods, 35.
Food, 25.

Goshawk, 39.

Grouse, 39.

Growth of young, 16.

Habitat, 30.

control, recommendations, 56.

improvement, 46.

requirements, 30.
Hawk, duck, 39.

ferruginous roughlegged, 39.

Krider's, 39.

marsh, 39.

rough-legged, 39.
Heath-hen, 1.
Hunting, effects, 44.

Increase, 20.

Limiting factors, 31.

Management, 45.
Mating, 10.
Mortality, juvenile, 19.

Nesting, 14.

Nests, predation on, 37.

Overgrazing, effect of, 42.

pallidicinctus, Tympanuchus, 4.
Pasture burning, effect of, 41.

mowing, 43.
Population status, 8.
Prairie chicken, Attwater's, 1.

greater, 4.

lesser, 4, 5.
Predation, 37.
Predator control, 57.
Protection, 45.

Quail, 39.

Rainfall, 32.
Restocking, 58.

Seasonal movements, 21.
Sparrow, English, 17.
Starling, 17.
Storms, 35, 36.
Surplus, 57.

Turkey, bronze, 17.

Tympanuchus cupido americanus, 4.

cupido attwateri, 1, 4.

cupido cupido, 1.

pallidicinctus, 4.

Weight, 5.
Wolf, red, 38.
Woody vegetation, 36.

Young, development of, 16.
predation on, 38.

65

O

OCCIDENTAL COLLEGE LIBRARY

149 30:57 jocus

Attwater s prarne chicken, i/Lehmann, V

3 5043 00322 1677

DATE DUE

Demco, Inc. 38-5

293