HARVARD UNIVERSITY

Library of the Museum of

Comparative Zoology

I

i

i

f

/

OCCASIONAL PAPERS

TPIE MUSEUM Mcz

LIBRARY

TEXAS TECH UNIVERSITY

AUG 1 1 1986

Nl VIBER lOI

HARVARD,

UNIVERSITY

I’ST 1986

CRANIAL VARIATION IN THE BOBCAT (FELIS RUFUS) FROM TEXAS AND SURROUNDING STATES

David J. Schmidly and Jody A. Read

I'he bobcat (Felis rufus) is distributed Irom southern Canada across tiie continental United States into southern Mexico. It occurs throughout Texas and is adapted to a wide variety ol habitats, ranging from limbered swamps to broken forests, rocky or brushy arid lands to mountain ranges, and broken farmland.

Ck)nsiderable geographic variation exists in external and cranial features of the bobcat as evidenced by the recognition of 11 subspecies (Hall, 1981). However, in the 200 years the bobcat has been known to science, there has been no comprehensive assessment of geographic variation in the species, although several partial reviews have treated variation in selected portions of the range. CTrinnel! and Dixon (1924) and Peterson and Downing (1952) studied geographic variation and delineated subspecies in California and in the northeastern United States and adjacent Canada, respectively. Samson (1979) made morphometric comparisons of each of the subspcxies listed by Hall and Kelson (1959), but he made no assessment of geographic variation. Van Zyll de Jong (1974) studied variation in a race of a closely related spcxies, the lynx (Felis lynx snbsolana), in Newfoundland.

Phe present study is lirnitcxl to bobcats from Texas and the surrounding states of New Mexiccj, Oklahoma, Arkansas, and Louisiana. Four of the 1 1 ctirrently recognized subspecies cx cur in this area: F. r. texensis, F. r. baileyi, F. r. rufus, and F. r. floridanus (Hall, 1981); the first two C3f these are kne^wn from

9

OCCASIONAL LAPKRS IHK MIISKLM

Texas. The purpose of this study is to describe patterns of nongeographic and geographic variation in cranial characteristics of bobcats from this region.

Materials and Methods Measurements and Samples

A total of 956 specimens (595 males and 361 females) was examined. Animals were obtained from two different sources. The I’exas Parks and Wildlife Department provided 506 skulls that were obtained from trappers required to register their pelts during the 1978-79 trapping season; these specimens are deposited in the Texas Cooperative Wildlife Collections at Texas A&M University. The remaining 450 specimens were from the following museum collections: American Museum of Natural History, New York City; Zoology Department, University of Arkansas, Fayetteville; Museum of Natural and Cultural History, Oklahoma State University, Stillwater; Stovall Museum of Science and History, University of Oklahoma, Norman; Museum of Southwestern Biology, University of New Mexico, Albuquerque; Museum of Zoology, Louisiana State University, Baton Rouge; Midwestern University, Wichita Falls, Texas; Texas Cooperative Wildlife Collection, Texas A&M University, College Station; National Museum of Natural History, Washington, D.C.; The Museum, Texas Tech University, Lubbock.

Twenty-six cranial measurements were recorded from each specimen with dial calipers to the nearest .05 mm as follows: GSL, greatest skull length (greatest length of skull including incisors); ZB, zygomatic breadth (greatest width across the zygomata measured parallel to the long axis of the skull); SB, squamosal breadth (greatest width across squamosal immediately posterior to zygomatic arch and parallel with long axis of skull); POC, postorbital constriction (narrowest width of skull immediately posterior to the postorbital process measured at right angles to the long axis of the skull); LN, length of nasals (greatest distance from the anteriormost projection of nasal bones to the posteriormost projection along their medial suture); IC, interorbital constriction (the least width between the orbits); DUCk diameter of upper canine (diameter from anteriormost point to posteriormost point of tooth measured at the alveolus); PM, length of upper fourth premolar (greatest distance from anteriormost point to posteriormost point of P4 measured at

S( ;n \I 1 1) L^ A N I ) R K A D— ( ;R A N 1 A 1. \’A R 1 A HON

alveolus); VVR, width ol rosiruni (greatest width across losttum ininiediaiely anterior to zygoinatie plate); PM I’, pieniolar-iiiolar toothrow length (least distance Irom anterior lip ol alveolus ol P3 to the posterior lip ol M2); WMI', width across maxillary toothrow (measured at junction between P3 and Ml); Md’, maxillary toothrow length (measured from anteriormost edge ol canine to posteriormost point of M2); PL, palatilar length (measured from a point inmu'diately posterior to the incisors to the posterior edge of the palatine bone at its medial suture); Cdf, cranial height (least distance from dorsalmost portion of skull to ventralmost portion of {palatine); MB, mastoid breadth (greatest width of skull across the mastoid region); BL, basilar length (greatest distance from anterior surface of the premaxillae {protruding from between the incisors to the anteriormost lifp of the foramen magnum); CBL, ccpndylcpbasal length (greatest distance from anterior surface of premaxillae to {posteriormcpst {portion cpf the cpcci{pital ccpndyles); LM, length cpf mandible (measured frcpm {pcpstericprmcpst {pcprtion cpf ccpndylcpid {prcpcess tcp antericprmcpst {PCPrtion cpf inciscprs); CRH, ccpicpncpid height (measured frcpm the dcprsalmcpst {pcprticprp cpf the ccpronoid {process tcp the ventralmcpst {porticpn of the angular {Ptcpcess); DM, length cpf anterior mandibular tcpcpthrow (measured frcpm antericpr surface cpf 13 tcp antericpr alveolus of Ml); PPL, {pcpst{palatal length (measured frcpm antericprmcpst edge of fcpramen magnum tcp the {Post{palatal ncptch); NFI., nasal-frcpntal length (measured frcpm antericpr {porticpn cpf nasal bcpne at its medial suture tcp the right cpibital ccpnstricticpn immediately {pcpstericpr tcp the {pcpstcprbital {prcpcess); C.LB, greatest length cpf auditcpry bullae (measured frcpm {pcpstericprmcpst {pcpint cpf {parcpc c i{pital {prcpcess tcp the antericprmcpst {pcprticpn cpf the auditcpry bullae); WAB, width cpf the auditcpry bullae (measured acrcpss bullae frcpm {pcpstericpi edge cpf the extetiial auditcpry meatus at a right angle tcp the Icpng axis cpf the skull); DFM, diameter cpf fcpramen magnum (greatest distance frcpm the interli{p cpf the antericpi rncpst {porticpn of the Icpramen magnum to the li{p of the {pcpstericprmcpst {pcprticpn cpf the' fcpramen magnum); HOB, height of cpcci{pital bcpnc' (itic-asurc'd frcpin {pcpstericpimcpst li{p cpf foramen magnum tcp clorsalmcpst {pcpint cpI lamboiclal ridge).

S{Pt'cimens were examinc'cl from a{P{)rcpximately 331 Icpcaliiies. Fhese were grcpu{ped intcp (57 sam{ples fcpi {pur{pcpses of statistical analysis (see Fable 1). All majcpr regicpiis cpf Fexas were re{presented by sam{ples with the exre{Pticpn cpf the western half cpf the High Plains and the extreme ncpitheastern ccprner cpf the state

1

OCCASIONAL PAPKRS I HK MUSKIIM

I ABLE I. — Designatiotis, locations, and number of individuals for samples of male (M) and female (F) bobcats used in the univariate and multivariate analysis of

geographic variation.

S.miplc

no. Ciouiity oi pnii.sh ,S;nnpl( si/c

Texas (Males)

1	C4ierokee, Hardin, Liberty, Montgomery, Polk,
	San Jar into. Walker	13
2	Graysoti, Hopkins, Lamar	6
3	Burleson, Grimes, L.eon, Madison, Robertsoti	22
4	Brazoria, Harris, Jackson, Matagorda	14
5	Bastrop, Colorado, Fayette	6
6	Falls, McLennan, Milam, Williamson	13
7	Archer, Clay, Cooke, Montague, Wise, Young	1 1
8	Boscpie, Eastland, Eratb, Palo Pinto, Stephens	7
9	Brown, Burnet, Mills	5
10	Atascosa, Bexar, Goliad, Karnes	6
1 1	Aransas, Calhoun, Kleberg, Ntieces, Refugio	12
12	Brooks, Cameron, Kenedy	10
13	Bee, Duval, Jim Wells, lave Oak, McMullen	17
14	LaSalle, Webb	13
\5	Frio, Maverick, Medina, lAalde	22
16	Edwards, Gillespie, Kerr, Kimble, Llano, Mason
	Menaid, Sc hleicher, Sutton	15
17	Goleman, Nolan, lorn Greeti	6
18	Baylor, Gottle, Haskell, King, Knox, Wichita
	VVdlbarger	12
19	Armstrotig, Briscoe, Ghildress, Collingsworth	12
20	Hartley, Hemjrhill, Hutchinson, Moore, Ochiltree,
	Ratidall, Roberts, Wheeler	14
21	Bordon, Crosby, Dickens, Garza, Ffoward, Mitchell,
	Scurry	18
22	Cilasscock, Reagan, Upton	12
23	Crockett, Val Verde	10
24	Brewster, Pecos, Ferrell	13
25	Culhersoti, Fludspeth, Jell Davis, Reeves	7
	Texas (Females)
26	Andersoti	9
27	Chambers, Galveston, Jefferson	4
28	Vic toria, Wharton	10
29	Refttgio, Nueces, Kleberg	10
30	Starr, Cameroti	5
31	Webb, LaSalle, Duval	12
32	Me Mullen, Live Oak, Duval, Jim Wells	13
33	Caldwell, LaVaca	8
34	Brazos, Washington, Robertson, Madison	1 1
35	Cloryell	10
36	Blatuo, McCulloch, San Saha	10
37	Sierlitig	12

SCMMini.V AND Rl' AD— CRANIAL V'ARIAI ION

Iabi.k 1, — (Continued.

	(5aiu', Presidio, Ward	11
39	Kent	9
10	Callahan, laylor	8
n	Jac k, Parker	10
12	1 lardeinan	7
13	Motley	5
M	Armstrong, Ciol 1 ingsworth	8
If)	Cray, Potter	1 1
10	Zavala	12
	Louisiana
17	Nalt hilodies	2.M,	4F
18	Vermilion, West Baton Rouge, Iberville	6M
19	Hast Baton Rouge, Livingston, St. Helena,
	Langipahoa	2M,	3F
50	Catahoula, Cioneordia, Madison, Lensas	6M,	7F
	Arkansas
51	Arkansas, Ashley, Cross, Drew, Jackson, Lawrence,
	Lee, Phillips, Randolph	8M,	I3F
52	Bradley, Calhoun, Hemstead, Sevier	22 M,	5F
53	Cdark, Conway, Franklin, Garland, Hot Springs,
	Howard, Johnson, Montgomery, Nevada, Perry Polk, Pope, Pulaski, Scott, Sebastian, Yell	3LM,	I5F
51	Boone, Carroll, I/ard, Madison, Newton, Stone,
	Van Buren, Washington	36 M,	16F
	Oklahoma
20	Cimarron, Beavei	IM,	IF
55	•Atoka, l.atimer, Le Flore	8M,	5F
50	Bryan, Cdevelaiul, Cieek, Hughes, Lincoln, Logan,
	Lore, Mc(dain, Oklahoma, Osage, Pawnee, Pottawatomie, Pontotoc , Stephens, Washington	I5.M,	lOF
57	(iotnanche, Kiowa, 1 illman	IM,	IF
58	(laclclo, (ianaclian, Dewey, (.arlield, Majoi
	Roger .Mills	IM,	3F
	New Mexico
59	Cdiavcv, Dona Ana, Fddy, Lea, Lincoln, Otero	31 M,	27 F
00	Catron, (»rant, Hidalgo, Siena, Socoiro	13 M,	I8F
01	Bernalillo, (iollax, Mora, Arriba, .San Juan,
	■Sandoval, .Santa Fe, Faos, V'alencia	30. M,	16F
02	DeBaca, C.uadalupe, Harding, Quay, .San .Miguel,
	Union	15.M,	lOF
	.Arizona
63	Graham, Pima, A tuna	9.M.	IF
04	(ioconino, Mohave, Aavapai	7M,	3F

OCCASIONAL PAPERS EHK Ml’SEUM

(J

Table 1. — (Continued.

Florida

65 Dade, Pinellas, Puinian 3M

Georgia

66 Marion, 'I'alhoi 7M

67 Ihonias 3M

(Fig. 1). Slightly different sample groupings (1-25 for males; 26- 46 for females) had to be formed for the two sexes in Texas because specimens came from such a variety of localities. Sample groupings (and their attendant numerical designations) are the same for both sexes in Louisiana, Arkansas, Oklahoma, and New Mexico. In addition to the material from the study area, two reference samples of F. r. bailey i from Arizona (63, 64) and three samples of F. r. floridanus from Florida (65) and Georgia (66, 67) also were included.

Aging Technique

Each specimen was assigned to one of five age categories using the technique of Grinnell and Dixon (1924). In addition, one upper canine from each specimen collected by the Texas Parks and Wildlife Department was extracted and aged using the modified technique of Linhart and Knowlton (1967). LIsing this method, Blankenship (1979) was able to provide the estimated age in months for each Texas specimen. This information was combined with the aging technique described by Grinnell and Dixon (1924) to form the five age categories described in Table 2.

Statistical Analysis

Individual, age, and sexual variations were assessed in two samples of F. r. texensis from Texas, one of 64 individuals from Robertson-Madison counties (sample 3 of males and 34 of females) and another of 57 individuals from Webb-Duval counties (sample 14 of males and 31 of females), using the Statistical Analysis System (SAS) designed and implemented by Barr et al. (Helwig, 1976). Means were calculated for each character in the two samples. Analysis of variance (ANOVA) and Duncan’s Multiple Range Mean Test (DUNCAN) were employed to evaluate significant differences among age classes with sexes

s ( : 1 1 M 1 1 ) 1 A . \ N n R !• /\ I ) — ( ; R A N 1 . \ 1. \ ’. \ R I I I () N

7

Fig. I. — Cieographic lofaliiics foi fcniali* (lop) and male (hoiioin) Feli.s rufus from Texas, New Mexico, Oklahoma, Arkansas, and Louisiana. Ciionped samples used in the statistical analysis are outlined and nundrered. .See Fable 1 for designation and site of each sample.

8

o(;c;asi()nal papers i hk mi’skum

I'abi.e 2. — The five age classes, as defined by morphological characteristics and tooth ring counts, used in the study of nongeographic variation.

Age ( la.s.s	Age ill moiith.s'	Cranial and denial t harai ler.s
lA	0-9	basioccipilal suture open; palatal sutures not tightly lused; temporal ridges undetectable; molar teeth just breaking gum line
IB	9-24	basioccipital suture fused but highly visible; palatal sutures fused tightly but visible; temporal ridges visible but not pronounced; muscle attachments on frontals not detectable
II	24-31	basioccipital suture tightly fused but still visible; tem¬ poral ridges well defined but not pronounced; muscle attachments on frontals detectable
III	31-36	basioccipital suture no longer visible; temporal ridges sharp; muscle attachments on frontals well developed
IV	36-48	temporal ridges pronounced; muscle attachments on frontals highly developed with distinct valley between frontals
V	48+	all ridges and muscle attachments highly pronounced with a massive coarse bone siriu ture

'Derived from tooth ring counts by Blankenship (1979).

separated. A Mest was used to test for significant differences between sexes of the same age class in both samples. Coefficients of variation (CV) were calculated to determine the extent of variability for each character.

Geographic variation was analyzed using a variety of univariate statistics (mean, range, standard deviation, standard error, and coefficient of variation) calculated for each character in every sample. Statistical analysis employed ANOVA and DUNCAN test for all samples treating males and females separately.

lb assess clinal patterns of variation, north-south and east- west transects were constructed among selected samples, and modified Dice-Leraas diagrams (Dice and Leraas, 1936) were drawn for four variables in males (greatest length of skull, squamosal breadth, cranial height, and length of the upper fourth premolar). A single north-south transect was established as follows; Transect A, from Oklahoma Panhandle to southern tip of I'exas (samples 20, 19, 18, 17, 16, 10, 13, 12). Four west-to-east transects were established as follows: Transect B, from western Texas to eastern Arkansas (samples 25, 21, 18, 7, 2, 55, 53, 51); Transect C, from northwestern Arizona to southern Texas (reference sample 64; samples 60, 59, 25, 24, 23, 15, 13, 11);

SCHMini.V AND RKAD— CRANIAl. \'ARlAri()N

Transect I), irDiii northwesiei n Arizona to souih-tentral Florida (relertMue sample 63; samples 60, 59, 21, 18, 7, 2, 52, 51; and retereiue sample 66); and Transect T, from northwestern Aiizona to eastern Arkansas (reference samjile 61; samj)les 61, 62, 20, 58, 56, 53, 51).

Several multivariate statistical technicjues, using the Numerital Faxonomy Program (NF-SYS) of Rohlf and Kishpaugh (1972), were em})loyed to cluster samples according to phenetic affinity. (Muster analysis was performed on standardized character means using average taxonomic distance as a measure of similarity and the UPCiMA cluster option. Fhe first three principal components were extracted from a matrix of correlation among characters, and projections of the samples onto the first two components were made.

Results

Nongeographic Variation

d'he pattern of individual, age, and secondary sexual variation was almost identical in the two samples of F. r. texensis; therefore, only the data for the Robertson-Madison County sample are presented and discussed. No individuals of age class lA were included in this sample; therefore, all reference to age class I is to individuals in category IB (see Table 2).

Age variation. — ANOVA and DUNCAN test revealed that significant differences exist in most measurements among age classes of males. Generally, suhaclults (age class I) were significantly smaller than were older adults (age clas.ses II, III, lY, \') in most measurements. Age class I in males completely separated from the other age classes in 18 of 26 characters ( Fable 3). In breadth of scjuamosal, males of age class I were significantly different (F = 3.55; p <.()5) from age class but not age classes II, III, and IV. In length of the maxillary toothrow, age class I males differed significantly {F = 8.93; p <.0001) from age classes III, IV, and \' hut not age class II. In mastoid hteadth of males, age classes I and II .separated distinctly {F = 1.38; p <.002) ftom age classes III, IV', and V. I here were no significant differences among age classes in lour characters of males (postorbital constric tion, length of uppc'i fourth premolai , premolar-molar toothrow length, and diameter of foraitien magnum).

Fhere was little significant age variation among females. Except for width of rostrum and length of mandible, none of the

10

()C:CASI()NAL. PAPERS EHE MUSEUM

Table 3. — Variation with age in cranial measurements of Felis rufus from Robertson and Madison counties, Texas. Age classes are defined in Table 2. Statistics given are number, mean, two standard errors of mean, range, coefficient of variation, F, and Fs. Symbols alongside age classes indicate nonsignificant

subsets according to DUNCAN test.

Srx anti Duncan

ago ( lasses N Mean ± 2 sk Range CIV F/F, resiills

Greatest Length of Skull

Male
IV	8	129.94 ± 3.98	1 18.45-137.60	4.30
V	5	128.71 ± 5.80	1 19.50-135.05	5.04	12.18
III	5	128.48 + 4.04	122.95-133.50	3.52	0.0001
II	2	125.05 ±11.80	119.15-130.95	6.67
I	7	109.10+ 5.98	99.20-123.60	7.30
Female
II	3	122.75 ± 6.20	119.40-128.95	4.38	0.63
\'	2	121.80 ± 17.30	113.15-130.45	10.04	ns
III	9	121.00+ 3.60	119.20-122.80	2.10
I	5	117.42+ 3.40	112.60-123.25	3.23

Zygomatic Breadth

Male
IV	9	89.92 ±	2.54	82.45-95.50	4.25	26.10
III	6	88.80 ±	2.32	84.15-91.80	3.20	0.0001
V	5	88.21 ±	3.26	82.45-91.90	4.13
I	12	75.51 ±	2.98	66.95-86.10	6.83
Female
V	2	83.,55 ±	5.. 50	80.80-86.30	4.66	1.76
III	3	83.17 ±	2.94	81.70-86.10	3.06	ns
II	4	82.64 ±	4.68	76.80-88.15	5.65
I	7	78.29 ±	3.26	72.10-82.50	5.45

Squamosal Breadth

Male

V	4	54.54 ±	0..58	53..50-.55.2.5	1.18	4.65
II	2	.54.00 ±	3.20	52.40-55.60	4.19	0.0057
III	6	53.88 ±	0.96	.51.90-.55.20	2.18
IV	9	53.47 ±	1.12	.50.20-.55.20	3.13
I	9	.52.. 50 ±	1.04	49.50-54.05	3.02
Female V	2	52.98 ±	1.46	52.25-53.70	1.94	0.44
II	3	52.77 ±	1.74	52.45-.54.40	2.84	ns
III I	1 5	52..50 51.74 ±	1.48	49.6.5-53.70	3.20

SCHMIDIA AND RKAD— CRANIAL N'ARIAI ION

I'abi.k 3. — Cotitmued.

Postorbital Constriction

Male
III	6	37.86 ±	2. 12	35.65-42.95	6.87	0.69
I	10	37.38 ±	1.36	32.80-39.70	5.72	ns
\'	5	36.98 ±	2.22	34.20-39.50	6.72
II	2	36.78 ±	0.06	36.75-36.80	1.09
1\'	9	36.02 ±	1..52	31.90-38.40	6.35
Female
II	3	37.98 ±	1.34	37.10-39.30	3.06	1.91
\’	2	37.03 ±	2.86	35.60-38.45	5.44	ns
I III	5 1	3.5.46 ± 34.,5,5	1.60	33.70-38.20 Length of Nasals	.5.06
Male
II	3	27. .52 ±	1..52	26.00-28.35	4.78	6.83
	6	27.06 ±	1.94	24. .50-3 1.20	8.74	0.0003
I\’	13	26.87 ±	1.16	23.90-31.45	7.72
III	8	25.89 ±	l.,50	23.25-28.85	8.15
I	13	23.55 ±	0.78	22.20-27..50	5.98
Female
III	3	25.48 ±	1.44	24.25-26.75	4.91	0.19
\'	3	25.47 ±	2..52	23.25-27.60	8.. 55	ns
I	6	25.01 ±	1.08	23.35-28.15	6.43
II	5	24.54 ±	2.. 54 21.40-27.60 Diameter Upper Canine	11.60
Male
I\’	13	7.67 ±	0.34	6.75-8.70	7.87	6.53
\'	6	7.66 ±	0.74	6.40-8.70	1 1.84	0.0005
II	3	7.20 ±	o:98	6.35-8.05	1 1.80
III	8	7.04 ±	0.24	6.40-7.60	4.90
I	10	6.35 ±	0.46	5.00-7.85	1 1.45
Female
\’	3	7.10 ±	0.80	6.40-7.80	9.86	1.56
II	f)	7.03 ±	0.24	6.70-7.45	3.92	ns
III	3	7.02 ±	0.68	6.65-7.70	8.44
I	9	6.56 ±	0.34 5.60-7.30 Interorbital Constriction	7.83
Male
	0	24.51 ±	0.86	23.30-26.00	4.29	19.41
I\'	1 1	23.90 ±	0.97	21.65-26.35	6.52	0.0001
III	8	23.25 ±	1.20	21.20-25.10	7.25
II	3	22.38 ±	1.26	21.15-23.25	4.90
I	13	19.14 ±	1. 00	16.40-23.35	9.39

12

OCCASIONAL PAPERS THE MUSEUM

Table 3. — Continued.

Eemale
V	2	23.55 ±	2.30	22.40-24.70	6.91	3.02
II	4	22.69 ±	1.70	21.05-24.65	7.50	ns
III	3	22.10 ±	1.84	20.50-23.70	7.24
I	9	19.91 ± 1.50 15.70-22.85 11.35 Length of Upper Fourth Premolar
Male
V	6	14.28 ±	1.04	12.35-15.55	8.88	1.26
IV	13	14.04 ±	0.32	13.30-14.80	4.11	ns
II	3	13.97 ±	1.10	13.00-14.90	6.80
III	8	13.96 ±	0.48	12.95-15.15	4.85
I	12	13.45 ±	0.54	10.90-14.60	6.87
Eeinale
III	3	14.02 ±	0.70	13.40-14.60	4.29	2.. 59
II	5	13.94 ±	0.66	13.10-14.75	5.26	ns
V	3	13.25 ±	0.76	12.50-13.65	4.91
I	8	13.16 ±	0.36	12.30-13.90 Width of Rostrum	3.95
Male
IV	13	33.07 ±	1.12	29.70-36.55	6.08	8.85
V	5	33.01 ±	1.70	30.95-35.55	5.75	0.0001
III	8	32.56 ±	1.24	28.60-33.85	5.40
II	3	31.88 ±	3.00	29.40-34.60	8.18
I	14	29.29 ±	0.82	26.60-32.05	5.23
Eemale
V	3	32.03 ±	1.44	30.70-33.20	3.93	6.49
II	5	31.84 ±	1.20	30.15-33.50	4.28	0.0044
III	3	31.23 ±	1.22	30.10-32.20	3.39
I	9	29.56 ± 0.62 28.25-30.75 3.22 Premolar-Molar Toothrow Length
Male
V	6	25.33 ±	1.92	21.75-28..50	9.24	1.72
IV	13	25. 1 1 ±	0..54	23.85-26.50	3.84	ns
III	8	24.84 ±	0.80	22.50-26.20	4.60
II	3	24.62 ±	1.42	23.25-25.60	4.99
I	14	23.89 ±	0.78	21.35-27.30	6.14
Eemale
II	5	24.79 ±	1.20	23.35-26.45	5.40	0.64
III	3	24.32 ±	0.02	24.30-24.35	0.12	ns
I	9	24.16 ±	0.45	23.00-25.80	3.50
V	3	23.90 ±	1.32	22.65-24.90	4.79

SCHMIDI.V AND RKAD— C:RANIAI. X'ARIAHON

l.'j

r A B 1 .K 3 . — Contniued.

Width Across Maxillary Toolhrow

Male
IV	13	38.30 ±	1.00	34.40-42.40	4.73	4.45
\'	(i	37.97 ±	1.60	35.40-40.45	5.14	0.0048
111	7	37.80 ±	1.10	34.85-39.35	3.91
11	3	37.32 ±	2.20	35.30-39.10	5.12
1	11	3.’i.26 ±	1.32	30.00-38.74	7.01
Ft’iiuile
11	5	37.32 ±	1.48	35.70-39.95	4.44	1.65
\’	3	37.15 ±	1.96	35.20-38.20	4.55	ns
111	3	36.62 ±	1.44	35.25-37.70	4.31
1	9	35.79 ±	0.74 34.55-38..50 3.09 Length of Maxillary Toothrow
Male
IV	13	39.18 ±	0.88	36.50-41.00	4.08	5.66
V	6	38.82 ±	2.66	34.25-43.65	8.42	0.0011
III	8	38.13 ±	1.14	35.40-39.80	4.25
II	3	38.12 ±	3.06	35.20-40.40	6.97
I	14	3.5.75 ±	0.94	31.95-38.10	4.88
Female
II	5	37.52 ±	2.06	35.40-41.00	6.16	0.40
V	3	36.92 ±	2.06	35.05-38.60	4.83	ns
III	3	36.83 ±	1.06	36.30-37.90	2.. 50
I	9	36.52 ±	0.90	34.15-38.60 Palatilar Length	3.67
Male
IV	13	49.82 ±	1.40	45.40-.52.65	5.04	1 1.71
V	6	48.61 ±	3.18	44.00-53.80	8.01	0.0001
III	7	48.28 ±	2.30	43.90-51.75	6.32
II	3	47.27 ±	2.90	45.65-50.15	5.30
I	13	42.54 ±	1.50	37.70-45..50	6.35
Female
V	3	46.87 ±	2.33	44.55-48.70	4.52	1.06
III	3	46.35 ±	0.76	45.95-47.10	1.40	ns
II	5	46.34 ±	2.76	42.25-50.90	6.67
I	9	44. .52 ±	1.70	40.75-49.00 Ciranial Height	5.69
Male
I\'	12	45.45 ±	1.70	41.00-.50.70	6.45	1 1.96
III	5	45..50 ±	2.26	41.60-47.80	.5.. 55	0.0001
V'	6	45.09 ±	1.94	42.25-48.00	5.27
II	3	42.78 ±	1.84	41. 15-44. .35	3.74
I	12	39.20 ±	1.26	35.90-42.95	5..57

OCCASIONAL PAPERS 1 HE MUSEUM

1’abi.e 3. — CoJitinued.

Eemale
V	2	45.03 ±	6.14	41.95-48.10	9.66	1.68
III	3	42.50 ±	2.66	40.45-45.00	5.43	ns
II	4	42.34 ±	3.60	40.40-46.40	6.45
I	8	40.43 +	1.84	35.56-42.70	6.43
				Mastoid Breadth
Male
IV	9	56.48 ±	1.18	53.70-58.90	3.16	13.24
III	6	55.73 +	1.30	54.00-58.45	2.98	0.0001
V	5	55.14 ±	1..50	52.60-56.75	3.03
II	1	.50.70
I	8	50.48 +	1.58	47.75-53.90	4.42
Eeinale
II	3	52.97 ±	3.60	50.10-55.30	5.90	0.37
III	1	52.45				ns
V	1	49.85
I	5	48.82 ±	5.98	37.40-55.00	13.67
				Basilar Length
Male
IV	9	108.11 ±	1.92	103.80-113.05	2.67	13.93
V	5	106.38 ±	5.20	96.95-112.60	5.45	0.0001
III	5	106.22 ±	3.66	100.45-110.85	3.85
II	2	102..58 ±	8.20	98.65-106.50	5.41
I	6	90.70 ±	5.08	81.70-99.60	6.87
Female
III	2	102.00 +	2.40	100.80-103.20	1.66	0.83
II	3	101.20 ±	5.52	97.50-106.60	4.72	ns
V	2	100.50 ±	12.80	94.10-106.90	9.01
I	4	95.40 ±	6.30	87.45-102.80	6.60
			Condylobasal Length
Male
IV	8	118.45 ±	2.52	1 12.55-124.30	3.02	12.69
V	5	117.63 ±	5.10	108.70-123.75	4.85	0.0001
III	5	117.32 ±	3.90	110.45-120.90	3.73
II	2	1 13.00 ±	8.20	108.90-117.10	5.13
Female
II	2	1 14.20 ±	8.80	109.80-118.60	5.45	1.12
III	2	1 11.83 ±	1.36	111.15-112.50	0.85	ns
V	2	111.38 ± 12.66	105.05-117.70	8.03
I	4	105.21 ±	6.38	97.10-112.60	6.05

SCiUMlDI.V AND RKAD— CR AN lAI . VARIATION

I ABi.K. 3. — Continued.

Length of Mandible

Male
IV	1 1	76. bf) ±	2.31	69.35-81.00	5.08	1 1.32
\	b	76.27 ±	3.60	68.95-79.85	5.86	0.0001
III	7	71.16 ±	3.01	68.65-78.25	5.13
II		73.00 ±	■1.76	70.55-77.75	5.63
I	13	65.68 ±	2.66	56.80-73.75	7.28
Female
\'	2	74.58 ±	1.06	71.05-75.10	1.00	3.06
III	3	71.33 ±	1.56	69.80-72.35	1.89	0.0629
II	5	51.33 ±	3.10	66.85-75.70	1.85
I	8	68.13 ±	2.30	62.15-71.35	1.77

Height of Coronoid

Male
IV	12	38.57 ±	1.80	33.25-42.55	8.07	8.31
\’	5	37.17 ±	1.26	32.30-44.85	12.82	0.0001
III	6	36.23 ±	1.98	33.10-39.15	6.68
II	3	35.92 ±	2.72	33.35-38.00	6.58
I	13	31.63 ±	1.56	26.90-37.15	8.90
Female
	2	36.30 ±	1.70	35.45-37.15	3.31	2.12
III	3	35.73 ±	2.00	34.10-37.70	4.87	ns
II	5	33.96 ±	2.16	30.75-38.20	8.12
1	8	32.17 ±	1.58	28.60-35.00	6.91

Length of Anterior Mandibular Toothrow

Male

I\'	1 1	17.06 ±	0.66	15.10-18.55	6.31	10.32
V	6	16.35 ±	0.70	15.20-17.30	5.23	0.0001
II	3	16.05 ±	1.62	11.90-17.60	8.68
III	7	16.01 ±	0.58	15.00-17.15	1.72
I	13	14.58 ±	0.50	12.75-16..55	6.29
Female
\’	3	16.18 ±	0.60	1 5.60- 1 6.55	3.16	2.79
II	5	16.00 ±	1.12	11.60-17.95	7.81	ns
III	3	15.67 ±	0.11	15.10-16.10	2.12
I	8	11.83 ±	0.56	13.70-15.90 Postpalatal Length	5.30
Male
IV	9	57..59 ±	1.06	55.50-61.25	2.75	15.96
V	5	57. 1 1 ±	2.02	53.20-58.95	3.91	0.0001
III	5	57.06 ±	2.20	51.20-61.00	1.31
II	2	54.78 ±	3.96	52.80-56.75	5.10
I	7	19.81 ±	1.76	47.70-.51.25	1.68

16

OCXJASIONAL PAPERS I’HK MUSEUM

Eable 3. — Continued.

Female

III	2	55.43 ±	3.66	53.60-57.25	4.66	0.76
V	2	54.08 ±	7.96	.50.10-58.05	10.40	ns
II	3	54.00 ±	1.92	52.80-55.90	3.08
I	4	51.46 ±	3.34 46.90-54.60 Nasal-Frontal Length	6.48
Male
V	6	54.61 +	2.78	.53.30-61.95	5.90	10.72
III	8	56.82 ±	2.00	53..55-60..50	5.00	0.0001
IV	13	56.61 +	1.76	51.05-61.15	5.63
II	2	55.77 +	4.66	53.05-60.40	7.23
I	11	49.76 ±	1.62	45.70-54.15	5.37
Female
II	5	.54.97 ±	1.36	53.40-57.45	2.78	1.87
V	2	53.15 +	7.70	49.30-57.00	10.24	ns
III	3	53.02 ±	2.40	51.05-55.20	3.93
I	9	51.46 +	1.82 47.45-56.05 Greatest Length of Bullae	5.30
Male
IV	10	32.06 +	1.08	29.50-34.40	5.33	14.32
V	5	31.21 +	0.86	30.00-32.40	3.08	0.0001
III	8	31.11 ±	1.14	28.80-32.80	5.20
II	3	30.63 ±	2.02	29.15-32.55	5.68
I	10	27.14 ±	0.98	24.70-29.55	5.73
Female
II	3	30.37 ±	1.86	28.70-31.90	5.28	0.72
V	2	29.72 ±	5.56	26.95-32.50	13.20	ns
III	2	29.10 +	0..50	28.85-29.35	1.22
I	6	28.39 ±	1.42	25.45-30.30 Width of Bullae	6.12
Male
IV	10	15.78 +	0..52	14.65-16.90	5.1 1	9.60
III	8	15.43 ±	0.44	14..55-16.10	3.96	0.0001
V	5	15.22 ±	0.60	14.60-16.30	4.46
II	3	15.13 ±	0.54	14.60-15.45	3.07
I	12	14.10 ±	0.36	13.15-15.15	4.46
Female
III	2	14.85 ±	0..50	14.60-15.10	2.38	0.15
I	6	14.60 ±	0.94	12.90-15.70	7.90	ns
V	2	14. .58 ±	2.24	13.45-15.70	10.92
II	4	14.28 ±	0.76	13.60-15.30	5.26

SCHMIDI.V AND RKAD— CRANIAL VARIAI ION

17

I abi.k 3. — (luntinued.

			Diameter Foramen Magnum
Male
\'	b	1 3.69 ±	0.54	13.25-14.75	4.44	0.87
I\’	9	13.52 ±	0.46	12.55-14.55	4.98	IKS
III	6	13.10 ±	0.68	12.65-15.00	6.24
I	7	13.39 ±	0.84	11.75-14.80	8.44
II	2	12.38 ±	1.96	1 1.40-13.35	11.14
Female
	2	13.98 ±	0.06	13.95-14.00	0.25	3.20
II	3	13.80 ±	0.26	13.60-14.00	1.45	ns
III	2	12.48 ±	1.96	1 1.. 50- 13. 45	1 1.05
I	•I	12.31 ±	0.92	1 1.. 50- 13.30 Height of Occipital	7.51
Male
I\'	8	21.83 ±	0.88	20.00-24.00	5.70	8.90
III	6	21.27 ±	0.52	20.45-22.20	2.95	0.0002
\'	5	21.24 ±	0.74	20.15-22.40	3.93
11	2	20.73 ±	2.86	19.30-22.15	9.72
I	7	18.65 ±	0.90	17.20-20.40	6.44
Female
III	2	20.13 ±	1.56	19.35-20.90	5.45	0.35
	2	19.95 ±	1.70	19.10-20.80	6.03	ns
II	3	19.72 ±	2.38	17.75-21.85	10.42
I	3	18.87 ±	1.44	17.45-19.75	6.57

measuremetus showed a significani F-value in the ANOVA. Age class I females differed significantly from the other age classes in only a single measurement (width of rostrum). In length of mandible, age class I differed significantly from age class but not age ( lasses II and III.

Secondary sexual variation. — Males averaged larger than females in all cranial measurements except diameter of the foramen magnum. The average percent of differente was 5.75, ranging from a low of 1.6 percent for postorbital constriction to a high of 10.34 percent for height of ottipital bone. Males and females differed significantly (p <.()5) in 17 ol 26 (tanial measurements ( Table 4). Measurements relieving length, width, and depth of skull were signili( antly different except for scjuamosal breadth and cranial height. Postorhital and interorbital constriction, as well as measurements of dentition (premolar-molar toothrow length, length of upper fourth premolar, diameter of the upper canine, and length of the diastema), did not differ significantly between sexes. Because of

18

()Cx;asional papers the museum

Table 4. — Secondary sexual variation in 23 cranial and three mandibular characters of adult P'elis rufus from Robertson and Madison counties, Texas. Statistics given are number, mean ± two standard errors, range, and l-value. See

text for character abbreviations.

Character (N)		Male.s		Ke male.s		(-value
GST (24;22)	128.43	± 5.34	(119.70-138.45)	122.88	±	5.13	(1 15.20-135.80)	3. .59**.
ZB (24; 14)	86.86	± 3.62	(78.15-92.05)	84.19	±	3.99	(79.05-92.40)	2.12*.
SB (17;13)	.53.19	± 1.44	(50.60-56.05)	52.42	±	2.25	(49.00-57.40)	1.14
POC (24; 19)	36.97	± 2.18	(31.75-40.20)	37.26	±	1.92	(33.95-40.75)	-0.45.
UN (32;25)	25.88	± 1.73	(22.10-29.80)	25.07	±	1.68	(21.20-27.40)	1.79
IC (31;21)	23.20	± 1.79	(18..55-27.55)	22.76	±	1.67	(20.20-26.20)	0.90
DUG (31;24)	7.25	± 0.59	(6.30-8.60)	6.62	±	0.48	(5.75-7.75)	4.23**
PM (32;24)	13.59	± 0.75	(12.20-15.25)	13.37	±	0.83	(H.30-14..55)	1.04
WR (30;23)	32.64	± 1.66	(29.80-35.85)	31.19	±	1.78	(28.05-34.15)	3.05**
PMT (32;24)	24.53	±1.15	(22.20-27.35)	24.39	±	1.10	(21. .55-25. 95)	0.46
WMT (30;22)	37.89	± 1.36	(35.00-40.,50)	36.08	±	2.. 56	(27.80-39.70)	3.02**
MT (31;24)	38.20	± 1.65	(35.20-41.70)	36.95	±	1.76	(33.80-40.30)	2.71**
PL (31;22)	47.73	± 2.28	(43.50-52.80)	46.55	±	1.99	(43.35-51.75)	1.95**
GH (29;20)	44.27	± 1.90	(41.35-47.95)	43.49	±	1..58	(40.00-45.90)	1.49
MB (24;17)	55.18	± 2.66	(51.75-61.05)	.53.63	±	2.51	(.50.10-58.35)	1.88
BL (21;16)	105.06	± 4.20	(97.2.5-113.65)	100.31	±	4.36	(93.00-111.35)	3.36**
GBL (23;17)	115.57	± 5.06	(102..50-124..55)	110.84	±	4.64	(103.60-123.45)	3.03**
LM (29;24)	74.72	± 3.35	(66.95-80.20)	71.38	±	4.19	(58.05-80.20)	3.22**
GRH (28;24)	36..50	± 2.41	(32. 20-41.. 55)	34.88	±	2.21	(30.80-40.05)	2.51*
DM (31;24)	16.49	± 1.11	(14.25-19.00)	15.43	±	0.91	(13.65-17.25)	3.80**
PPL (19; 1,5)	57.65	± 2.81	(53.70-65.10)	54.54	±	2.33	(51..50-.59.75)	3.44**
NFL (30;22)	57.12	± 2.85	(.50..50-62.40)	.55.40	±	2.47	(51.35-.59.75)	2.27*
GLB (28;21)	31.44	± 1.76	(27.40-34.90)	30.10	±	1.29	(27.85-32.35)	2.94**
VVAB (29;22)	15.22	± 0.91	(13.15-17.60)	14.70	±	0.77	(13.30-16.15)	2.17*
DEM (23; 19)	13..54	± 0.91	(1 1.90-15.70)	13.53	±	0.89	(12.55- 15. .55)	0.01
HOB (24;22)	21. .56	± 1.38	(19.25-24.40)	20.30	±	1.04	(18.80-22.90)	3.48**

*Signifi<ani at Of)-. 01. ••Sigiiifitanl at .01- 001.

the extent of sexual variation, the sexes were separated for analysis of geographic variation.

Individual variation. — The average CV for 26 measurements in the Robertson-Madison county sample (see Table 3) was 5.57, and CV’s of females (5.52) averaged slightly smaller than those for males (5.69). Average CV’s for males of the age classes were: I, 6.42; II, 6.12; III, 4.85; IV, 5.09; and V, 5.99; for females, these

same values were 5.97, 5.49, 3.81, - , and 6.42, respectively.

Thus, younger males (age classes I and II) were slightly more variable than were older males (age classes IV and V), whereas the opposite was true in females. Age class III exhibited the lowest CV in both sexes.

SCHMIDLY AND Rl- AD— CRANIAl, \'AR1 A I ION

19

The nieasurenieiUs with tlie lowest C^V’s lor males were scjuamosal breadth, mastoid breadth, postpalaial length, condylohasal length, and width across the auditory bullae; for females, the measurements with the lowest CV’s were scjuamosal breadth, width across the auditory bullae, mastoid breadth, width across the maxillary toothrow, and length c:)f mandible. I'he most variable (highest ClV’s) measurements for males were diameter of upf)er canine, coronoid height, length of nasals, diameter of foramen magnum, and postorbital cciiistriction; for females, the most variable measurements were postpalatal length, interorbital constriction, length of nasals, diameter of upper canine, and height of occipital bone.

Geographic Variation

Because of small sample sizes and the greater variation in size characters of females, emphasis on the analysis of geographic variation was given to males, although some information for females is presented for comparison.

Univariate Analysis

Patterns of univariate variation along the north-to-south and west-to-east transects for males, as depicted by Dice-Leraas diagrams, are illustrated in Figs. 2-6. dlie overall pattern of univariate variation along these transects is erratic for most characters, with alternating sections of increasing and decreasing size. Smooth dines are rare in most measurements, and clinal changes from north to south and from west to east are evident oidy in localized {)arts of the range.

Transect A. — Fhe four measurements do not follow a concordant pattern along this transect, although the deviations among them are relatively minor (Fig. 2). Proceeding from north to south, there is a slight size decrease in scptamosal breadth anci length of upper fourth premolar, whereas a slight increase in size is evident in cranial height anci skull length. Individuals from sample 17 (central Fexas) are significantly larger than those Irom adjacent samples in skull length and length of the upiter fourth premolar, hut this trend is not apparent in the othei two measurements. Fhus, variation along this transect is erratic with few concc^ndant breaks or charac ter shifts.

Transect B. — Fhe smallest individuals are found in the c'astern samples of this transect, with size gradually incieasing in samples

20

OCCASIONAL PAPERS I HK MUSEIIM

H - ^ ^ ^ - 1 - 1 I I ( I t I I I I I - ^ - 1 - 1 - 1 I M M I I -t" 4 4' tH

118 125 135 145 50 52 54 56 58 12 14 16 39 41 43 45 47 49

Fig. 2. — Geographic variation, expressed by Dice-Leraas diagrams of selected characters, among samples of Felis rufus along transect A in the study area. Sample designation (Sam.) appears along the left margin. See Fig. 1 and Table 1 for location of samples. The horizontal line represents the range; vertical line, the mean; open rectangle, one standard deviation; and closed rectangle, tvs'o standard errors of the mean.

progressing toward the southwest (Fig. 3). The magnitude of variation is much less in length of the upper fourth premolar than in the other measurements. In most geographic regions, character transitions follow a pattern of smooth, gradual change with few significant breaks. Exceptions include sample 25 (northern Trans-Pecos Texas), which has a significantly larger skull length than the adjacent sample (21) from the High Plains; sample 7 (north-central Texas), which has a significantly smaller squamosal breadth than the two adjacent samples in central Texas (18, 2); and sample 21 (High Plains of Texas), which has a significantly greater cranial height than the adjacent sample (18).^

Fig. 3. — Geographic variation of Felis rufus, expressed by Dice-Leraas diagrams of selected characters, along transect B in the study area. See Pig. 1 and Table 1 for location of samples and P ig. 2 for an explanation of the diagrams.

21

SC;nMIl)lA AND RKAD— c;ranial \'ARIA1 ion

	GLS	SB	LPM	CH
SAM
64		-
			t m ) 1 — ^ — 1
99 25	-		1 — IPLJ
23			-a*B—
15			-am-	-am-
13	-	—
11			3m~	*

H — i 1 1 — I — I 1 h-t— I — I — I — I — I — I — I I - 1 - 1 - 1 - 1 I f I M M-n M i I I

113 120 130 140 50 52 54 56 58 12 14 16 37 39 41 43 4 5 47 49

Fig. 4. — (ieot^raphic variation of Felis rufus, expressed by Dice-Leraas diagrams ot selected characters, along transect C in the study area. .See Fig. 1 and Fable 1 for location of samples and Fig. 2 for an explanation of the diagrams.

Transect C. — The pattern of variation along this transect is more erratic than that observed for the other transects (Fig. 4). .Samples 64 and 60 (Arizona and southwestern New Mexico) have significantly smaller skull lengths than other samples in the transect. Sample 25 (northern Trans-Pecos Texas) averages larger in skull length than the two adjacent samples (59 from southeastern New Mexico and 24 from the Big Bend region of Texas), whereas samp^les 13 and 15 (southern Texas) average smaller in this measurement than adjacent samples (11 from the Fexas Ck)a.st and 23 from the Stockton Plateau). Length of the upper fourth premolar shows an almost identical f)attern of variation except that individuals in sample 64 (Arizona) are significantly larger than those of sample 60 (southwestern New Mexico). Fhe same general pattern is also evident in cranial height; samples 64 and 60 are significantly smaller than other samples in the transect, and sam})le 1 1 averages larger than sample 13.

Transect 1). — Fhree of the four measurements (skull length, length of u[)per fourth premolar, and cranial height) exhibit a similar pattern of variation along this transeit (Fig. 5). Fhesi' measurements in the western samples (sample 63 from southeastern Arizona and sample 60 Irom southwestern New Mexico) are small and a significant size incrc'ase occurs lietwcTU samples 60 (southwestern New Mexico) and 59 (southc'astern New Mexico). Beginning with sample 59 and continuing east through I'exas and Arkansas, size gradually declines. Fhe reference sample of floridanus (66) is significantly smaller in skull length and length of upper fourth premolar than samples from Arkansas.

22

OCCASIONAL PAPERS 1 HE MUSEUM

Pig. 5. — Geographic variation of Felis rufus, expressed by Dice-Leraas diagrams of selected characters, along transect D in the study area. See Fig. 1 and Table 1 for location of samples and Fig. 2 for an explanation of the diagrams.

Squamosal breadth gradually increases over the first three samples (63, 60, 59) and then, with the exception of sample 2, decreases in size through the remainder of the transect.

Transect E. — There was no consistent pattern of variation in the four measurements along this west-to-east transect across the northern portion of the study area (Fig. 6). Little significant variation was evident in skull length and cranial height, and the pattern for length of the upper fourth premolar was highly erratic. The only measurement to show a definite pattern was squamosal breadth, which exhibited a classic pattern of clinal change beginning with smaller size in eastern samples and gradually increasing in size toward the west.

Fig. 6. — Geographic variation of Felis rufus, expre.ssed by Dice-Leraas diagrams of selected characters, along transect E in the study area. See Pig. 1 and Table 1 for location of samples and Fig. 2 for an explanation of the diagrams.

SCHMIDI.V AND RKAD— CRAN I AI . VARIAI ION

Males. — The munber oi maximally iionsigiiitu am suhseis t^eneraled tor the 26 characiers in the DTNC^AN analysis ol males varied trom two (diameter ot the foramen magnum) to 10 (mastoid breadth) with all hut four of the measurements (diameter of the foramen magnum, length of nasals, length of upper fourth premolar, and greatest length of auditory bullae) reriuiring at least six nonsignificant subsets to cover the range of vaiiation. Considerable overlap was evident among the arrays of std)sets for each character, and there were ncj instances where one array was completely segregated from the others. In overall size, the smallest male bobcats were from southern Louisiana (47-49) and west of the Continental Divide in New Mexico (60, 61) and Arizona (63, 64). The largest individuals were from southeastern New' Mexiccj (59) and western (17, 22-25) and southern (11, 12) Texas.

Females. — A similar but slightly different pattern was evident in the DUNCAN analysis of females. The arrays of maximally nonsignificant subsets ranged from two (diameter of the foramen magnum) to 12 (mastoid breadth) with all but three characters (length of nasals, width of rostrum, and diameter of the foramen magnum) having six or more arrays of subsets, all of which overlapped substantially. I'he smallest females were from Arizona (63), southeastern Oklahcjma (55), Arkansas (51-54) and eastern Louisiana (49). The largest bobcats in most measurements w'ere from central Oklahoma (57, 58), western Texas (38, 40, 42, 58), and southeastern New Mexico (59).

Midtivariate Analysis

CAuster analysis. — A distance phenogram was generated using all samples of males and females sejxtrately, and the residts w'ere substantially different. Lhe phenogram for males (Fig. 7) separates into two groups (A and B), w'ilh the exception of sample 47 (comprised of only tw'o individuals from Natchitoches Parish, Louisiana), which segrc'gates by itself, (.roup A includes reference samples of F. r. floridanus (65, 66) and F. r. baileyi (63), two samples (60, 61) from west ol the Continental Divide in New Mexico, and a sample (49) from east of the Mississippi River in Louisiana. Lhese bobcats are the smallest in overall size, (iroup B contains the remaining samples from the study area plus single samples of the jloridanus (67) and haileyi (64) reference samples. Ibis group can be further divided into subgroups I and II. Subgroup I, which includes bobcats of large overall size, is made

21

ocx;asional papers i hk museum

MALES

2.55 2.25 1.95

H - ^ - 1 - 1 - 1

1.65 1.35 1.05 .750 .450

Eig. 7. — Distance phenograrn of the cluster analysis for samples of male Felis rufus. The cophenetic correlation coefficient for the phenograrn is 0.772.

SCHMIDLV AND RKAD— CRANIAl . N'ARIAHON

25

up ot samples from the coastal bend of J'exas (1, 5, 11, 12), western I’exas (22-25), southeastern New Mexico (59), and central Oklahoma (56, 57). Subgroup) II, which includes bobcats of intermediate size, consists of sam[)les from central d'exas (18-21, 1-3, 6, 7, 9, 10, 13-16), southeastern Oklahoma (55), and Arkansas (51-54).

'I'he phenogram for females (Fig. 8) also separates into two major groups. Group A, which comprises those females of relatively small size, includes reference samples of F. r. baileyi from Arizona (63, 64) plus samples from western Arkansas, southeastern Oklahoma, and the lower Mississippi River Valley in l.ouisiana. Group B, which includes the remaining samples, is further divided into two subgroups. Subgroup I includes female bobcats of intermediate size from western New Mexico, a single sample (51) from eastern Arkansas, and a series of samples stretching from central Oklahoma southward into central and southern Texas. Subgroup II, which includes females of relatively large size, is comprised primarily of samples from western and southern Texas.

Principal components analysis. — The first three principal components were computed from the matrix of correlation among the 26 characters. For males, the first principal component expresses 60.72 percent of the phenetic variation; the second, 11.11; and the third, 5.05; for females, these values are 69.59, 10.82, and 3.80, respectively.

Loadings (Table 5), which indicate the correlations of characters with the first three principal components, indicate that component I is es.sentially a general size factor with high positive correlations for all characters except postorbital constriction and diameter of the foramen magnum. The six characters with the largest loadings in both sexes are skull length measurements. For males, these are basilar length, condylobasal length, length of mandible, greatest length of skidl, palatilar length, and premolar-molar textthrow length; for females, they are length of mandible, greatest length of skidl, basilar length, ( ondylobasal length, maxillary loothrow length, and palatilar length. With respect to positioning of samples along (omponent I, samples containing specimens that were smallest in the.se measurements are located on the far left; from that point, .samples are arranged in ascending order relative to size, with those containing the largest individuals on the far right of the plot (Fig. 9). For males, the smallest bobcats are from Arizona, western New Mexico, and

26

OCCASIONAL PAPERS I HE MUSEUM

FEMALES

n

B

c

c

c

£

{

26

28

37

29

45

32

39

59 62

43

38 42

58

27

30

31

33

34

35

44 41

46

36

51

56

47

40

60 61

57

49

52

50

53

- 54

- 55

- 64

- 63

h

+

0.6

0.4

- H

1.8 1.6 1.4 1.2 1.0 0.8

Fig. 8. — Distance phenograin of the cluster analysis for samples of female Felis rufus. 'Lire cophenetic correlation coefficient for the phenogram is 0.709.

SCllMIDLY AND RKAD— CRANI Ai. VARI VI'ION

27

I ABi.F. f). — Character loadings on the first three principal components of interlocality phenetic iniriation in males (M) and females (F) of f elis ruins. See

text for character abbrexnations.

			Pi iiu ipal	( oinponcnts
		1		II		III
Ch.iia< tfi	M	K	M	V	M	F
(;si.	.939	.961	-.01 1	-.013	.019	.110
ZB	.785	.867	.314	-.358	.162	-.016
SB	.166	..568	.722	-.791	-.202	.064
POC	-.071	-.274	.764	-.868	.173	-.023
I.N	.819	.642	-.289	.201	-.483	-.075
IC	.776	.61 1	.570	-.445	.326	-.103
DUC	.747	.737	-.454	.466	-.074	-.274
P.M	.7.53	.712	.009	.324	-.372	-.019
VVR	.851	.840	-.320	.335	.028	-.214
PMT	.814	.864	.076	.240	-.147	.020
VVM I	.904	.842	-.191	.034	.018	-.258
MI	.948	.834	-.145	.111	-.092	-.035
PL	.951	.893	-.334	-.032	-.032	-.016
CH	.832	.812	-.023	.003	.080	-.223
MB	.880	.838	.423	-.359	-.053	.123
Bl.	.958	.952	-.007	-.044	-.028	.074
CBL	.948	.950	-.014	-.046	-.067	.090
LM	.978	.942	-.162	.012	.122	-.028
CRH	.915	.844	-.023	.043	.176	-.112
DM	.841	.717	-.258	.005	.025	-.067
PPL	.926	.843	.159	.013	.032	.120
NFL	.932	.868	.042	-.129	.221	.116
GLB	.901	.838	.087	.006	-.229	.055
VVAB	.668	.542	.589	-.461	-.120	.148
DFM	-.326	-.349	.372	-.438	-.678	-.763
HOB	.919	.779	-.048	.073	.254	.184

llie southeastern United States (Georgia, Florida, Louisiana, and Arkansas); bobcats of medium to large size occupy a geographic area including eastern New Mexico, Fexas, and rnosi of Oklahoma. The same general ger)graphic trend is evident for females with the smallest bobcats being from the western (Arizona and New Mexico) and eastern (Arkansas and Louisiana) parts of the study area and bobcats of medium and large size occurring in the intervening areas.

Component II for males has high positive loadings for characters affecting shape of braincase and width of skull, including postorbital constriction, scjuamosal breadth, width of the auditory bullae, interorbital constriction, and mastoid breadth, and a high negative loading for diameter of the upper

28

OC:CA.SIONAL PAPERS 7'HK MUSEUM

n

.8-

.6-

.4-

.2-

0-

,2-

-.4-

,6-

.8-

-1.0-

47*

•61

•63

•60

•66

• 49

•65

•64

•58

• 62

1 ' •ao

•19 *21 18^^2 •9^3

• 20 •21 •23

•55 *1 ^15 *^®10

• 8 *12'

54* •I

53

•14 •O

• 50l

• 67

•52

• 51

*48 56 •»4

• 5

• 57

• 24

• 25

•II

•17

MALE

1 - \ - 1 - 1 - 1 - 1 - ^ - r

“1 - 1 - 1 - 1 - 1 - 1 - 1 - \ - 1 - ^ - 1 - 1 -

2.4 2.2 -2.0 - 1.8 -1.6 1.4 -1.2 -1.0 -.8 -.6 -.4 -.2 0 .2 .4 .6 .8 1.0 1.2 1.4 1.6

n

.8-

.6-

.4

.2-

0-

-.2-

-.4-

-.6-

-.8-

1.0-

• 47

• 55

»54 ^52

•49

.•53

*50

•36

•56

•51

33.

•27

••30

*28 *2^45'

•26

• 64

|•34 *31

• 44.. -•44 937 •A2

•35^46 - .£29.^ -

•OO

I ^43

, ,58 *38 *57

•59

•62

• 60

• 63

.el

FEMALE

“I - 1 - 1 - 1 - \ - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - r-

2.4 -2.2 -2.0 1.8 -1.6 1.4 1.2 1.0 -.8 -.6 -.4 -.2 0 .2 .4 .6 .8 1.0 1.2 14 1.6

Fig. 9. — Two-dimensional projections of the samples of male and female Felis rufus onto the first two principal components.

canines (Fig. 10). The loadings for females on this component are a mirror image of those in males with high negative loadings for the same characters affecting braincase shape and skull width, and a high positive loading for diameter of the upper canines. With respect to positioning of samples along component II, the pattern is similar for both males and females. Most samples of males with positive values and those of females with negative values are from the shrub-grassland habitats in the western half of the study area (west of longitude 99° W). These bobcats are characterized by skulls with a wide squamosal breadth, postorbital constriction, interorbital constriction, mastoid breadth, and auditory bullae, and a narrow diameter of the upper canines. Most samples of males with negative loadings and those

SCHMIDI.V AND RKAD— CRANIAL \ ARIA1 ION

29

.5- .4 - .3 - .2 - .1 - 0- - .1 - -.2 - 3- 4 - -.5 - -.6 -

• 47

• 67

• 10

• 58

• 14

• 25

• 2

•16 •a

• 57

,15 I •12

0*18

.19

• 6

• 62

-•3-

• 13

• 59

-•6a

• 64

m

r

-.8

-•17-

• 5

• 51

• 48

•11

• 20

• 49 • 52

• 1

56^«^4 •eo 50 . •$

53

• 21 • 23

• 22

• 24

• 65

•55

~r

.5

•61

MALE - 1 -

.7

r

-.6

T

-.5

-.3

I

.2

0

n

.3

2

.1

0

n 1 2 .3 .4

•64 • 62

• 38

• 57

•35*^^ •42 • 46

•^5

• 36 • 26

•56 • 30

•63

• 61 • 60

•40 •41 ,28

— — — — — I — *44 — — — — —

•59 •29 *52 ^51

• 58

• 32 •39

•43

• •53

50

• 49

• 34

•55 • 33

• 27

~i - r

.7 -.6

“T

-.5

— I - 1 - r

-.4 -.3 -.2

— *47

FEMALE

~r

.5

n

Kif.. 10. — I'wo-climen.sioiial projections of the samples of male and female Fe/is rujus onto the second and third princ ipal components.

of females with positive loadings are from the broadleaf and needleleaf eastern forests and the grassland-forest habitats in the eastern half of the study area (east of longitude 99° W). Ihe skulls of the.se hoheats have a narrow stjuamosal hreadth, postorhital constriction, mastoid hreadth, and auditory bullae, and a large diameter of the upper c anines.

Ciomponent III in males has high negative' loadings for nasal length, length of ujtper fourth jtremolar, and diameter ol the foramen magnum. Interorhital constriction and hc'ight ol ilu' occipital hone are the only characters with a high positive loading. Cdiaracters with high negative loadings on this component in females include diameter of the foramen magnum, diameter of the upper canine, interdentary hreadth, cranial height, and rostral width. Fhe only character with a high

30

OCXJyVSIONAL PAPERS I HK MUSEUM

positive loading is height of the occipital bone. Most samples of males with negative values for component III (narrow interorbital constriction, large diameter of foramen magnum, and short height of the occipital bone) are distributed in the eastern half of the study area, whereas those with positive values (character trends opposite those above) are from the western half of the region. Samples of females with negative values for component III have a slightly larger diameter of upper canines, wider rostrum, greater cranial height, and greater diameter of the foramen magnum than samples with positive values for this component. However, there is no obvious geographic trend with respect to the positioning of samples of females along component III.

Discussion

Nongeographic Variation

The greatest nongeographic variation is with age and involves proportions as well as general size. Analysis of age variation in male bobcats suggests that adult size is reached at about 24 months of age at which time the temporal ridges and muscle attachments on the frontals become well defined. Male bobcats do not increase appreciably in size from 24 to 48 months of age, but the temporal ridges and muscle attachments become more {pronounced. Females seem to attain adult size much earlier than males. In most cranial measurements of females, there was no significant difference between age classes I and II; however, age class I females lacked tem{Poral ridges and frontal muscle attachments. These results are consistent with those of Grinnell and Dixon (1924) who noted that bobcat teeth are not subject to much wear or breakage and that com{)arative age is best determined by degree of development of the attachments for muscles and by the stage reached in the effacement of sutures than by degree of wear shown by the teeth.

With res{pect to sex, the skulls of males are significantly larger, longer, and more shar{)ly ridged than those of females of the same age. These residts agree with those of Grinnell and Dixon (1924) who found that male bobcats in Galifornia were roughly one-fourth larger than females. Samson (1979) re{Ported that nine measurements useful in distinguishing the 11 subspecies of F. rufus showed no clear sexual dimor{phism and he combined sexes for {Pur{POses of making {phenetic comfparisons amcpiig the subspecies. I'o the contrary, our data, together with the

SCHMIDIA AND RKAD— CRANIAI . \ARIA1 ION

.SI

(oiichision.s ol Cirinncll and Dixon (1924), .strongly suggest the .sexes shonld be treated separately in studies of geogiaphit variation. Cloinbining sexes tor purposes of geographic comparisons con Id jjrodnee erroneous resnlts.

With respc'ct to individual variation in cranial characters, bobcats from the sonth-central Ihiitc^cl .States sc*em to be slightly more variable than those from other geographic regions. Long (1968) re{)c)rtecl that F. rufus from Wyoming had CV's ranging from 3.45 (greatest sknll length) to 5.96 (interorbital breadth). The Robertson-Madison (bounty sample showed a much wider range of values, with 16 of the 26 measurements having CA'^’s larger than the upper limit reported by Long (1968).

Geographic Variation

I he trend of geographic variation in cranial characteristics c:)f bobcats from the study area, considering both univariate ancf mrdtivariate analyses, may be summarized as follows; 1) bobcats from the eastern deciduous forests c:)f Louisiana and yVrkansas have small rounded skulls; 2) progressing westward into (Oklahoma and Texas skulls increase in size so that bobcats from this region are medium to large in cranial measurements; and 3) continuing westward across the Continental Divide skidls decrease in size and the smallest bobcats occur in this portion of the study area.

I he region of greatest phenetic divergence is across the Ciontinental Divide in New Mexico. .Specimens from west of the Divide are significantly smaller in most measurements than those from c^ast of the Divide. I'here is a lack of distinct .separation among populations in most other geogra[)hic areas and intergraclation appears to be evenly progressive instc'ad of step¬ like. A broad zone of intergraclation exists betwc’en bobcats of intermediate and large size, which occur in Oklahoma and central Texas, and the smaller bobcats found in Arkansas and Louisiana. Ihdike the spc’cimens from New Mc'xico, which display a step-like brc'ak as.sociatc'd with a physiographic bairica ((iontinental Divide), ifie individuals from the c'astein j)oriion of the study area seem to respond to major vc'getativc' lypc-s, with large- to medium-sizc'd bobcats having cwolvc'd in the more open habitats and smaller bobcats in the forc'sted situations. The zone of iniergradation betwc^en the.se two size groups corresponds to tlie broad ecotone created by the transition from the broadleaf and pine forests of the eastern Lttited .States and the prairie habitats of central I'exas and Oklahoma.

32

()CCy\SI()NAL PAPERS EHE MUSEUM

The variational pattern described is understandable in light of the continuous and broad distribution of bobcats. The degree and magnitude of geographic variation probably results from selective pressures exerted by the ecological conditions characteristic of the most distinctive habitats that bobcats occupy. However, the complex array of environmental pressures that have produced the observed variational patterns cannot be individually dissected given the information base presently available.

Subspecies

We find little justification for recognizing more than one subspecies of the bobcat in Texas. Although regional trends are evident in the size and shape of bobcat skulls over the state, there are no geographic regions where populations are significantly distinctive and strongly demarked from populations in nearby areas by sharp phenetic breaks indicative of reduced gene flow. Therefore, we refer all populations from Texas examined in this study to the taxon F. r. texensis (Merriam).

The western boundary of F. r. texensis corresponds with the location of the Continental Divide in New Mexico. The most distinctive bobcats we examined are those from west of the Continental Divide in New Mexico and Arizona, and to this group the trinomial F. r. baileyi (Merriam) applies. None of the Texas bobcats we studied can be appropriately assigned to the latter, although Hall (1981:1053) previously mapped the Panhandle and Trans-Pecos portions of the state as occurring within its range. Specimens from eastern New Mexico (samples 59 and 62) are indistinguishable to us from those of F. r. texensis in western Texas.

Our reference samples of F. r. floridanus from Florida and Ceorgia are well differentiated by their smaller size from Texas samples of F. r. texensis. Lowery (1974:470) and Hall (1981:1053) referred specimens from eastern Louisiana and Arkansas to F. r. floridanus. Our analysis, however, suggests that specimens from these states represent intergrades between F. r. texensis and F. r. floridanus. Because their overall cranial characteristics resemble the former subspecies more than the latter, we have tentatively referred them to the subspecies F. r. texensis. Bobcats from east of the Mississippi River in Louisiana (sample 49) are especially like F. r. floridanus, suggesting this river may form an ap[)ro{)riate boundary for this subspecies.

S(;11M11)L^ AND RKAD— CRANIAL CARIAI ION

riu* laxonoinic assignment ol material from Oklaiioma and Arkansas to F. r. texensis must he considered tentative because we did not examine sjiecimens ol the subspecies t. r. rufus, to which Hall (1981:1053) reterred specimens from eastern Oklahoma and western Arkansas. Our analysis reveals that samples trom cential Oklahoma (56-58) are virtually indistinguishable from samples ol F. r. texensis from I'exas and that specimens from southeastern Oklahoma (55) and western Arkansas (52-54) are phenetically c lose to Louisiana samples of the same subspcxies.

F.arly authors (Baird, 1858; Merriam, 1890) stressed the importance of coloration in recognizing subspecies and varieties of the bobcat. Although we did not analyze pelage coloration cjuantitatively , general color descriptions and comments were noted, particularly among specimens housed in the United States Museum of Natural History. Considerable color variation was evident in pelts from the same locality, and this seemed to ccjrrelate with age, season of year, and degree of pelage wear. Pelage color often differed substantially among specimens from nearby localities, whereas other individuals from widely removed Icicalities were identical in color. For these reasons, pelage coloration was considered unsuitable as a taxonomic character for bobcats. Grinnell and Dixon (1924) and Peterson and Downing (1952) reached similar conclusions for the geographic areas in which they studied bobcats.

Samson (1979) used cranial measurements to make a morjihometric comparison of the 1 1 recognized subspecies of F. rufus as listed by Hall and Kelson (1959). However, Samson used a statistical approach (stepwise discriminant analysis) in analyzing his data that retpnres an a prion assignment ol spec imens to grou[)S (in Samson’s case, subspecies). I his technitpie maximizes the detection of morphological separation among groups, and it is not surprising that the subspecies ol h. rufus could be discriminated from one another using this approach. More conservative multivariate techniiiuc's (cluster analysis and principal components analysis), recjuiiing no a priori assumptions rt'garding the data, wc'ie usc'd in this study to provide a representation of the distances among samples. Use of these more appropriate and conservative techniciuc's did not produce results consistent with the existing taxonomic arrangement for bobcats in the study area. I herefore. we doubt the validity of Samson’s conclusions concerning bobcat stibs|)ecies

M

OCiCASIONAL PAPKRS I HE MUSEUM

and suspect that there are far fewer valid intraspecific taxa than are currently recognized.

Management and Legal Im.plications

In 1979, Defenders of Wildlife filed suit in U.S. District Court to prohibit the export of pelts from certain states (including all those in the study area) on the grounds that available data were not adequate to define population trends in certain subspecies and geographic areas (Defenders of Wildlife, Inc., vs Endangered Species Scientific Authority et al., Civil Action no. 79-3060, U.S. District Court for the District of Columbia, 12 December 1979). The court dismissed the plaintiff’s claims as to Arkansas, Louisiana, and Oklahoma, but enjoined export of bobcat pelts from New Mexico and the High Plains of Texas on the grounds that this geographic area roughly corresponds to the range of Felis rufus baileyi and that export from that area would threaten the survival of baileyi as a subspecies (Memorandum Opinion, p. 7 in the transcript on the hearing in District Court). Our analysis reveals that bobcats in Texas belong to a single subspecies, F. r. texensis, and that the threatened subspecies, F. r. baileyi, is restricted to the region west of the Continental Divide in New Mexico and Arizona. Therefore, there seems to be little justification for prohibiting the export of bobcat pelts collected in Texas unless it can be demonstrated that the survival of F. r. texensis would be threatened by such activity.

Synopsis

Felis rufus texensis (J. A. Allen)

1829. Felis rnaculata Horsfield and Vigors, Zool. Jour., 4:381, pi. 13, type Irom Mexico. Not Felis (Lynx) vulgaris maculatus Kerr, 1792.

1895. Lynx texensis J. A. Allen, Bull Amer. Mus. Nat. Hist., 7:188, based on the description of a bobcat by Audubon and Bachman, The viviparous cjuadrupeds of North America, 2:293, 1851.

1897. Lynx rufus texensis Mearns, Preliminary diagnoses of new mammals of the genera Lynx, Urocyon, Spilogale, and Mephitis, from the Mexican boundary line, p. 2, 12 January 1897 (preprint of Proc. U.S. Nat. Mus., 20:458, 24 December 1897).

Flolotype. — None designated; two syntypes, one of which is figured (Plate XCII), were described by Audubon and Bachman (see synonomy above); they are from the vicinity of Castroville, headwaters of Medina River, Medina Co., Texas.

SCillMlDlA AND Rl AD— ( ;R AN 1 A1 . VARIATION

Distribution. — New Mexico east ol the (a)niiiiental Divide; throughout I'exas, Oklahoma, Arkansas, and Louisiana west ol the Mississippi River.

Diagnosis. — A medium-sized, reddish brown or grayisli subspecies of F. rufus characterized by a large deep skull with a relatively narrow braincase, medium-sized auditory bidlae, and relatively large canine teeth.

C.oniparisons. — For a comparison with F. r. baileyi, see account of that subspecies. From F. r. jloridanus, the subspecies texensis differs in being larger in overall size, with a mc3re rounded and higher skull (as reflected in the measurements depth of cranium and height of cjccipital bone).

Measurements. — The following are mean values (in mm) of 23 cranial and three mandibular measurements for seven males (sample 8) and 11 females (sample 35) of this subspecies from central Texas (mean values lor females are in parentheses and follow thc3se of males): greatest length of skull, 129.59 (121.40); zygomatic breadth, 92.31 (83.76); scjuamosal breadth, 54.63

(52.52); postorbital constriction, 35.48 (36.84); length c^f nasals, 26.36 (24.23); interorbital constriction, 23.43 (21.43); diameter of upper canine, 7.69 (6.46); length of upper fourth premolar, 14.10 (12.95); width of rostrum, 33.35 (30.55); premolar-molar toothrow length, 25.27 (23.63); width acrc3ss maxillary toothrow, 38.78 (35.78); maxillary toothrow length, 37.45 (36.50); palatilar length, 48.90 (46.05); cranial height, 44.54 (42.24); mastoid breadth, 56.28 (52.21); basilar length, 108.53 (98.98); condylobasal length, 118.69 (111.42); post-palatal length, 60.38 (54.35); nasal-frontal length, 57.78 (53.66); greatest length of auditory bullae, 31.65 (28.92); width of auditory bullae, 15.22 (14.90); diameter of foramen magnum, 13.30 (13.22); height of occipital bone, 21.75 (19.49); mandibidar length, 76.36 (70.34); coronoid height, 38.75 (34.19); mandibidar diastema, 16.64 (15.29).

Samples.— Comprised of specimens in our study from the following sam[)les ( Fable 2 and Fig. 1): New Mexico: 59, 62; Oklahoma: 55-58; Arkansas: 51-54; Louisiana: 47, 48, 50; Texas (males): 1-25; (females): 26-46. Iticluded, in addition, aie the following specimens deposited at Fhe Museum, Fexas Fech llniversity: Texas: Armstrong To.: 29 mi SF. Cdaude, 1; Brewster Co.: Arnette Ranch, 20 mi S Marathon, 8; Crosby Co.: near Crosbyton, 2; Dickens Co.: 6 mi N, 16 mi F Dickens, 1; Foard Co.: no specific locality; Garza Co.; 10 mi N Post, 1; 4 mi N Post,

36

()Cc;asionai. papers i hk museum

1; Hardeman Co.: near Lazare, 1; Haskell Co.: 25 mi SE Haskell, 1; Jeff Davis Co.: 13 mi NW Marfa, 1; 9.5 km E, 9 km N Ft. Davis, 1; Lubbock Co.: no specific locality, 1; 4.8 mi NW Lubbock, Hwy 82, 1; McCulloch Co.: 12.6 mi S Winchell, 1; Motley Co.: 1 mi E Matador, 1; Pecos Co.: 16.4 mi N Sheffield, 2; Stephens Co.: 13 mi NW Breckenridge, 1.

Felis rufus baileyi (Merriam)

1890. Lynx baileyi Merriam, N. Amer. Fauna, 3:79.

1901. [Lynx rufus] baileyi, Elliot, Field Columbia Mus. Publ. 45, Zool. Set., 2:297.

1905. Fells rufa baileyi, Elliot, Field Columbia Mus. Publ. 105, Zool. Set., 6:372. 1978. Felis rufus baileyi, Anderson, Bull. Amer. Mus. Nat. Hist., 148:388.

Holotype. — U.S. National Museum, Biological Survey Collection, no. 5214/5909; from Moccasin Spring, north of Colorado River, Coconino Co., Arizona. Type examined.

Distribution. — Arizona and New Mexico west of the Continental Divide in the study area; also known from southern Utah and Nevada as well as the southeastern arid region of California (Hall, 1981:1053).

Diagnosis. — A pale-colored (yellowish gray in winter and pale yellowish in summer) subspecies of F. rufus characterized by a short, narrow, and shallow skull with a large braincase, well rounded auditory bullae, and relatively small canines.

Comparisons. — Compared to F. r. texensis, the subspecies baileyi averages smaller in skull length and width (as reflected by greatest length of skull, zygomatic breadth, and mastoid breadth), but is larger in width of the braincase (as reflected by squamosal breadth and postorbital constriction) and width of the auditory bullae.

Felis rufus baileyi is similar in overall size to F. r. floridanus, but its skull is slightly shorter, flatter, and more angular with wider measurements in the braincase region (as reflected by squamosal breadth and mastoid breadth), narrower rostral measurements (as reflected by width of rostrum, width across maxillary toothrow, and zygomatic breadth), and a shorter cranial height and height of the occipital bone.

Measurements. — The following are mean values (in mm) of 23 cranial and three mandibular measurements for 43 males and 18 females of this subspecies from sample 60 in southwestern New Mexico (mean values for females are in parentheses and follow those of males): greatest length of skull, 123.55 (117.80);

SCHMIDI.V AND Rl- AD— CRANl A1 . \ ARIAI ION

zygonialic breadth, 86.92 (89.26); s(}nainosal breadth, .59. .50 (59.12); postorbital constrit tion, 38.70 (38.86); len.Rth oi nasals, 25.29 (23.99); interorbital eonstrir tion, 23.56 (22.69); tliaineter ol upper canine, 6.98 (6.58); length of up{)er fourth preniolar, 13.53 (13.23); width o( rostrum, 30.88 (29.61); premolar-molai toothrow length, 29.31 (23.70); width across maxillary toothrow, 36.68 (35.26); maxillary toothrow length, 37.73 (36.20); palatilar length, 96.78 (99.88); cranial height, 92.38 (91.03); mastoid breadth, 59.08 (52.79); basilar length, 101.51 (96.68); condylobasal length, 112.32 (107.29); post-palatal length, 55.92 (52.13); nasal-frontal length, 56.92 (53.55); greatest length of auditory bullae, 30.13 (29.16); width of auditory bidlae, 15.39 (15.25); diameter of foramen magnum, 13.92 (19.35); height of occipital bone, 19.63 (17.99); mandibular length, 73.02 (69.77); coronoid height, 35.99 (33.28); mandibular diastema, 16.31 (15.81).

Samples. — Comprised of specimens in our study from the following samples (Table 1): New Mexico: 60, 61; Arizona: 63, 69.

Felis rufus floridanus (Rafinesc|ue)

1817. Lynx floridanus Rafiiies(]ue, Amer. Monthly Mag., 2(l);-46.

18.58. Lynx rufus var. floridanus, Baird, Mammals in Repis. Kxpl. ,Sui\. . . , 8( l):91.

Holotype. — Philacielphia Acad. Sci., no. 12763; frtjm Biscayne Bay, 6 mi. .S Miami, Dade Co., Florida. Fype not examined.

Distribution. — Confined to the southeastern Ihiited States, east of the region of concern for this study.

Diagnosis. — A small, dark subspc'cies of F. rufus characterized by a short, narrow, relatively deep skull with a small braincase, narrow auditory bidlae, and relatively large canine teeth.

Comparisons. — For comparison with F. r. haileyi and F. r. texensis, see accc)unls of tho.se subspcx ies.

Measurements. — I he following are mc'an values (i?t mm) lor 23 cranial and three mandibular measurements ol 7 male F. r. floridanus (sample 66) from (»eorgia: grc'aiesi length of skidl, 123.76; zygomatic breadth, 85.96; scjuamosal brc'adlh, 52.56; postorbiial constriction, 37.05; length of nasals, 26.56; interoibital constriction, 21.62; diameter upper canine, 7.01; length of uppc'i fourth premolar, 13.21; width of rosttum, 31.29; premolar-molai toothrow length, 23.79; width across maxillary toothrow, 36.38; maxillary toothrow length, 37.13; palatilar length, 96.89; cranial height, 93.26; mastoid breadth, 52.98; basilar length, 101.26; condylobasal length, 111.67; post-palatal length, 59.79; nasal-

38

OCCJyVSIONAL PAPERS THE MUSEUM

frontal length, 54.94; greatest length of auditory bullae, 30.52; width of auditory bullae, 14.86; diameter of foramen magnum, 13.91; height of occipital bone, 19.66; mandibular length, 72.41; coronoid height, 35.64; mandibular diastema, 16.57.

Samples. — Comprised of specimens in our study from the following samples (Table 1): Florida: 65; Georgia: 66, 67; Louisiana: 49.

Acknowledgments

Many people helped in the course of this research. Special thanks are extended to Michael Tewes, Terry Blankenship, Brian Barnett, and Donna Morgan for assistance in specimen preparation. Thanks are due the many museum curators who permitted us to examine specimens in their care. We are also appreciative of the assistance from the Texas Parks and Wildlife Department (especially Mr. W. C. Brownlee of that agency) in obtaining specimens from trappers in Texas. Financial support was provided by the Texas Agricultural Experiment Station and the Caesar Kleberg Program in Wildlife Ecology at Texas A&M University. This paper represents contribution no. TA-20285 of the Texas Agricultural Experiment Station, Texas A8cM University.

Literature Cited

Baird, S. E. 18.58. Mammals. In Reports of explorations and surveys. ..from the Mississippi River to the Pacific Ocean..., 8( 1 ):xxi-xlviii + 1-757.

Bi.ankenship, T. 1979. The reproduction of bobcats in relation to prey populations. Ihipublished M..S. thesis, Texas A&M ITniversity, College Station, 108 pp.

Dice, L. R., and H. H. Teraas. 1936. A graphic method for comparing several sets of measurements. Contrih. l.ab. V^ert. Cienetics, Univ. Michigan, 3:1-3.

CfRINNei.e, J., and J. Dixon. 1921. Revision of the genus Lynx in California. Univ. California Publ. Zool., 21:339-354.

Hale, E. R. 1981. I’he mammals of North America. Wiley-Interscience, New York, 2:vi+601-181 l+9fA

Hale, E. R., and K. Keeson. 1959. The mammals of North America. Ronald Press, New York, 2:viii+547- 1083+79.

IIeewic, j. E. 1976. A user’s guide to SAS-76. Sparks Press, Raleigh, North Carolina, 329 jjp.

UiNiiART, S. B., AND F. E. Knoweton. 1967. Determining age of coyotes by tooth cementum layers. J. Wildlife Manage., 31:362-363.

UoNt;, C.A. 1968. An analysis of patterns of variation in some representative Mammalia Part I. A review of estimates of variability in selected measurements. I'rans. Kansas Acad. .Sci., 7:201-227.

SC:HMI1)1.^ and rkad— cranial x ariai ion

I.owKRY, (;. H., Jr. 1971. I lu* mainiiuils of Louisiana and its atljatcni

waiers. I.ouisiana .Slate Univ. Press, Baton Rouge, xxiii+.‘j6.5 pp.

Mkrri.wi, (;. 11. 1890. Results of a biological survey ol the San Krantisco Mountain region and desert ot the Little Colorado, Arizona. N. Amer. Fauna, .S: 1- 130.

Pktkrson, R. L. , AND S. (;. Downino. 19,52. Notes on the bolxais (Lynx rufus) of eastern North America with a description of a new race. Contrih. Royal Ontario Mus. Zool. Palaeo. , 333:1-23.

Roiii.F, F. J., AND J. Rishpaugh. 1972. Numerical taxonomy system of multivariate statistical programs. State Univ. .New York at Stony Brook (“Manual” printout from program].

.Sa.m.son, F. Ci. 1979. Multivariate analysis erf cranial characters among berbeats with a preliminary discussion of the number of subspecies. Pp. 80-86, in Bobcat research conference proceedings (L. G. Blum and P. Ck Kscherich, eds.), Natl. Wildlife Fed. Sci. Tech. Ser. , 6:1-137.

\’an Zyi.i. de Jong, C. Ci. 1974. Differentiation of the Canada lynx, Felis (Lynx) canadensis subsolana, in Newfoundland. Canadian J. Zool., .53:699-70.5.

Acfdres.ses of authors: Departrneyil of Wildlife and Fisheries Sciences, Texas A6-M University, College Station, 77843. Received 2 January 1985, accepted 18 March 1985.

PUBLICATIONS OF THE MUSEUM TEXAS TECH UNIVERSITY

Three publications of The Museum of Texas Tech Univer¬ sity are issued under the auspices of the Dean of the Gradu¬ ate School and Director of Academic Publications, and in cooperation with the International Center for Arid and Semi- Arid Land Studies. Short research studies are published as Occasional Papers whereas longer contributions appear as Special Publications. Papers of practical application to col¬ lection management and museum operations are issued in the Museology series. All are numbered separately and pub¬ lished on an irregular basis.

The preferred abbreviation for citing The Museum’s Occa¬ sional Papers is Occas. Papers Mus., Texas Tech Llniv.

Institutional subscriptions are available through Texas Tech Press, Texas Tech University, Lubbock, Texas 79409. Individuals can purchase separate numbers of the Occasional Papers for $2.00 each from Texas Tech Press. Remittance in U.S. currency check, money order, or bank draft must be enclosed with request (add $1.00 per title or 200 pages of publications requested for foreign postage; residents of the state of Texas must pay a 5 per cent sales tax on the total purchase price). Copies of the ‘‘Revised checklist of North American mammals north of Mexico, 1982” (Jones et al., 1982, Occas. Papers Mus., Texas Tech llniv., 80:1-22) are available at $1.00 each in orders of 10 or more.

ISSN 0l49-17.'iX

Eexas Let h Press Lubbock, Eexas 79409

OCCASIONAL PAPERS

THE MUSEUM

TEXAS TECH UNIVERSITY

NliMBER 102

26 AUGUST 1986

THE GREATER HORSESHOE BAT, RHINOLOPHUS FERRUMEQUINUM (SCHREBER, 1774), IN GREECE, WITH DESCRIPTION OF A NEW SUBSPECIES

Joan Iliopoulou-Georgudaki and John C. Ondrias

The greatei horseshoe bat, Rhinolophus ferrumequinum, is widely distributed in Europe, Asia Minor, Persia, and the Arabian Peninsula, eastward through southern Asia to Japan. In North Africa, it is known from Algeria and Morocco and may occur elsewhere on that continent as well. A number of subspecies have been described (Felten et al, 1977), of which the most acceptable seem to be R. /. ferrumequinum (Schreber, 1774), R. /. martinoi Petrov, 1941, and R. /. proximus Andersen, 1905. I he recognition of other named races is questionable.

I he species has been recorded from Greece from the following localities; Parnassos and Syros (Miller, 1912); Rhodos (de Beaux, 1928; Festa, 1914); Olympus (Chaworth-Musters, 1932); Skyros (Pohle, 1953); Kimmeria (Thrace) (Linderg, 1955); Corfu (Niethamrner, 1962); loannina (Fpirus) (Felten et al, 1977); Crete (Bate, 1906; P'elten et al, 1977; Kahrnann, 1959; Miller, 1912; Pieper, 1977).

Materfaes and Meiifods

The present study is based on examination of 120 specimens from continental Greece and 17 from the Island of Cirete. Fhe localities of collection are listed in the accounts beyond. Most specimens are preserved in alcohol and are deposited iti the Zoological Museum of the Ihiiversity of Patras (ZMUP), Greece.