Giles: Multivariate Analysis of Pleistocene and Recent Coyotes 377 



Seven coyote mandibles from the tar seeps at McKittrick, California, were 

 measured in addition to the main groups studied for this paper. Schultz (1938) 

 assigned these coyotes to Canis latrans orcutti, the subspecies found at Rancho La 

 Brea, and, like other authors, commented on the great variability apparent in a 

 coyote population. The contents of this trap, like those of the Los Angeles pits, 

 have been referred to the Late Pleistocene, with the possibility that those of McKit- 

 trick are slightly younger. Some small ecological difference may have existed 

 between McKittrick and Rancho La Brea, which is 110 miles to the southeast, but 

 this difference probably had no effects on the free-ranging coyote. 



Six mandibles of coyotelike canids from Randall County, Texas, were also 

 measured. These animals, described by Johnston (1938) as a distinct species, 

 Canis lepophagus Johnston, came from beds assigned Blancan age. Johnston's 

 diagnosis of the species states that the C. lepophagus cranium is more slender, has 

 stronger crests and postorbital processes, stronger lower jaws, relatively longer 

 teeth, and is generally somewhat smaller than that of C. latrans. Although the 

 canid remains occur in two beds separated by a 10-foot stratum lacking mammal 

 remains, evidence indicates that the fauna is localized and represents no great time 

 interval (Johnston and Savage, 1955). 



Table 2 lists the specimens studied. Eighty-six per cent of the Rancho La Brea 

 specimens are from the Los Angeles County Museum, the rest from the University 

 of California Museum of Paleontology. The McKittrick specimens are from the 

 Vertebrate Paleontology Collection of the California Institute of Technology (now 

 in the Los Angeles County Museum), the Texas canids from the University of 

 California Museum of Paleontology, and the Recent specimens from the University 

 of California Museum of Vertebrate Zoology. 



SAMPLING CONSIDERATIONS 



The Recent coyote specimens were chosen to represent the smallest possible separa- 

 tion in time and space. The Canis latrans mearnsi sample was the most restricted 

 geographically: all the specimens used were collected in Cochise County, Arizona, 

 most of them from the southeastern part of the county, and were acquired over a 

 span of thirteen years (1919-1932). C. I. lestes specimens were collected in 1916 in 

 Lassen County, California. Unfortunately, the available specimens of C. I. ochropus 

 have been taken over a fairly wide area and in the course of about thirty-five 

 years (fig. 1). 



Grinnell (1933) suggested that the great variation among coyotes may arise 

 partly as an artifact of museum collections: a collection built up over time, even 

 though from the same locality, may represent two populations, or even two sub- 

 species — one having moved in rapidly after the first was exterminated. Jackson 

 (Young and Jackson, 1951) , however, discounts this argument, but samples for my 

 study were chosen to minimize this possibility. 



It is very likely that the specimens of Canis latrans mearnsi and C. I. lestes are 

 samples of distinct, potentially interbreeding aggregates. The time and geogra- 

 phical spread of the sample of C. I. ochropus makes this supposition less probable 

 for it. 



The nature of the group taken here as "Rancho La Brea Coyotes" is much less 



