378 University of California Publications in Geological Sciences 



clear than that of the Recent assemblages. The Rancho La Brea fauna is derived 

 from a single locality well within the geographic range of the modern Canis latrans 

 ochropus. Fossil evidence from Rancho La Brea, both plant and animal, corrobor- 

 ates a Late Pleistocene date, not only because of the large number of Recent species 

 and genera present but also by correlation with other deposits known to be of Late 

 Pleistocene origin (Gale, 1932; Schultz, 1938; Grant and Sheppard, 1939; Bailey, 

 1943; Stock, 1953; Durham, Jahns, and Savage, 1954). 



Stock (1953:15) states that the period during which entrapment took place was 

 "a relatively restricted time interval within this [Pleistocene] epoch." Hay (1926) 

 reports that estimates as short as 2,000 years have been made, though he con- 

 siders a much longer period more reasonable. The small variation in contents 

 among pits, which may or may not be statistically significant, may indicate a 

 valid facies difference. If so, such a difference would indicate a lengthy period 

 of accumulation, which might invalidate consideration of all the coyotes as one 

 sample. A change in environment would not necessarily be reflected in the taxon- 

 omy; modern subspecies range through many environmental zones. 



Faunal associations of the single find of human remains (pit 10) clearly indicate 

 an entrapment much later than that of the other pits. Although pit 10 is separate 

 from the rest, it is possible that in the other deposits a few Recent specimens 

 might overlie the Pleistocene material, or be mixed in by the movement of the tar. 

 Generally, the likelihood of contamination has not been considered great (Stock, 

 1953) . Here the Rancho La Brea canids have been considered a homogeneous Late 

 Pleistocene population representing a short period of time. 



In all cases, only adults, as evinced by permanent dentition, were accepted for 

 study: only males were used for comparisons of Recent subspecies, but both sexes 

 were measured for the more general comparisons with the fossil material. Very 

 large coyote samples taken by government hunters indicate that the sexes maintain 

 a 50 : 50 ratio in the wild state (Young and Jackson, 1951). A similar ratio is as- 

 sumed for the Rancho La Brea specimens. In some samples the number of Recent 

 males used was limited by the number of females available. To provide a larger 

 sample, equal numbers of known males and females were augmented by unsexed 

 specimens when these were available and satisfied requirements of time and lo- 

 cation. 



A limitation had to be set on the number of unsexed specimens used, since they 

 might not be half males and half females, and would thus upset the sex ratio in the 

 known-sex group to which they were added — in the extreme case all unsexed speci- 

 mens might be of one sex. The variation expected when drawing a random sample 

 from a population in which both sexes are equally represented can be calculated 

 from the mean and standard deviation of the binomial distribution. For example, 

 using the normal-curve approximation to the binomial distribution: when N -50, 

 the number of one sex may vary from 19 to 31 and still be within 95 per cent 

 confidence limits. In other words, a sample of 50 that contained 19 males and 31 

 females would not invalidate the null hypothesis that this sample was drawn from 

 a population in which the sexes were equally represented. Fewer than 19 males 

 (or females) would make the probability less than 1 in 20 that the sample came 

 from a group containing equal numbers of males and females. The number of 



