Race Crossing in Man (Eugenics Lab. Mem. XXXVI) [Review]

Race Crossing in Man (Eugenics Lab. Mem. XXXVI) [Review]

American Journal of Human Genetics
Volume 6, Number 1 (March 1954)
pages 195–196

Kenneth S. Brown
University of Chicago

By J. C. Trevor, New York: Cambridge University Press, 1953, Pp. 45

This brief monograph is a mixed blessing. On one hand it demonstrates what a wealth of metrical material has been collected on human hybrid populations, while on the other it presents methods of analysis which are both inefficient and ineptly applied.

In his preferatory note, Dr. L. S. Penrose points out that this work was done prior to the start of hostilities in 1939, but that the value of the data presented is timeless. This is very true, however the analysis made of the data is rapidly showing signs of age. The data are a compilation of published records of nine outstanding cases of biracial crossing; Hybrid American Negroes, Jamaican ‘Browns’, Half-Blood Sioux, Ojibwa-Whites, Yucatecans, Rehoboth Bastaards, Kisar Mestizos, Norfolk Islanders, and Anglo-Indians. The mean and standard error are recorded for stature and seven cranial measures for most of these populations. Additional measurements are noted for many. In each population studied the sample includes 25 or more adult (20 years or over) individuals of each sex. The values for each sex are recorded separately.

The mean of the hybrid population is compared with that of each of its propositus population groups by the use of Student’s t test. Unfortunately this test requires the assumption that the variances of the populations being compared be the same. Nowhere in the presentation is this recognized. It would have been eminently desirable to determine the significance of the variance ratio for each parameter for each pair of populations compared before the t test was applied. For cases of significant difference in variance between the populations the Fisher-Behrens method for the use of t with samples of unequal variance would be applicable.

The variability of the propositus and hybrid populations is considered separately by the method of Mourant in which the variance ratio of each character for each population pair is found and then the mean variance ratio for each pair determined. This analysis indicated that the variance of the hybrid population is greater, but was found, by a t test, to be significantly greater in only two cases. Here it would have seemed desirable to look up the values in a table of F to get a more powerful estimate of the difference between these populations.

The material presented in this monograph provides a good addition to the blood group, dermatoglyphic, and taster frequency data which are currently used in the analysis of population dynamics, and should serve to attract the attention of interested workers to this relatively undeveloped body of information.

Tags: ,