° ' 192 o J Figgins, Races of Branta canadensis. 95 



Consultation of the several authorities appears to be of scant 

 assistance, other than to reveal an apparent acceptance of the varie- 

 ties as a means of escaping a troublesome question, or a seeming 

 attempt to defend them upon purely geographical grounds; al- 

 though this course necessitates a denial of the evidence at hand, 

 questioning of the accounts of authorities of high degree and an 

 appeal to "chances of error" and the "misunderstanding of data." 



Although the various authorities disagree to some extent on the 

 measurements of the several subspecific forms, they are unanimous 

 in concluding that hutchinsi is smaller than canadensis and that 

 occidentalis is larger than hutchinsi. A critical examination of 

 such statements might lead to the conclusion that occidentalis 

 being larger than hutchinsi, its measurements would fall within 

 the extremes of canadensis. Investigation proves this to be true, 

 if the largest and smallest measurements of the various authors are 

 employed for comparison. Continuing the experient further, one 

 finds that only .07 of an inch separates the maximum length of the 

 wing of occidentalis from hutchinsi, according to the early authori- 

 ties. The statement that occidentalis is larger than hutchinsi, is, 

 therefore, based on .07 of an inch in the maximum wing measure- 

 ment — all other wing measurements being within the limits of the 

 latter race. 



On page three of 'A Study of a Collection of Geese of The Branta 

 Canadensis Group From the San Joaquin Valley, California,' 

 Swarth states, in a discussion of thirty-six specimens considered 

 as hutchinsi, "twenty-five are males." Without an explanation 

 of his reasons, he employes but ten of that sex as representative 

 of the differences he describes on page fourteen. It is, therefore, 

 not unreasonable to conclude that the differences he finds in the 

 minimum and maximum measurements of wing, culmen and tarsus, 

 as compared with the findings of other writers, may be due to the 

 elimination of the remaining fifteen males belonging to the series. 

 It is the present writer's experience that the measurements of the 

 tarsus and middle toe obtained from dry skins are not always sat- 

 isfactory and reliable and consequently some doubt may be en- 

 tertained as to the importance of Mr. Swarth's comparisons and 

 conclusions. The same authority shows the number of rectrices 

 in canadensis varies from 14 to 20. The variation is the same in 



