XVI NOTE. 



(6) Width of hill, at iase, measured across the chin l)ctween the out- 

 side of the gnathidea at their base. 



(7) Tarsus^ measured from the tibio-tarsal or heel joint 07i the outer 

 side to the lower end. This last point is often more or less indefinite 

 and sometimes very difficult to locate, but in such cases may usually 

 be ascertained by flexing the toes. 



(8) Middle toe, measured from the lower end of the tarsus to the base 

 of the claw, the length of the latter not being included, unless so stated. 



(9) GradiMtion of tail, measured from the extremity of the outer- 

 most rectrix to that of the middle or longest, the tail being closed. 



Owing to the considerable individual variation in measurements in 

 almost all birds, measurements of a single specimen are of little value as 

 part of a description of a species or subspecies. In such a case the meas- 

 urements are quite as likely to represent the maximum or minimum 

 dimensions of the form as the average. Therefore, in connection with 

 the present work, a series of specimens of each form has, whenever 

 practicable, been carefully measured, and the minimum, maximum, 

 and average of each separate measurement given with the descriptions. 



The average measurements, as given, do not always express actual 

 difference or agreement, as the case may be, between allied forms 

 because, in the first place, the series measured are too often unequal and 

 inadequate, and in the second, determination of sex by collectors is 

 not always to be relied on, measurements of one sex sometimes almost 

 certainly figuring among those of the other. Besides, a certain allow- 

 ance must be made for errors of measurement, it being a well known 

 fact that the same person can rarely measure the same specimen twice 

 and obtain exactly the same results. It may be stated, however, that 

 great care has been taken to have the measurements given represent the 

 facts as nearly as possible, by selecting as nearlj^ equal a series of each 

 allied form as the available material would allow, and by eliminating 

 immature specimens and those with abraded primaries and rectrices, 

 as well as those whose sex had not been determined by the collector or 

 is otherwise not obvious. 



In describing the wing-formula or relative lengths of the primary 

 remiges, these are counted from the innermost one outward, not con- 

 versely, as has been the general practice. Hence, in "ten-primaried" 

 birds the outermost primary is designated as the tenth, and in "uine- 

 primaried" birds as the ninth. The advantage of this innovation, 

 which no doubt will at first cause inconvenience, is that in the case of 

 "nine-primaried" birds the outermost obvious primary (the ninth) 

 receives its correct numerical designation, whereas according to the old 

 inethod of numeration it was designated as the first while being in 

 reality the second.' 



•See Forbes, Proc. Zool. Soc. Lond., 1879, p. 256, note 2; Gadow, iu Newton's 

 Dictionary of Birds, pp. 118, 741, 780. 



