26 Phillips, Problems in Migration. [jan. 



up from a flock of two to a flock of five, the apex being at the 

 number seven. Then the curve falls to a right hand base at the 

 number nine. Above this number there are small rises in the curve 

 at various points, notably ten, fifteen and twenty, which may have 

 some significance, but as we go up into the larger flocks, the numbers 

 (number of flocks counted) are smaller, and the counts themselves 

 perhaps not so accurate. The "small flock" frequency curve 

 (flocks of one to fifteen in number) is composed of counts of 262 

 different flocks. It gives us the usual size as six or seven birds, 

 the next commonest number being five. The scarcity of flocks 

 of eight and nine is remarkable. Only ten flocks of nine each were 

 observed, while there were thirty-seven flocks of seven each and 

 thirty-five flocks of six each. The actual average size of a flock 

 of autumn migrants in Massachusetts is not considered here. 

 It is much larger, nearer thirtj^-five, because very large flocks of 

 100 to 300 are not rare. These large flocks need not concern us. 



It is of interest to note that the small flocks, when they are 

 captured entire, show from inspection of various external age 

 characters — size, roughness of soles of feet, development of wing 

 spurs, etc. — a pair of old birds and several young, two to five, or 

 even six. It is especially easy to pick out the young birds from 

 the old in early October flights, because the young have had less 

 time to develop, but it is not always possible to do this with 

 certainty. We are speaking, of course, of the fall migration only. 

 Whether these facts, special behavior and individual make-up of 

 the " small flock," hold for the vernal migration is unknown to me, 

 and it would be harder to ascertain. 



I give below a curve of frequency for flocks of geese of from one 

 to thirty in number. The peaks of the curve at ten, fifteen and 

 twenty may mean combinations of two, three, or four families: 

 but as noted above the observations are hardly numerous enough 

 for the larger flocks. The striking fact is the very sharp rise of 

 the curve to six and seven. The frequency of ones and twos and 

 even threes and fours is much too high, because I have had to 

 include in the curve a number of geese undoubtedly left over from 

 shot-up flocks. In the natural state of affairs there would probably 

 be very few flocks below the number four. Flocks of eight are 

 perhaps mostly one large family, while flocks of nine are composed 

 of small families. 



