226 



PROBLEMS OF RELATIVE GROWTH 



the largest. It is suggested that it would be most advantageous 

 to have a linear relation (exponent = i-o), but that physio- 

 logical difficulties connected with surface-area (which would 

 give a limiting value of 0-67) interfere with this, and that they 

 interfere increasingly with increased absolute size. It is inter- 

 esting to find the large Ratites lie very close to the theoretical 

 curve, in spite of their relatively 

 huge bulk and their flightless 

 habits. (Fig. 97 ; see also p. 264.) 



Body-weight, q. 

 300 400 600 BOO 1.000 



SO 



Body-weight, q. 

 8 iol"PP«"»""20 



1 ■ 



1 1 r 



4 000 6000 8.000 10.000 70000 30.000 40,000 100.000 



Body-weight q 

 (lower cur.el 



Fig. 97. — Graph showing change of egg- weight with body-weight in 432 species 

 of birds : Means by weight-classes, logarithmic plotting. 



o, Carinatae ; X, Ratitae (excluding Apteryx, which is plotted separately) ; K mean for last class 

 of Carinatae combined with Ratitae (except Apteryx). 



The graph begins with the upper curve ; k = i-o. The value of k then sinks (dotted line), finally 

 approximating to 0-67 (solid line, below). 



Each large group has its own curve, differing chiefly in the 

 fractional constant b. The modifiability of relative egg-size 

 in relation to adaptive needs is clearly seen in the abnormally 

 low position of the value for species which are reproductive 



