758 



SCIENCE 



[N. S. Vol. XXXVI. No. 935 



left end of the polygon of frequency further 

 to the right than the right end would be 

 shifted. This produces a narrowing of the 

 range and decrease of variability (Fig. 17a). 

 The result is a decrease of natural selection, 

 unless the survival curve is more inclined at 

 the new position (Fig. 166). In the third 

 case, those with the characteristic in ques- 

 tion innately larger sustain the greatest modi- 

 fication. This might be illustrated in the case 

 where only some moths which could reach the 

 nectar of a particular flower would make con- 



FiG. 17. The straight solid line is the survival 

 rate; the dotted line, the polygon of the indi- 

 viduals as modified. To illustrate action of co- 

 incident selection in decreasing and increasing 

 variation. 



tinned effort to do so. By increasing the range 

 (Fig. 176) an increased efficiency of natural 

 selection is produced, even though the polygon 

 is not transferred to an increased inclination 

 of the survival curve. It is probable that the 

 greatest importance of coincident selection 

 lies in its efficiency to promote incipient or- 

 gans in this way, for it is easy to see how im- 

 portant it would be in such a caise as the 

 transformation of a somewhat prehensile 

 snout to the elephant's trunk. On the other 

 hand, it must, in many cases, slow up the effi- 

 ciency of natural selection, when an organ has 

 passed the " infant industry " condition, and 

 thereby enters the second category referred to. 

 The study of the analyses which have been 

 presented brings out some facts as to correla- 

 tion of characters which should be mentioned 

 here. It is often assumed that every character 

 that is selected has a direct survival value. 

 Some authors have recognized that a char- 

 acteristic may be carried along by correlation 

 with some related characteristic which has 

 survival value. In these cases, the correlated 



characteristics are generally thought to be 

 two distinct structures, for the correlation of 

 which a reason is evident. It will be noticed, 

 however, that in the case of the sparrow 

 catastrophe, the several dimensions, aside 

 from the length, are all selected in the posi- 

 tive direction and that they are all dimen- 

 sions of head or limbs except the length and 

 one body measurement, which is taken in the 

 cephalo-caudal direction. The one character- 

 istic not of this sort, weight, had an indiffer- 

 ent survival value. If we try to reason out 

 particular advantages in each of these char- 

 acter changes, we are hard pressed. 



Is it not probable that they are all corre- 

 lated with a physiological characteristic or 

 two, which has its external manifestation in 

 a general increase of peripheral parts and 

 slenderness? This is not fantastic, for we 

 have something analogous in human pathol- 

 ogy. The disease of acromegaly, caused by 

 some condition of the pituitary body, produces 

 a great increase in height and in the size of 

 the hands and the feet. The survival value of 

 length of antennse in moth pupffi, the number 

 of rostral spines of a shrimp, affected by 

 changed salinity, are inexplicable on the basis 

 of direct adaptation. We are driven to resort 

 to correlation with internal features. But the 

 frequency with which this resort to correla- 

 tion is necessary in the discussion of the ex- 

 periments on survival values seems signifi- 

 cant. Even in case of the frontal breadth of 

 the crab, where a definite value has been 

 postulated, namely, that the narrowness of 

 slit which permits water to reach the gills may 

 be of importance in excluding mud, we are 

 still dealing with a correlated character. 



The multiplicity of structure that we find 

 in animals, particularly in sculpture and 

 markings, is simply bewildering, if we con- 

 sider them all as adaptations. But with an 

 appreciation of the great role played by corre- 

 lation in natural selection, we should expect 

 just such a wonderful maze of characteristics. 

 Just as our modern insistence on the impor- 

 tance of the variations that offer themselves 

 reduces the creative role of natural selection, 



