Experimentally Fused Larva of Echinoderms , etc. 117 



and independent of the other except for the common surface of attachment. 

 There is no reciprocal influence of one upon the other. When, however, 

 the agglutinated surfaces extend over a large area the blastulae tend to 

 fuse, to have a common blastocoele, and the gastrulae tend to approximate 

 the form of a single embryo. It can not be stated definitely which is cause 

 and which effect, whether greater common surface makes for more complete 

 fusion or whether the internal forces making for fusion also affect the surfaces 

 of the agglutinated embryos, 



(2) Mass of agglutinated or fused embryos: When more than four or five 

 eggs are agglutinated the disintegration of one or more of them at an early 

 stage in their development brings about the disintegration of the entire 

 cluster before they can fuse. When three blastulae are agglutinated they 

 may develop into three perfect and practically independent larvae, or they 

 may fuse together, under which circumstance the reciprocal influences are 

 exceedingly complex and the resulting larva is quite atypic. For this reason 

 I have omitted all consideration of such complex fusions and limited myself 

 to fusions of two eggs, so that the regulative changes may be observed with 

 the least number of disturbing factors. 



(3) Differential rate of development: When two gastrula are fused and 

 each has developed at an equal rate, twin larvae like that shown in figures 

 5 and 6 result, in which case each larva is practically complete and equal in 

 size, and has little or no influence upon the other. When two gastrula have 

 not developed equally the conditions are then most favorable for form regu- 

 lation. 



At least two factors are involved in form regulation, namely, the sup- 

 pression of structures and the disintegration of those already differentiated. 



I have elsewhere shown, from the study of living fusions, that the 

 archenteron of one of the fused gastrulae may be entirely inhibited or may 

 never develop beyond an early stage of its differentiation. Driesch and 

 de Haan have also observed aborted or dwarfed archentera. I have shown 

 in this paper that certain parts of the skeleton also may never be developed, 

 that such suppressed parts are nearly always the last to be differentiated 

 and confined to but one of the fused larvae. 



It has been urged by de Haan that the suppressed larva is a sick larva, 

 that there is no regulation but that a union of a healthy and an unhealthy 

 larva occurs. Under such circumstances we should expect the sick larva 

 to be atypic and some relation between diminutive size and degree of ir- 

 regularity in development. But the fused larvae are quite perfect and often 

 of normal size. It might be urged that only very sick larvae would be atypic 

 and would not complete their development. But how can one account for 

 the lack of development of the right or the left sides of the skeleton on such 

 an assumption? It would also be exceedingly difficult to account for the 

 absence of the oral half of one side (fig. 14), particularly in the face of the 

 fact that the skeletal parts that are differential are characteristic and normal 

 and show no trace of irregularity or sickness. 



