EMBRYONIC RECONSTITUTIONS 537 



Later they form a compact rounded mass of cells with more or less definite 

 outer layer but without any indication of definite embryonic structure or 

 polarity. Still later, outgrowths from this primary embryo develop, some- 

 times becoming elongated, finger-like extensions; from the tips of these 

 outgrowths cell masses, varying somewhat in size, separate as secondary 

 embryos. These may develop or in certain species give rise to further 

 outgrowths from which tertiary embryos develop.^" There is no indication 

 of polarity in the early stages of these masses. According to Harmer, how- 

 ever, the outgrowths develop only toward the distal end of the ovicell in 

 certain species of Crista. This perhaps indicates presence of a polar pat- 

 tern determined by some differential in the ovicell. In other species of 

 Crista Robertson finds outgrowths developing in any direction. Each out- 

 growth probably represents a localized region of increased cellular activ- 

 ity like a bud, the locahzing factor being apparently more or less fortui- 

 tous, and each mass separating from the outgrowth may develop polar pat- 

 tern from the part of the longitudinal pattern of the outgrowth persisting 

 in it; but, even if this is the case, the origin of ventrodorsality remains ob- 

 scure. That axiate physiological pattern is acquired at some time in the 

 course of this polyembryonic development and in relation to some differ- 

 ential or differentials in conditions and that a definite egg pattern is not 

 necessary for development of an axiate individual in these forms is highly 

 probable. 



Development of two whole embryos from a single egg and various de- 

 grees of embryonic twinning, involving anterior or posterior duplications 

 or both, are not infrequent in ohgochete annelids.^' These duplications, 

 like some experimental duplications discussed in the following chapter 

 (pp. 556-61), are of special interest as indicating that annelid development 

 is not as completely a ''mosaic" of self-differentiating parts, as often 

 assumed. 



In certain parasitic hymenoptera there are extreme degrees of polyem- 

 bryony, although maturation and early cleavage suggest presence of axiate 

 pattern of some sort and degree. If such pattern is present, there is no evi- 

 dence that it persists or is effective in later stages. By continued cleavage 

 the primary embryo becomes a solid mass of cells, apparently without 

 definite arrangement or visible pattern. Constrictions divide this embryo 

 into a number of cell masses, constituting the polygerm; and these pri- 

 mary masses divide to form secondary masses; these, in turn, to form 



^"Harmer, 1893, 1896, 1898; A. Robertson, 1903. 



" Kleinenberg, 1879; Vejdovsky, 1883-92; Weber, 1917; Welch, 1921; Penners, 1924a. 



