MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 421 



large spbernles ou angular blocks, wliicb are abundantly perforated with round lacunaB. Crangou 

 agrees essentially at this stage with at least three other species, namely, with Alpheus saulcyi, 

 Pontonia domestica, and Steitoi>us hispidits. 



It is probable that Figs. .3 and 4 ol Kingsley's paper do not correspond, the latter representing 

 the older egg, that is, older than the sixtcen-cell stage. In the segmenting eggs of Homarus there 

 is a striking individual variation, which finds expression chiefly in the external characters of seg- 

 mentation, but 1 have never observed this in any related forniR, and can not say to what extent it 

 occurs. Jiulging from analogy, I think it will be found that as a rule all the protoplasm in the egg 

 of Crmif/oH TuUjaris reaches, or very nearly reaches, the surface, as in Homnrus and Alphciis snulcyi, 

 and that toward the ch»se of the yolk-])yrainid stage dclamination 0(;curs and some of it wanders 

 back into the depths of the yolk. If this is true, the "belated" cell near the center of I'ig. 4 of 

 Kingsley's i)aper may represent some of the jirotoplasm which has taken this roundabout. journey. 



(While this memoir was in press a paper has appeared by W. F. K. Weklon, on "The Forma- 

 tion of the (ierm layers in Orangon vulgaris." {Quart. Jour. Mir. iS'ci., Vol. xxxiir, ])t. .'i, March, 

 1S92). Lie makes no mention of the biulding of cells from the blastoderm before invagination, 

 nor of the presence of migrating cells at a later period not derived from the invaginate cells or 

 ventral plate, hence it is probable that in Crangon primary yolk cells do not occur. In this case 

 the suggestion just ottered does not let us out of the difhculty.] I have not succeeded in obtaining 

 the segmenting eggs of Callianassa for comparison with the early stages of CaUianasna mediterranea, 

 described by Mereschkowski (40). It seems to me, however, that the diagrams in Mercschkow- 

 ski's paper are misleading, and that the process of segnu'utation in Calliana.ssa, instead of being 

 peculiar as one might infer, is essentially typical. According to this observer's account the "blas- 

 toderm" ari.ses without yolk segmentation. Nuclear division is at first central, and the resulting 

 cells, sixteen in luimber, pass gradually to the surface and form a deep layer of |irotoplasm 

 inclosing the yolk. This lajer, at tirst raised into hillocks corresponding to the nuclei, tinally seg- 

 ments into sixteen parts around the latter. These segments are formed simultaneously over the 

 egg, but the yolk is not involved. The cells multiply rapidly and form a layer of tall prismatic 

 elements, which gradually flatten out as division proceed.s, into ordinary blastoderm cells. 



If there is any analogy between this egg and that of related forms the broad "protoplasmic" 

 layer is really the protoplasm plus the i)eripheral yolk. It would be remarkable in any case if 

 a segmenting egg could accjuire such a mass of proto])lasni, not to speak of the suddenness with 

 which the accjuisition is nuide. That this layer, comprising more than one-half the contents ot the 

 egg, is largely yolk is indicated by the fact that the nuclei which occur in it are representt>d as 

 surrounded by a protoplasmic reticulum as normally occurs. If this is true, the j>cis»in/ic cells are 

 yolli pyrainiih, and their line of separation from the central yolk is purely imagiuarj'. 



Ishikawa found, in his studies upon a Japanese prawn, Atyephyra compreiisa (27), that the egg 

 underwent at first a total and regular segmentation. At the end of the fourth phase the yolk is 

 centralized and the protoplasm surrounds it, and in the next stage, after thirty two blastomeres 

 have been formed, the central end of each separates ofi' to form a yolk segment. The yolk .seg- 

 ments, which till the center of the egg and correspond to the common yolk mass with whic^li the 

 apices of the yolk pyramids blend in other Decapods, are of unequal size and contain nuclei which 

 do not take part in the "blastoderm." These nuclei probably correspond to the delaminated cells 

 of Homarus and Alphens. In the latter they appear at the close of segmentation. Ft is (juite 

 probable that the time at which these cells are budded ofi' may vary considerably in different 

 species. 



In Extpagurnn pridcauxii (39) Mayer found that the protoplasm .segmented tirst in the center 

 of the egg, as in other forms, until eight nuclei were present. When this stage is reat^hed the yolk 

 now .segments not simultaneously into eight blastomeres, as in the ca.se with the Isojiod Asrllus 

 aquaticm, but according to its inherited primitive manner, first into two, then into four and eight 

 spherules. Segmentation of the yolk is thus at first total, but after the fourth pha.se the s])heres 

 unite in a central yolk core as in other forms. 



In Alphens saulcyi I did not find any eggs which showed a progressive segmentation of the 

 yolk between tlwe stages represented in Figs. 9 and 10, and hence I infer, as already stated, that 

 the segmeutAtion of the yolk is (here a simultaneous process for each of the sixteen segments 



