260 



THE AMERICAN NATURALIST [Vol. XL 



the details of the formation of prothallial cells, but in the species 

 of Zamia studied there were always at least two prothallial cells 

 cut off at one side and projecting into the grain as in Ceratozamia 

 and Macrozamia, and different from Cycas and Ginkgo, where 

 the walls are straight across the grain and not arching into it. It 

 seemed to Webber, however, that three prothallial cells were 

 occasionally formed, and in this case the first one was resorbed 

 as described by Strasburger and others in Pinus, Ginkgo, etc., 

 remaining as a dark refractive layer in the wall of the pollen grain, 

 situated at the point of contact of the other cells. In many in- 

 stances in mature grains, and in later stages, during germination, 

 no indication of this resorbed prothallial cell can be observed, 

 but in some cases it is unmistakable. A careful investigation of 

 the development of the pollen in Zamia will have to be made before 

 it can be determined whether three prothallial cells are regularly 

 formed or whether the remnants of a third cell, occasionally ob- 

 served, are to be considered as cases of rare and somewhat abnor- 

 mal development. 



^yebber (:01, p. 24) clears up a point in the origin of the stalk 

 cell and the central cell (body cell of Strasburger) showing that 

 they originate by the division of the inner prothallial cell. Up to 

 tlu^ time of Webber's full publication this point was quite obscure 

 and rhis brings it in agreement with the facts in other gj^nosperms 

 investigated, except that the sterile cell bulges strongly into the 

 stalk cell and this had led to a wrong interpretation. Webber also 

 shows that Ginkgo is probably similar to Zamia on this point, and 

 Seward and Gowan (:00, p. 130) show that this is actually the 

 case. They also state that in Ginkgo the pollen grain at maturity 

 contains a prothalliinn of from 'A to 5 cells. 



Lang ('97) rcpoit> that Sfanr/rria paradoxa has two prothallial 

 cells when poUcii is sh<'(l, l)iit liis Figure bS, Plate 22, seems to 

 show three. 



An interesting <,,iestion is suggested by th<' .iilVerent results 

 observed in two >|H-rir> of the getius F.phe.h-a. Jaecanl i^) 

 reported that in A'. Iir/nflni at the time of anthe^is thi-re were in 

 the pollen grain three free nuclei, but not separate eelU, aud he 



ative nucleus, and the tube nucleus. Land (:04) in Ephedra 



