THE FORAMINIFERA 63 



the shell is dotted over with minute tubercles (not visible in 

 Fig. 7). 



A keel-like thickening runs round the margin of the test, and 

 in some specimens small spines, like the points of a spur, project 

 from it at the places where the septa join the keel. These 

 are more frequently present in the earlier than in the later 

 convolutions. 



The pores traversing the walls of the chambers are in this 

 species exceedingly minute. They are hardly visible when the 

 test is seen from without, but they may be detected when a broken 

 piece of the wall is highly magnified and seen by transmitted light. 



On examining the external characters of the tests of a number 

 of Polystomellas, they are found to form a uniform series, presenting 

 such gradations of size from small to large as may be seen, for 

 example, in a sample of the shells of any Mollusc which contains 

 young and old. If, however, a batch of living Polystomellas is 

 killed by some reagent which dissolves the shell but preserves the 

 protoplasm filling its chambers, the protoplasmic casts of the 

 shells no longer form a uniform series but fall into two sets (Fig. 8). 



In the great majority of them the series of chambers, when 

 traced to the centre of the spiral, is seen to take its origin in a 

 large spherical chamber, having a diameter generally between 60 

 and 100 p.. In the others a small central chamber, with a dia- 

 meter of about 10 yu,, 1 occupies the centre of the test, and the suc- 

 ceeding chambers of the series are at first correspondingly small, 

 so that for a given diameter of test these specimens have a greater 

 number of chambers than the others. 



It is clear that though in Polystomella there is no marked differ- 

 ence in the size attained by the two forms, we have here the same 

 phenomenon of dimorphism which is exhibited in the nummulites. 



The large central chamber is known as the megalosphere, the 

 small one as the microsphere, and the two sets of individuals of the 

 species are known as the megalospheric and microspheric respectively. 

 The numerical proportion of the two kinds of individuals probably 

 varies with the season, but the megalospheric form is here also 

 always the more abundant. In a large batch of several hundred 

 specimens the megalospheric forms were found to be thirty-four 

 times as numerous as the microspheric. 



In the protoplasmic casts obtained in this manner the form 



1 The diameter of the microsphere varies in the specimens of P. crispa which I 

 have seen from 6'5 to 13 jJ~ That of the megalosphere from 165 to 35 /tt. These 

 dimensions fall, however, a little short of the actual diameters of the chambers, owing 

 to the shrinkage of the protoplasm produced by the reagents. When comparing the 

 size of the microsphere in specimens preserved in this manner and in those examined 

 by sections of the test, it is well to bear this cause of difference in mind. 



The number given as the diameter of a central chamber, in this article, is to be 

 understood as the mean between the long and short diameters as presented for 

 observation in the specimen. 



