Structural Variation among the DifflugianRMzopods. 1^1 



more frequent in the mitriform than in the globular series. 

 It also takes place more frequently in certain varieties than in 

 others of the globular series, as will be seen on reference to figs. 

 17, 18, & 27 on the one hand, and figs. 19, 20, 22, & 23 on 

 the other. 



But the selection of mineral particles is not confined to strictly 

 inorganic substances. The Diffiugia seem to know that in 

 the valves of the Diatomacese are combined the properties best 

 suited to their wants — that is to say, transparency and forms capa- 

 ble of being easily arranged, — at the same time that the diatoms 

 occur as epiphytes on the aquatic plants upon which they them- 

 selves frequently find feeding-ground. And it is a remarkable cir- 

 cumstance that we can generally tell whether diatoms are or are 

 not plentiful in a given locality by observing the share taken by 

 them in the composition of the Difflugian tests. Thus, to cite an 

 example close at hand, in the Hampstead pools the predominant 

 diatoms are Pinnularia and Eunotia — the former of very large 

 size, the latter extremely minute *. Hence the first is but rarely 

 seen impacted into the tests, and when present it is of medium 

 size. The second, however, constantly occurs, and, in the curious 

 variety of test referred to as D. spiralis, and likewise in the glo- 

 bular and lageniform series, they often constitute a very large 

 percentage of the mineral matter (figs. 18, 18 a, 246, &32t). 



In the Greenland mountain streamlets from which I obtained 

 some of my specimens, Eunotias occur very abundantly, and under 

 a remarkable variety of forms. In some tests they constitute the 

 entire mass of mineral matter, and, as in the Hampstead ma- 

 terial, pervade nearly every shape of test, though much less pre- 

 dominant in the depressed series (figs. 9 & 27). 



I have now to speak of several novel modifications in the out- 

 ward characters of the Difflugian tests which have heretofore been 

 observed by me, for the most part amongst the varieties of the 

 mitriform and lageniform series, and which I have been able, only 

 within the past few months, to trace as an unbroken chain from 

 the forms already described to those in which the chitinoid matrix 

 presents no appreciable admixture with mineral matter. 



The first indication of this very remarkable and instructive 



* This diatom would seem to have been hitherto undescribed. It is 

 stipitate, the valves being generally from tj^'o? to TeW of an inch in length, 

 and occurs in crowds around a filamentous " frond," appearing to be very 

 generally distributed. An extremely minute Navicula also occurs, but is 

 rarely emjiloyed in the construction of the test — thus aflfording another 

 example of selective power. 



t The two figures 24 & 24a are not intendfcd to show these diatom- 

 valves ; but certain cylindrical or, as I formerly called them (' Annals,' 

 June, p. 451), pellet-shaped cylinders, which will be more particularly re- 

 ferred to presently. 



