FORAMINIFERA OF THE ATLANTIC OCEAN. 5 



the spicules are few and 1 have never seen one with a short or broken 

 spicule, but always with a very long uninjured one. 



P. testacea builds its test of other foraminifera and lives especially 

 as would 1)6 expected in globigcrina ooze. The tests are not alike 

 nor of the same size nor shape, but sand grains are almost never used, 

 while in the same dredge haul may be other genera and species 

 largely made up of sand giains. 



In P. howmanni there is a selection by which only mica flakes are 

 used, these being cemented together by their edges, making a weak 

 and irregidar test. Such specimens, however, rarely show any sand 

 grains and the selective power must be considerable, for in most 

 bottom material the amount of mica flakes is not great. 



Lastly, in P. nistica is a species with an even gieater ingenuity'. 

 It uses large acerose spicules for the main lines of its polygonal test, 

 tlien tills in tlie sides with broken spicules, fitting each to the poly- 

 gonal area l^etweeu the three or more borders of that surface. The 

 long edge spicides are the only ones that extend beyond the face of 

 the wall, the others being fitted as though cut off at the various 

 lengtlis. Tlie only explanation of the budding of sucli a test as this 

 is that the material is ingested in the protoplasm and then at a certain 

 stage carried to the outside of the protoplasmic body to form the test , 

 and that the distribution of the inner broken spicules is mechanically 

 arranged and the whole cemented. 



In the genus Teclinitella there is also a marked selection. T. melo, 

 for example, has a rounded test built entirely of sponge spicules, 

 these placed lengthwise of the test and firmly cemented. In T. 

 legumen, which is sometimes found with the former species, fine 

 amorphous white material is also used with the spicules and two 

 layers of spicules are distinguished, the inner running transversely 

 and the outer lengthwise. As a result a strong test is developed 

 when the amount of spicules is considerable. In T. thompsoni there 

 is a very unique condition in which the test is made up of the dis- 

 integrated plates of a r>rittle star-. The amount of these plates in any 

 given area can not be very great, yet the animal obtains sufficient 

 numbers of them to build its test from these entirely, using probably 

 hundreds of individual ])lates in the process. 



