Mar. II. 1918 Sterility in the Strawberry 645 



that these stages bearing bivalent loops are presynaptic. Gradually the 

 synaptic mass tightens until it is close and compact, occupying a very 

 small portion of the nuclear cavity (PL B, 4). During this contrac- 

 tion there are refractive particles present both in the mass, some of which 

 are in contact with the nucleolus, and outside of the nuclear membrane, 

 which give the same staining reactions as the nucleolus. Similar masses 

 to these Digby (jj) has considered to be synaptic extrusions. 



The synaptic stage is of long duration. Gradually the chromatin in 

 synapsis takes on the appearance of being made up of a closely tangled 

 mass of threads. Soon loops are pushed out from it, which are bivalent 

 often for their entire length (PI. B, 5). The fact that these loops are 

 often double from the point at which they leave the mass to the point of 

 entrance, and can sometimes be traced through a portion of the synaptic 

 mass gives the impression that they are made up of two continuous 

 threads which closely approximate each other over their entire length. 

 This view is supported by the later stages, especially those at and follow- 

 ing segmentation, which in Fragaria spp. are very clear. 



Loops continue to push out from all sides of the synaptic mass, often 

 shifting it to the center of the nucleus. Gradually the bivalent thread 

 becomes more or less regularly distributed about the nuclear cavity, 

 usually having, however, a somewhat tangled center near the nucleolus. 

 The spireme thread at this stage often appears to be a single strand 

 due to the dose approximation of its univalent portions (PI. B, 6). 

 However, no anthers at this stage of development have been found which 

 do not contain many portions of the spireme which are double for eon- 

 siderable distances. It is probable that, during the post synaptic stages 

 up to segmentation of the spireme into chromosomes, the univalent 

 portions of the thread are never fused throughout their whole length to 

 form a single spireme. It is even possible that no fusion takes place, 

 but that the imivalent threads only approach each other so closely that 

 in such delicate threads the line of demarcation can not be distinguished. 



There is no distinct second contraction, but there is a semblance of 

 one following the loosely-distributed spireme stage. The thread con- 

 tracts gradually but unevenly throughout its whole length, its univalent 

 portions as a consequence becoming separated from one another and 

 appearing thicker (PI. B, 7). The portions of loops which are in 

 contact with the nuclear membrane remain so and often extend long 

 distances on the periphery and then turn at relatively sharp angles and 

 again extend in fairly straight lines toward the central mass, still situ- 

 ated usually near the nucleolus. Many of the loops in this way form 

 equilateral triangles. There is no doubt at this time of the double 

 nature of these loops. The paired threads are evidently identical with 

 those which appeared while passing out of synapsis and may be identical 

 with the bivalent loops seen extending out from the loose presynaptic 

 38325°— 18 3 



