52 2 The Ohio Naturalist. [Vol. VI, No. 7, 



The contraction of the spirem continues and is manifested 

 not only by a shortening in length but also by a drawing together 

 of the sprit glanules. This usually goes so far that the longi- 

 tudinal split becomes very difficult or impossible to observe 

 (figs. 6-8). The two loops destined to become the tertads now 

 become more and more definite and begin to break apart (fig. 6) 

 sometimes becoming twisted (fig. 7) in a way that is inconceiv- 

 able were they due to a longitudinal split of the original thread. 

 Sometimes (fig. 8) the two arms of a loop are of different lengths 

 and this again seems to me fatal to any interpretation of their 

 origin by a longitudinal split. In this stage the longitudinal 

 split becomes visible again by the moving apart of the two strands 

 so that we see anew the four chromatids which form the te- 

 trad (figs. S-9). So marked a contraction as is shown in figs. 7 

 and 8 is not universal. Sometimes the four strands are visible 

 through the whole process. Fig. 10 represents a stage but little 

 later than fig. 6, where the spirem has not yet broken as in that 

 case though the granules are more closely packed together to 

 form the four elements in each of the tertads ; but the approxi- 

 mation of the two sides of the loop is still incomplete in the upper 

 tetrad. The chromosomes now become more compact and grad- 

 ually take their position in the tetrads (fig. 11.) All traces of 

 the linin thread may have disappeared by this time or the orig- 

 inal linin may persist between the two tetrads as in fig. 11 

 and Ijy the attachment of its split ends show plainly which chro- 

 mosomes are the result of conjugation and which of splitting. 

 From a stage represented by fig. 11 it is an' easy step to the ma- 

 ture tetrad ready for the first division (fig. 12). The only change 

 consists in a further shortening of the chromatids. 



In the process just described the ends of the loops which form 

 the tetrads are connected by two double linin threads which 

 twist or pass close together at a common point, corresponding to 

 the bases of the two original loops. Because of their being thus 

 drawn together the resultant tetrads nearlv always stand at an 

 angle to each other instead of extending in the same straight line, 

 see especially figs. 6 and 7. This angle persists until just before 

 the separation of the dyads in the first mitosis (fig. 12) and is 

 very noticeal)le. While it cannot be regarded as positive evi- 

 dence either way, it is not easy to explain such an angle on any 

 assumption of double longitudinal splitting but it corresponds 

 with and helps to corroborate the looping shown to take place 

 in tetrad formation. 



After the tetrads are well formed the facts of the process of 

 reduction are so well known as to require no amplification here. 

 From each tetrad by the two maturation divisions are passed out 

 successively a dyad and a monad, leaving one monad from each 

 of the tetrads to form the resulting female producleus. Since the 



