1901.] NATURAL SCIENCES OF nilLADELrillA. 209 



Hemiptera (described iu my previous paper, l. e. ), I have counted 

 iu the testes of the Woods IIoll individuals the chromatin ele- 

 ments in all the cases of spermatogonic monasters which were 

 favorable for such counting, with the following results: Nine 

 spermatogonia showed exactly thirteen elements; in one case I 

 could not determine whether thirteen or fourteen were present. 

 These cases from four different testes, as well as those from four 

 testes of Pennsylvania individuals previously described by me, are 

 sufficient to show that the uneven number is not an individual 

 variation, due e.g. to some pathological condition, but is probably 

 characteristic of every individual of the species. 



The uneven normal number of chromosomes which I have 

 demonstrated also for Protenor belfragei, Ahjdus eurinus and 

 (Edancala clorsalis, represents a stage in the change of the number 

 of chromosomes from one even number to the next successive even 

 number. For Protenor I have shown that the uneven sperma- 

 togonic number is produced by a failure of two of the sperma- 

 togonic chromosomes to separate from one another. This I can now 

 prove for Harmostes also. For Avhile in most of the monaster 

 stages, as in fig. 22, all eleven chromosomes appear more or less 

 simply rod-shaped, in a few cases, as in fig. 23, one of the eleven 

 shows a well-marked transverse constriction. Were this constric- 

 tion a complele division, there would be the even number twelve. 

 Hence, for Harmostes the ancestral number of chromosomes 

 must have been twelve, and if, as is the case in Protenor, the odd 

 bivalent chromosome is "destined to change from a chromosome 

 into a chromatin nucleolus, in the course of time ten chromosomes 

 will be the number for the species. 



In conclusion, I would again call attention to the importance of 

 studying the chromosomal relations comparatively in a large num- 

 ber of species of a group. By such investigations not only may 

 much of importance be obtained regarding the evolution of cell 

 structures themselves, but by implication a criterion may thereby 

 be obtained for testing genetic relationships. In opening up this 

 line of research, I have drawn attention so far mainly to the 

 numerical relations of the chromosomes, and to the chromatin 

 nucleoli as representing chromosomes on the way to disappearance 

 during progressive evolution. These are the facts most easily 



