194 



SCIENCE 



[N. S. Vol. XXVII. No. CSS 



an irre^larity in the chromosome numbers 

 of the next generation. This is especially 

 common in the hybrids. An irregularity of 

 this kind is probably the cause of the 15 

 chromosomes found by Miss Lutz' in one 0. 

 lata plant. Such irregularities will probably 

 also explain other variations of one or two in 

 the chromosome numbers, including the 20 

 instead of 21 chromosomes in what I have 

 now shown to be an 0. latay^O. gig as 

 hybrid. 



The long-believed tendency to wide vari- 

 ability in the number of chromosomes in root- 

 tips made it important that the count made by 

 Miss Lutz in 0. gigas should be confirmed. I 

 have recently examined the reduction mitoses 

 in the pollen mother cells of 0. gigas and also 

 find the number to be 28, the reduced number 

 being 14. The material for this study was 

 obtained from A. C. Life, now of the Uni- 

 versity of Southern California. It was col- 

 lected by him at Woods Hole, Mass., in the 

 summer of 1905, from plants growing side by 

 side with those from which my first material 

 was collected. 



I have also recently examined material from 

 another plant in the cross which was believed 

 to be pure 0. lata X 0. Lamarchiana, but 

 which appeared to have all the characters of 

 0. gigas and was found to have 21 chromo- 

 somes. This, with other evidence, furnishes 

 definite proof that the cross which was believed 

 to be pure had been partly pollinated with 

 0. gigas. The interesting feature about this 

 plant is the manner in which the chromosomes 

 segregate in the heterotypic mitosis. 



Almost invariably 10 chromosomes go to 

 one pole of the spindle in the heterotypic 

 mitosis in the pollen mother cell, and 11 

 chromosoifles to the other pole. It is difiicult 

 to see how this can be interpreted in harmony 

 with current views of the homology of ma- 

 ternal and paternal chromosomes. For even 

 although we assume that in the reduction of 

 chromosomes in the hybrid the descendants of 

 the 7 lata chromosomes in the cross always 

 pair with the descendants of 7 of the 14 gigas 



' International Zoological Congress, Boston, 

 August 22, 1907. 



chromosomes, the remaining 7 gigas chromo- 

 somes are invariably distributed as though 

 they paired with each other, with the excep- 

 tion, of course, of the unpaired chromosome 

 which (when one is present) may go to either 

 pole of the spindle according to chance. In 

 other words, the chromosomes are invariably 

 distributed as though 10 of them were homolo- 

 gous with the other 10, which is not the case 

 when their origin is considered. 



How the 0. lata'X^O. gigas hybrid, which 

 matures an abundance of pollen, will behave 

 in later generations, is not known. A little 

 thought ■will show that if the chromosomes 

 always segregate in germ-cell formation in 

 the manner described above, by crossing the 

 hybrid with the parental types and judiciously 

 re-crossing, in later generations plants may be 

 obtained having all sporophyte numbers of 

 chromosomes between 28 and 14, or reduced 

 numbers between 14 and 7; and whenever the 

 sporophyte number was an uneven one the 

 plant would produce two kinds of germ cells, 

 difiering by one in their chromosome number. 

 The writer hopes to carry some of these crosses 

 forward in the future, as it is believed that 

 they should throw light on the nature of a 

 chromosome in its relation to somatic char- 

 acters in heredity. 



The early history of the bodies designated 

 heterochromosomes is not yet fully worked 

 out, but they are evidently not directly con- 

 nected with the chromosomes, although fre- 

 quently almost identical with them in appear- 

 ance in the metaphase of the heterotypic 

 mitosis. It is believed that these bodies will 

 be found to be more nearly related to the 

 ordinary nucleoli than to the chromosomes. 



Resuming briefly, we may say that 0. 

 Lamarclciana, ruhrinervis, nanella, as well as 

 both the lata and Lamarchiana types arising 

 from 0. lata X 0. Lamarchiana, have 14 

 chromosomes, with perhaps occasional de- 

 partures of one from this number. 0. gigas 

 has 28 chromosomes, and certain plants ap- 

 pearing in the first hybrid generation of 0. 

 lata X 0. gigas having almost or quite the 

 identical appearance of 0. gigas have 21 

 chromosomes (20 in one plant), the latter 

 segregating equally in reduction, with the ex- 



