6 COMPARATIVE EMBRYOLOGY OF THE RUBIACEAE 
‚sac-mother-cell. Against the adoption of such an interpretation it 
may be urged that, where each of four spores form the primordium 
potentially or actually of an embryo-sac, the gametophyte becomes 
spilt up into four individuals. To be sure there are analogies to 
come to our aid, such as the constant development of four embryos 
from a single proembryo or the much more widely distributed 
phenomena of vegetative reproduction. 
But in spite of these analogies, it would seem more natural to 
regard the gametophyte as an individual by coalescence, having 
its origin in four like vegetative cells whose primitive function has 
been lost. These four cells become grafted on each other, so to 
speak. The chief difficulty seems to be due to our formal concep- 
tional idea of an individual, which in nature, outside of our con- 
sciousness, does not exist, : 
In comparison with the cases at present before us we may re- 
call the very important results obtained by Juel * and by Murbeck.t 
Juel found that in Antennaria divica a tetrad division takes 
place, and his account of the process differs in no essential detail 
from mine. [m Antennaria alpina, however, in which partheno- 
genesis is now known to obtain, the tetrad division does not occur, 
and the embryo-sac-mother-cell develops directly into the embryo- 
sac without any reduction in the number of the chromosomes or 
the usual chromosome-ring formation. Nor is this process to be 
compared with that in some plants in which megaspores are not 
delimited in the usual manner, but in which parthenogenesis does 
not occur (Ziium). In such plants the first two divisions are the 
same in character as the two divisions occurring when four mega- 
spores are formed. 
In Alchemilla, as shown by Murbeck, parthenogenesis also oc- 
curs, and in such event the megaspore mother-cell divides some- 
what irregularly, forming at most three cells, because of the failure 
of one of the daughter-cells to divide. These divisions are not 
different in character from ordinary divisions ; or, in other words, 
the heterotypic and homotypic divisions fail, and the chromosome 
number remains constant. 
* Vergleichende Untersuchungen über typische u und و سس‎ he Fortpflan- 
zung bei der Gattung Antennaria. K. Sv. Vet. Akad. Handl. 33: No. 5. 1900. 
1 Parthenogenetische Embryobildung in der Gattung pente dd Acta Reg. Soc. 
Physiogr. Lund. 11: No. 7. 1901. 
