130 Sargant . — The Reconstruction of a Race of 
tilization contains eight nuclei, without any trace of cell-formation and with- 
out the usual very characteristic orientation. The two which later act as 
synergids are sometimes distinguished by their smaller size, and the lower 
polar nucleus is from the first conspicuously different from the rest. The 
remaining five are precisely similar to each other : no one of them can be 
picked out as the future egg-nucleus. Tulipa Gesneriana , however, has a 
perfectly normal embryo-sac. 
This fact alone would indicate that the peculiarities of the species de- 
scribed can hardly be a survival of primitive characters. Indeed, the devia- 
tion from the usual structure is not very great, and may be more apparent 
than real. The irregular orientation clearly depends largely on the absence 
of a central vacuole. It does not follow that no difference exists between the 
five nuclei, which later behave quite normally as egg-nucleus, upper polar 
nucleus, and antipodal nuclei, because we cannot perceive such difference. 
To trace the formation of these nuclei in greater detail than Guignard has 
done would be interesting. In Tulipa Gesneriana Ernst has shown that 
each nucleus in the chalazal tetrad is built up of more chromosomes than go 
to each of the micropylar tetrad. If this should prove to be the case in 
Tulipa sylvestris and T. Clusiana also, three of the five similar nuclei must 
be marked out as antipodals from their first formation. The egg-nucleus 
must be one of the two formed with the reduced number of chromosomes. 
A similar but less marked absence of polarity is shown in certain excep- 
tional embryo-sacs of Trillium grandiflorum according to Ernst (23), and 
of Juglaits nigra according to Karsten (50 a). But in the large majority of in- 
dividuals examined both observers found the embryo-sacs nearly if not quite 
normal. These genera should certainly not be included in such a list as this. 
The species of Cypripedium described by Miss Pace (63) are C. specta- 
bile and C. parviflcrum , with corroborative sections from two other spe- 
cies. The embryo-sac mother-cell divides once : the upper segment soon 
perishes, the lower becomes the embryo-sac. After division of its primary 
nucleus, both daughter-nuclei divide again : the mature embryo-sac con- 
tains four nuclei. At the time of fertilization three nuclei are surrounded 
by delicate cell-walls, and orientated like an egg-apparatus : the fourth lies 
near the chalaza (63, Figs. 29, 30). 
In entering the embryo-sac the pollen-tube seems to push the nucleus 
of one synergid-like cell in front of it. This nucleus passes to the chalazal 
end and fuses with the fourth nucleus and with one of the generative nuclei 
from the pollen-tube. The other generative nucleus fertilizes the ovum in 
the usual way. 
This is an important variant on the normal embryo-sac structure, but 
clearly a reduction from it. The exact homologies of the nuclei are obscure. 
If we consider the embryo-sac of Cypripedium to represent the upper half 
of the Lilium sac, its four nuclei are truly homologous with the upper tetrad 
