Pollen Development in Lactuca. 38 5 
plasmodium with active nuclei, ( 2 ) in which the tapetal cells are emptied 
of their contents without the breaking down of the cell-walls. Between 
these extremes every type of intergrade occurs. A plasmodium was found 
in Anthurium , Lav at era, Cobaea , Lonicera , Valeriana , and Knautia . In 
Galium the tapetal cells form pseudopodium-like incursions between 
pollen mother-cells. The other type, in which the cell-walls remain intact 
and the cells are finally resorbed, occurs in Hyacinthus , Galtonia , Iris , Tilia , 
litmus , Gaura , Sambucus , Cucnrbita,* %lc. In Arabis and Linum the 
condition described closely approaches that found in Lactuca . The tapetal 
cell-walls disappear very late, producing in the former a plasmodium with 
disorganized nuclei and in the latter a plasmodium-like disorganized mass. 
Tischler (1915) finds that in the Commelinaceae the walls of the tapetal 
cells are lost during synapsis. The plasmodium then wanders in among 
the pollen mother-cells, its nuclei undergoing changes in form and structure. 
It is finally resorbed, so that only a trace remains in the ripe anthers. In 
Commelina coelestis Tischler figures the plasmodium in direct and intimate 
connexion with the tips of the thickenings on the sculptured walls of the 
pollen grains. 
Pickett (1916) has described in Arisaema how the tapetal cell-walls 
disappear, allowing the periplasm to spread through the sporangial cavity. 
He says : ‘ The tapetal nuclei for a considerable period show peculiarities 
of structure, and take an amoeboid form suggestive of active migration 
among the developing pollen spores.’ While the type of tapetal history 
is generally the same throughout a large group, there are exceptions in 
which the tapetum of related families may show quite different behaviour. 
Discussion. 
The Method of Synapsis . 
From the preceding account it is clear that the method of chromosome 
reduction in Lactuca is telosynaptic, and essentially in accord with the 
scheme of Farmer and Moore (1905). There is no indication of a pairing 
of threads previous to or after synizesis, but the delicate univalent' leptonema 
of synizesis gradually condenses into a relatively short and thick pachynema, 
which arranges itself into as many loops as the gametic number of chromo- 
somes. Clearly the pachynema is a single thick filament composed of the 
somatic chromosomes arranged tandem, the homologous paternal and 
maternal chromosomes alternating. The two arms of each loop of the 
pachynema constitute a pair of homologous chromosomes lying side, by 
side, and owing to torsion these arms or sides of the loop frequently become 
wrapped around each other. The looped condition apparently corresponds 
in time with the second contraction phase of other forms. It differs from the 
bouquet stage (which apparently occurs in the presence of a centrosome) in 
