No. 472] 



POLLEN GRAIN VARIATION 



273 



views of the same pollen grain, and Figs. 13 and 14 (PI. 1) are the 

 corresponding views of another grain. A study of the structure 

 shown in Figs. 12 and 14 as well as others more or less similar to 

 them finally suggested what the writer believes is the correct inter- 

 pretation of the extra cells along the dorsal wall of pollen grains 

 as shown in Figs. 10, 11, and 13. In Fig. 12 it will be noticed at 

 once that the cells labeled a and h are like the stalk and central 

 cell in a typical pollen grain, while cell c has the same relation to 

 the others as a tube cell, and cell d has no corresponding part in 

 the typical pollen grain. No cells appear to correspond to the 

 prothallial cells in the typical grain. In Fig. 14 the correspond- 

 ence to the typical pollen grain is less clear than in Fig. 12. Never- 

 theless there is a distinct suggestion of a similar structure. In 

 short, a study of the dorsal view of as many of these grains as possi- 

 ble led to the conclusion that grains of this type are not so very 

 different from the type shown in Fig. 8. They are double pollen 

 grains, but the division wall, instead of running from dorsal to 

 ventral side and dividing the grain into two equal parts, divides 

 it into two very unequal parts, and cuts off the .smaller portion alone/ 

 the dorsal wall. This interpretation was made more rcrtain on 

 looking over all the rough drawings made <hn-ing the study, when 

 it was found that there was a very complete series of transitional 

 stages between the structure sliown in Fig. <S and that in Fig. 12. 

 In these transitional stages the (Hvisioii wall ran from some point 

 on the dorsal side obli(|uely toward one of the wings. Fig. 9 is an 

 example of one of these transitional stag€\s, and Fig. 15 is a diagram 

 of what was seen in a dorsal view of a pollen grain with an oblique 

 division wall. The dotted line .r/y is presinnably the line of jtnietion 

 of the oblique division wall with the dorsal wall of the grain. This 

 was of course not plain, owing to the oblique position of the reced- 

 ing wall. Cell a bulges into cells b and c like the cell which divides 

 to form the stalk and central cell. Cells h and c taken together 

 represent the tube cell. If, as is generally supposed, the tube cell 

 represents an antheridiuin wall, then we liav<' hwe a ca^e where 

 the antheridium wall is two-celled, a very uinisual thing among 

 the gymnosperms. One other case was observed almost exactly 

 like this one, except for a slightly different position of the wall 

 separating cells b and c. It is just possible that the two tube 



