490 AMERICAN JOURNAL OF BOTANY [Vol. 9, 
particles agreed favorably with the size of the crushed starch particles 
when examined under the microscope. Figure 9 shows the resulting photo- 
FiG. 9. Finely pulverized cane sugar. 
graph. The lines are fully as strong and even clearer in this, and the 
exposure was fully 20 percent less than in the previous cane-sugar picture 
(fig. 5). The first and obvious conclusion concerning the starch is that 
the lines produced by the whole grains are due to the structure of the 
grain. The second conclusion is that the structure is neither crystalline 
nor amorphous. 
Before speculating further on a possible structure of the whole grain, 
it is interesting and suggestive to compare with this work on starch some 
recent work done on wood and cotton cellulose fibers. In 1920, Herzog and 
Jancke (38) published a table of figures which they obtained by exposing 
wood and cotton cellulose to X-rays, in a similar manner to that which 
has just been described for starch. They published only the values of 
(sin d)l2. These have been converted into ratios, and are given in table 8 
along with those for starch. 
Table 8. Comparison of Ratios of Atomic S pacings for Cellulose and Starch 
Cellulose 
Computed 
Starch 
Values of 
Cotton 
Ramie 
Wood 
Cube 
Potato 
Cassava 
Corn 
1. 00 
1. 00 
1. 00 
1. 00 
1. 00 
1. 00 
1. 00 
.670 
.677 
.682 
.707 
•745 
•758 
.765 
.557 
•558 
•554 
•577 
•575 
•587 
.592 
.486 
•499 
.496 
.500 
•507 
•530 
•524 
•443 
.448 
•443 
•447 
•452 
■465 
.456 
.366 
•367 
•372 
.408 
.404 
.416 
•339 
•346 
•343 
•354 
•387 
One of the very few well-established facts about starch is that it is 
formed only when in contact with a protoplasmic body. Likewise, cell 
walls are formed only when in contact with protoplasm. Cellulose and 
starch have long been recognized as being chemically very closely related. 
Here in table 8 appears again evidence of this close relationship ; also there 
appears a very consistent difference, all the values for starch being above 
the corresponding ones for a cubical arrangement, while all those for cellu- 
lose are below. This would indicate a slight difference in the atomic 
