28o 
R. A. HARPER 
Starch formation in the plant has been studied chemically much more 
than cytologically, and our various theories of light assimilation have been 
based very largely on the chemical study of chlorophyl and obse|vations 
on the appearance and disappearance of starch grains in their relation to 
the income and outgo of CO2 and O2 in the cell or leaf. Timberlake's 
studies of the relation of the pyrenoids to starch formation in Hydrodictyon, 
in which he shows that each grain begins as a segment of a pyrenoid which 
undergoes a series of microscopically visible changes in becoming a starch 
grain, indicate that starch formation in these cases at least involves physical 
transformations of masses which can be studied with the microscope as 
the atomic molecular readjustments with which the chemist deals. The 
work of Lutman on Closterium and of McAllister on Anthoceros have 
added a further series of observations in the field of the cytology of meta- 
bolism which indicate that the microscope is to play an increasingly im- 
portant role in the study of these problems which have hitherto seemed open 
only to the methods of chemical analysis. It may be of interest to note 
that while the most widely current chemical theories of light assimilation 
seem to favor a katalytic theory of carbohydrate formation, all these 
cytological data point just as clearly to the acceptance of a metabolic theory. 
What is known cytologically of leucoplasts and chromoplasts agrees equally 
well with the conception of the plastid as a region of the protoplasm special- 
ized with reference to carbohydrate metabolism. The leucoplast without 
its starch grain or grains is, as noted above, apparently a very slightly dif- 
ferentiated body indeed. The loss of chlorophyl has left it with very little 
to distinguish it from the adjacent cytoplasm, and yet the layer about a 
large stratified grain of storage starch is the seat of highly characteristic 
chemical transformations which, as Denniston's studies show, involve the 
formation of a cytologically demonstrable differentiated zone between the 
plastid and the stratified portion of the grain. 
If Schimper is correct that the chromoplasts are derived from the other 
plastids by chemical transformations, we have in the frequently crystalline 
appearance of the latter the evidence that here again the plastid in the older 
tissue cells is a region in which the deposit of pigment crystals takes place. 
The carbohydrate-forming plastids have gained a greater degree of 
permanency in the cell than the elaioplasts, so far as we know them now, 
and it is of the greatest significance that, associated with the growth and 
division of the cell, they so regularly arise by division. This division is a 
simple constriction both in the case of the plastids and of the pyrenoids. 
It is clear that pyrenoids may arise de novo in the cytoplasm. Timberlake 
could not convince himself that they certainly persist through swarmspore 
formation in Hydrodictyon, and Gilbert Smith has recently shown that they 
certainly arise de novo in three of the four cells produced by the double division 
of a mother cell in Scenedesmus and certain other algae. As to the chloro- 
plasts, as noted, Schimper's evidence of their persistence through the egg 
