660 STATE BOARD OF AGRICULTURE. 



The above table emphasizes the similarity of the check and inoculated 

 plants. Starting with the same average leaf surface and same transpira- 

 tion per unit of area, even after more than ten days invasion by the 

 fungus, we find the average area and average transpiration approxi- 

 mately the same. The difference in growth was not great. 



If we examine the individual plants we are struck by the ranges in 

 each of these relations; but careful comparisons will show that no 

 greater variation occurs in the inoculated plants than occurs in the 

 check plants. Moreover, both sets, before the inoculation took place, 

 showed a similar variability. Our experiments cannot hope to avoid 

 errors which come under this head. The number of plants possible for 

 manipulation is so few in comparison to the number needed to eliminate 

 error from this source, that it can hardly be hoped that experiments of 

 this type can ever be entirely free from this factor. The following points 

 may be made: 



The method of obtaining the area eliminated a large source of error 

 commonly present in area determination. The method of obtaining the 

 amount transpired per plant was accurate to at least I/2 gram, if not 

 1-10 gram. Moreover, the amount used is an average of several readings. 

 The amount transpired per unit of area in the unit of time is very sim- 

 ilar in each average. 



Considering that in the inoculated plants, % of each alternate leaflet 

 was diseased, we have the striking conditions of almost identical per- 

 formance. 



Necessarily the interpretation of this condition affects the whole ex- 

 periment. We can have the following possibilities: 



(1) Healthy and dead areas transpiring alike. 



(2) Dead areas allowing more evaporation. 



(3) Dead areas hindering water loss. 



Were the first case true, the similarity of ratios is very readily 

 accounted for, but then the death of the leaf, and the reduction of vital- 

 ity must be explained on grounds other than mere water loss. 



In event of the second, the yellowing of the leaves, the leaf-cast, etc., 

 could readily be attributed to excessive water loss. Since, however, the 

 ratios of transpiration for the leaves are almost the same, we need to 

 postulate a repression of transpiration from the living areas. 



In case the third condition were true, we need to postul'atie an aug- 

 mented evaporation from the unaffected portion of the leaves, and at- 

 tribute this to a chemical stimulation perhaps, from the by-products of 

 the fungus. 



To decide which of these conditions really existed, the following ex- 

 periment was performed, with leaflets in early stages of the disease, and 

 with leaflets which had been diseased for some time. Leaflets showing 

 a few well defined spots were attached by the petiole to small rubber 

 tubes. Other parts of the leaf were cut away and the cut surfaces cov- 

 ered with vaseline. All manipulations of attaching leaf to the rubber 

 tube were carried on under water. The rubber tube was then attached 

 to a potometer set up as shown in diagram. (Figure 2). In order to 

 have conditions of humidity and temperature constant, the work was 

 performed in an enclosed chamber (approximately 2 feet in each dimen- 



