ever, was dry and there was no rain for at least two days after the seed was sowed on 

 April 13. The weather continued cold and the seed came up very poorly. When it 

 did come up there was a striking difference between the treated and untreated rows. 

 The checks were much thicker and appeared so throughout the season. There was 

 a very obvious injury from the formaldehyde, irrespective of the formula used. The 

 writers were informed by onion growers who were in a position to know the facts of 

 the case, that growers all through the section had the same trouble with formalde- 

 hyde during 1921 and that many fields of onions were plowed under because of the 

 reduced stand. 



No jdeld data were taken in 1921 because the onions were harvested by mistake 

 in the absence of the writers and no records were kept. 



After the experience of 1921 it was decided that further progress in the search 

 for the best formula could not be made until the factors which cause formaldehyde 

 to injure were determined. The investigation of this problem was not finished 

 until 1922, but since it has an improtant bearing on the subsequent experiments, 

 we shall consider the findings at this time. 



Formaldehyde Injury 



At first it was thought that the injury of 1921 was due to some difference in the 

 composition of the chemical. Samples from all available sources were therefore 

 collected, including among them samples of the same material which we had used 

 during the previous years and wlfich had not caused injury. Chemical analyses of 

 the samples by Dr. Holland of the Chemistry department, revealed no differences in 

 composition which were of sufficient size or character to warrant a suspicion of their 

 toxic effect. The manufacturers assured us that there had been no change in the 

 method of manufacture of formaldehyde. The most apparent difference was in the 

 percentage of methyl alcohol contained in the various samples. The sixteen sam- 

 ples were next tested simultaneously on onion seed sowed in greenhouse benches, 

 using for each the formulas 1-50-5000, 1-50-3000, 1-100-3000 and leaving un- 

 treated rows between the plots. In order to see what effect variation in the per- 

 centage of methyl alcohol would have, samples containing known percentages from 

 0.4 per cent, to 16 per cent, alcohol were included in the tests. Without discussing 

 the results in detail, they may be briefly stated as follows: — 



1. The source of the formaldehyde was without influence on the percentage of 

 plants which came up. 



2. Variation in the percentage of methyl alcohol made no difference. 



3. With aU samples and every formula used, formaldehyde retarded germina- 

 tion for a day or two and 



4. With all samples and formulas, formaldehyde prevented the germination of a 

 certain percentage of the seed, this loss being more apparent when a more concen- 

 trated solution was used or where the quantity of the solvition was increased. 



The results of these ex-periments eliminated the possibility that the losses of 

 1921 were referable to difference in the character of the formaldehyde. 



Since there had been no differences in the formulas or method of apphcation of 

 formaldehyde nor in the treatment of the soil in 1921, there appeared to remain only 

 the character of the season itself as a possible explanation of the trouble. The 

 character of the spring weather could obviously affect the soil — where the injury 

 occurs — in two ways: (1) in the percentage of moisture and (2) in the temperature. 

 Experiments were undertaken with the object of determining to what extent the 

 percentage of injury is influenced by variation in the percentage of moisture in the 

 soil. The effect of the soil temperature on injury has not been determined as yet. 



These experiments have been described in detail in a previous pubhcation (2). 

 For our present purpose it wiU be sufficient to quote from the summary of that 

 paper : 



"When a concentrated formula like 1-50-3000 is used, the amount of injury 

 depends on the moisture condition of the soil. Injury varies inversely as the per- 

 centage of moisture. 



The amount of injury may be reduced by diminishing the amount of the solution 

 apphed per unit of row, iDut the percentage of smut control is probably also reduced. 



The amount of injury in a dry soil may be reduced hy increasing the dilution of 



17 



