Two other modifications of minor importance have been added. In the enlarged 

 base of the stand-pipe there is a fioat (M), which consists of a hght hollow brass 

 cy Under which has a "play" up and down of one inch. Attached to the top of this 

 float is a slender brass wire stem (N) passing up through the narrow part of the 

 stand-pipe and projecting an inch above the top of it when the float is up, just even 

 with the top when the float is down. When the stream is running from the tank 

 and everything is in order, the float is down and the stem does not appear above the 

 top. If, however, there is any leak in the top of the tank or the operator has for- 

 gotten to screw down the cap, i.e., if the tank for any reason is not air-tight, the 

 water rises in the stand-pipe to the level in the tank, the float remains up and the 

 projecting stem warns the operator that something is wrong. The float is thus a 

 safety device and not an absolutely essential part. 



The second modification is a very smaU pipe (L) about two inches long inserted 

 on the upper side of the outlet pipe. This was added when it was discovered that 

 for each of certain intermediate disks there were two distinct rates of flow. If the 

 pipe below the union became immediately filled with water when the stream was 

 turned on, it continued to come out in a steady stream and we had the faster rate of 

 flow. Under these conditions the effective head would be (P) in the diagram. But 

 at other times, even with the same disk, air gained access to this pipe and the water 

 dribbled out. Then we had the slow rate of flow due to the head (H) . Either flow 

 remained constant but one never knew when he opened the valve which speed he 

 would have. This difference in flow was ehminated when air was admitted to the 

 outlet pipe through the small opening and pipe just mentioned. 



With this improved tank just described all chances of error due to variation are 

 eliminated except one. Variation in the rate of application might arise from varia- 

 tion in the speed at which the workman pushes the drifl; e.g., he might walk more 

 slowly when he was tired or in softer land, etc. Numerous tests of the speed of 

 different workmen at various times have convinced the writers, however, that a 

 man's speed is remarkably constant and the error from this source is small. Never- 

 theless, it would be a distinct improvement if even this small error could be elimi- 

 nated and the practical grower would be saved the trouble of measuring his speed 

 and calculating the disk which he should use. This could be done if the apparatus 

 were so constructed that the amount of solution delivered was proportionate to the 

 distance traveUed {i.e., to the number of revolutions of the drill wheel) and not de- 

 pendent on the length of time during which the valve is open. Then the rate of 

 application would be entirely independent of the speed at which the drill was pushed. 

 Such a regulating apparatus, geared to the drill wheel, is mechanically possible and 

 its construction has been planned at various times by the writers. The construc- 

 tion has been abandoned for various reasons but principally because it would be too 

 complicated and expensive and therefore less suitable for the average onion grower. 

 For all practical purposes, and even for a fairly high degree of accuracy for experi- 

 mental work, the apparatus described above is quite satisfactory. 



MATERIALS OTHER THAN FORMALDEHYDE 



One objection to the use of formaldehyde is the danger of injury to the seed 

 under certain soil conditions which have been chscussed previously. Although this 

 can be largely avoided by changing the formula of appMcation to suit the soil con- 

 dition, growers cannot always be depended on to attend to this and some losses from 

 this source are inevitable. It would be much better if we had a substitute which 

 would give just as good control as formaldehyde, but without the attendant seed 

 injury. In a search for such a substitute the writers have experimented with a list 

 of chemdcal preparations, including Kalimat, Pythal, Fm-fural, Semesan, Uspulun, 

 Germisan and several disinfecting powders manufactured by the Du Pont Company 

 and by the Corona Chemical Division of the Pittsburgh Plate Glass Company. All 

 of these were found to have some merit in decreasing the amount of smut, but for 

 the most part the control was inferior to that secured by formaldehyde. KaUmat, 

 however, is a notable exception. It was tested for two years in the field and in the 

 greenhouse in pots where the soil moisture was controlled. These experiments and 

 the results are described in detail in a separate pubhcation (3). It is sufficient here 



23 



