298 Drs. D. A. Welsh and H. G. Chapman. On the [Apr. 30, 



precipitum obtainable in the latter instances. But it can be shown that the 

 test proteid is not all removed from the solution even when minute quantities 

 interact with antiserum to yield definite precipitates. On the contrary, when 

 the clear superfluid is transferred to a fresh tube and a second dose of anti- 

 serum is added, there is deposited a second precipitate, and this process has 

 been repeated six times, each fresh addition of antiserum resulting in a fresh 

 precipitate. In these circumstances it is hard to conceive an adequate source 

 of " precipitable " substance other than the antiserum. Supported by other 

 facts also, this conception of a precipitable content in precipitin antisera is 

 not only in harmony with the known phenomena of precipitin reactions, but 

 appears to offer a more consistent explanation of these phenomena and of 

 certain anomalies that arise on the assumption that the precipitable substance 

 is derived wholly or mainly from the test proteid. 



Our experiments are based on the exact measurement of the substances 

 concerned. The interacting quantities were often very minute, and we have 

 been aided in their measurement by the use of dried material. When it was 

 more convenient to employ undried material, a known volume was dried and 

 weighed, so that our results might be expressed in terms of the weights 

 of dried equivalents. We have independently determined the fact, recognised 

 by the majority of observers, that careful desiccation of the substances 

 concerned does not impair their efficiency. 



At a room temperature averaging 18° C. and in the dilutions employed by 

 us, we have repeatedly noted that a time interval of less than 48 hours may 

 be insufficient for complete interaction, though in certain circumstances the 

 reaction may be regarded as complete in 24 hours. 



For such prolonged exposures, and more especially when the same solutions 

 were tested again and again by repeated addition of proteid or of antiserum, 

 it became imperative to exclude bacteria as effectively as possible. This we 

 were able to do by collecting and storing our material aseptically, by sterilising 

 all apparatus and saline solutions employed, by filtration through porcelain of 

 suspected solutions, and generally by taking the usual precautions against 

 bacterial contamination such as plugging the tubes with sterile cotton wool. 

 The only stage at which contamination was likely to occur was that of 

 shaking the tubes to ensure thorough mixing of the solutions. In the 

 relatively few instances in which bacterial growths appeared we had no 

 difficulty in distinguishing the sediment due to bacteria from the flocculent 

 matter thrown out as a result of precipitin interactions. Nevertheless, when 

 small precipitates were in question, we rejected contaminated tubes. 



In reading the precipitates we have made a careful approximation to the 

 length of the tube occupied by the deposit, a graduated scale being held 



