454 REPORTS ON THE STATE OF SCIENCE. 



(ii) The calendar month, 

 (iii) The year. 



(iv) The quinquennial period : A group of the monthly values for 

 five years, e.g., 1901-5. 



These are directly comparable with normal meteorological tables. 



The monthly values are sufficiently close for nearly all purposes, and 

 they are also less liable to adventitious error than the weekly record, 

 where, except during an epidemic, or where the observer has an unusually 

 large population to deal with, accidental causes may have an important 

 influence on the weekly variations. From the monthly tables the com- 

 pilation of the other tables of longer period is a simple matter. 



The object of this memorandum is to inform observers of what material 

 is available and in what form, and to suggest the general lines which 

 new investigations should follow in order that they may be comparable 

 with existing records. If the suggestions in this memoi'andum are fol- 

 lowed, observers will usually find the interim results which they may 

 obtain will permit of easy rearrangement if that should prove necessary 

 for purposes of comparison. This is important, because (especially abroad 

 where reference libraries are scarce) it may happen that the opportunity 

 for observation must be seized at once or lost, while it is impossible to 

 ascertain immediately what has already been done. The observer may 

 in such a case compile his results on the lines indicated above, while he 

 is at the same time seeking information as to existing material on the 

 subject. 



The associated conditions, enumerated in Section II., may be as im- 

 portant as the actual climatic changes, if, indeed, they are not more so — 

 that is, the climatic effect is often indirect. 



Which of these shall be investigated in its relation to climatic and 

 disease variation, and how far this investigation shall be carried, are 

 matters which the individual observer must decide for himself. But 

 usually material relating to most of tliese points is already available, 

 many of which are of economic importance, and so have received early 

 attention ; this in many cases will be sufficient Every investigation 

 has a natural tendency to expansion ; this can never be entirely obviated ; 

 points crop up which need special inquiry. But a careful consideration 

 at the outset of the possible lines of expansion will often enable the 

 observer to obtain contemporaneous records which may prove useful, 

 which he could possibly never obtain later or from other sources. A 

 Avorking hypothesis is helpful, provided that it is regarded as a scaflFolding 

 to disappear in the finished work, and it is not allowed to influence the 

 selection of the observations. The difiiculty at the outset is to • dis- 

 tinguish between the essential and the adventitious ; this becomes easier 

 as the inquiry proceeds. But when the observer has no knowledge of 

 what auxiliary material is available, it is better to observe those asso- 

 ciated conditions which appear to be essential than to trust to the 

 existence of records which may be deficient on this very point. 



Of the associated conditions, probably those termed biological form 

 at present tne most promising field for investigation, especially as regards 

 the incidence of many forms of tropical disease. Much work has been 

 done in this direction, especially in regard to mosquitoes and ticks, but 

 fuller investigations are still useful, especially the relation (direct or 

 through vegetation changes) of climatic conditions to the life history of 

 the carrier and the development of the parasite carried, and hence, 



