76 Report S.A.A. Advancement of Science. 



The diagrams are capable of many interpretations, and a care- 

 ful examination of that for 1902 will show why the mealie crop is- 

 a comparative failure this year, and from a Boer point of view why 

 horse sickness in the Transvaal has been so rampant. 



Coming next to barometric pressure, we find that it is almost 

 impossible to present records graphically, as the changes, even in the 

 thunderstorm months, only amount to a few millimetres. In the 

 year 1901, for instance, the highest record we have is 642 m/m 

 and the lowest 626 m/m. The maximum is generally recorded in 

 June. Our records are taken twice daily, -i.e., 6 a.m. and 12 noon,, 

 from the barometer itself, but in addition to these direct readings,. 

 we have been in the habit of taking weekly diagrams of the baro- 

 meter pressure from the barograph. Naturally, the line is fiat, or 

 near]\ so, but one most curious fact is brought out, i.e., that the 

 highest points are always about 10 to 11 a.m. and p.m., and the 

 lowest about 4 a.m. and p.m. Perhaps someone who has made these 

 records a special study can throw some light on this point. 



The diagram Xo. III. which is apjjended shows the maximum 

 and minimum barometric pressure for the years 1 900-1 902, inclusive. 



The next series of observations for which we have records are- 

 those of temperature, and graphic diagrams in explanation are ap- 

 pended. They embrace the following : — 



1 . The average and the actual maximum temperature in^ 

 the sun. The " actual " maximum temperature naturally 

 occupies the top position. 



2. The middle curves represent the average and actual 

 maximum temperatures in the shade, and here again the actual 

 line takes top position. 



3. The lower curves indicate the actual and average mmi- 

 mum temperature, but here the " actual "" temperature takes the 

 lower position, as a moments reflection will show. 



It should be stated that all the readings are in Centigrade. 

 They are recorded in rather an unusual manner. Every point 

 plotted out on the diagram represents an actual reading, and, in order 

 to simi)lify matters, the month was s])lit up into weeks, each of six 

 davs. When there was one day over or under, it was simply in- 

 cluded in the last week. For instance, the " actual ' maximum 

 temperature plotted out was the highest for the week, but the 

 " average '' maximum was the average of the six maximum tempera- 

 tures recorded — one on each day— for the week. 



Diagrams IV.. IV. a, IV.b, IV.c. illustrate these records very 

 graphically. 



Needless to say, the general formation is much the same from 

 vear to year, but even a casual glance will shoAv large differences. 

 Taking things in their order, the greatest differences are naturally 

 shnwn bv the maximum temperature in the sun. In 1901, for 

 instance, there was a sudden drop from 48° C. to 28° C, in the 

 month of March. The lines naturally droop in the winter months, 

 but in the year 1901 the average sun temperature rose about the- 

 same right through the year. 



