ADVENTUKES IN SOITHEKX AFKICA. 13 



Oil Wediiesuay tlie 18th iilt., thcic was rain at Gettysburg, as also on 

 the next Wednesday the 25tli, when it continued lar into the night. 

 During the following week the rain occurred on Thursday late in the 

 day; during the next week on Friday early in the morning; and during 

 the week following that, on Thursday and Friday, but extending later 

 into Friday than in the week previous. 



Again; on Saturday evening the 21st of September ultimo, it rained; 

 then again on Saturday and Sunday the 28th and 29th, the rain extend- 

 ing over Sunday until near night. During the next Saturday in the 

 evening, Sunday, and early on Monday, the 5th, 6th, and 7th inst., the 

 sky was more or less overcast with clouds without any rain; on Mon- 

 day the 14th inst., the rain began at 6| o'clock a. m.; and on Monday 

 the 21st at 7 p. m. 



These examples have been given, not because they are at all peculiar 

 or the most remarkable that could have been given, for the same regu- 

 larity of succession can be perceived throughout the whole year, but 

 because they were recent and just at hand. 



5. We are therefore, if these things be facts, justified in inferring, (1) 

 that changes of weather correspondirg to each other occur at intervals 

 of about seven days and a half; the day of change in each month being 

 about two days later in the week, than in the preceding month; (2) that 

 there are generally two such corresponding changes in each week. What 

 has been said in reference to cloud, rain or snow, is tiue also in reference 

 to fair weather. Knowing then the state of the weather, especially the 

 time of cliange in one week, we may form a tolerably correct conjecture 

 with regard to that of the next two succeeding weeks. Thus if it have 

 been rainy or clear on some particular day as Wednesday of one week, 

 there is a strong probability that it will be rainy or cJear^ as the case 

 may have been, the next Wednesday, allowance being always made for 

 its occurrence from nine to twelve hours later than in the preceding week. 



An account of some other facts connected with this subject, and an 

 inquiry into the question how far the Moon may have an agency in pro- 

 ducing these alternations of weather must be deferred to another time, 



Gettysburg, October 22nd, 1814. 



STEKD.VIAN's WANDERIXGS and AnVENTUR]:S. 



" lyandcriujis and Mvcnlura in the Interior of Southern ^fru:a. 15v A.ndukw Stkedman." 2 vols. 

 I.onil'iii. 



'I'his work is the result of the travels of the author, during the years 

 J830 and 183], and abounds in interesting and instructive incitk'uts. It 



