461 



there is a decided tendency for exceptionally low temperatures to occur between, say, a.m. and 2 a.m., 

 then we may reasonably expect a lower mean value for the six lowest readings during this interval than 

 for the corresponding quantities for the other eleven 2-hour intervals. Table XI. shows the results thus 

 obtained for the twelve months, the year, and the usual seasons ; the data are means from two years. 



TABLE X. Extreme Monthly Temperatures and Monthly Ranges. 



* The values entered under January, 1902, belong really to January, 1904. 

 TABLE XL Means from Six Lowest Minima at each 2-hour Interval.! 



During the six months October to March, the data in Table XL show a very distinct diurnal inequality, 

 which resembles on the whole the inequality of the ordinary temperatures in these months. The lowest 

 minima are met with on the whole at or near the hours when the regular diurnal inequality of temperature 

 has its minima. 



For midsummer, Table XL shows a nearly steady rise from a minimum at 2 to 4 a.m. to a maximum at 



t All the temperatures in any one line have the signs shown in the first and last columns. 



