CLIMATE AND THE DISTRIBUTION OF MARINE ANIMALS 



297 



from showing roughly the same periochc fluctua- 

 tions, the July temperatures have remained very 

 nearly tlie same over the entire period. 



The imperfections of the New Haven tempeia- 

 ture record have been pointed out by Mitchell 

 (1953). He considers tiie most serious of these 

 to be tlie effect of urban development which may 

 have contributed as much as 1° F. to the average 

 temperature during tlie winter montlis. He cites 

 the temperature records at the Bluehill Meteoro- 

 logical Observatory of Harvard University, lo- 

 cated 10 miles south of Boston, as being quite 

 free from urban effects. Conover (1951) has pub- 

 lisiied mean winter temperatures (Decemt)er, 

 January', and February) for the Blueliill Observa- 

 tory covering the period 1849-50 to 1949-50. In 

 figure 4, we have compared the January devia- 

 tions at New Haven with those at Bluehill for tlie 

 winter months. 



The difference between the January means at 

 New Haven for the periods 1780-1900 and 1901- 

 53 is 2.94° F. The standard error of this differ- 

 ence is 0.6873. The difference is then iiighly 

 significant (/ = 4.28, 172 d. f.). If the difference 



were only 1.8°, it would still be highly significant, 

 and if only 1.35°, it would be significant. The 

 reality of the increase can hardh' be denied even 

 if one allows a maximum value for the effect of 

 urban development. With regard to tempera- 

 tures at the Bluehill Observatory over the period 

 1849-1950, Conover (1951, p. 9) states, "The 

 amount of warming up in general winter temper- 

 atuie over tiie last 100-year perioil has been 

 about 3M° Fahrenheit." 



To analyze further the pattern of temperature 

 changes, mean temperatures for each month liave 

 been averaged for the 20-year period 1780-1800 

 and by 25-year periods thereafter to 1950 for 

 New Haven, Conn., and for comparison, from 

 1876 to 1950 for Eastport, Maine (fig. 5). The 

 New Haven temperatures were adjusted to the 

 level of Eastport temperatures by subtracting 

 from them the mean difference between tempera- 

 tures at New Haven and at Eastport by months 

 for the period 1874-1923. The means for this 

 period for Eastport and New Haven are given 

 by Clayton (1927). 



Upward trends in temperature are noted fo 



22 



20 

 I 8 

 40 

 38 h 

 36 

 62 



lU 60 



I- 



58 



< 

 tt 

 UJ 

 Q. 



APRIL 



-8=- 



JULY 



24 - 

 22 - 

 20 



48r 



46 



44 



62 



60 



58 



36 



34 



1 1 r 



FEBRUARY 



MAY 



AUGUST 



30 

 28 

 26 

 56 

 54 

 52 

 56 

 54, 

 52 

 26 

 24 h 



JUNE 



SEPTEMBER 



1780 1801 1826 1S51 1876 1901 1926 1780 1601 1826 1851 1876 1901 1926 1780 1801 1826 1851 1876 1901 1926 



TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO TO 

 1800 1825 1850 1875 1900 1925 1950 1800 1825 1850 1875 1900 1925 1950 1800 1825 1850 1875 1900 1925 1950 



PERIOD 



Figure !y. — \ comparison of trends in monthly temperatures avcruKed over 21-year (1780-1800) and 25-ycar period.s; 

 New Haven, Conn., 1780-1950 (.solid line), Ea.stport, Maine, 1876-1950 (broken line). 

 •JI4171 O— .")- 2 



