REPORT OF THE SECRETARY. 19 



connexion of changes of the weather in the arctic regions with those 

 in the more southern latitudes of tiie continent of North America. 



The observations for temperature were made at equal intervals day 

 and night; in winter they were generally taken every two hours, in 

 the summer every four hours. The register extends over twenty- 

 seven months, and comprises a total number of 7,113 observations. 



The highest monthly mean temperature occurred in July. In the first 

 year, 1857, in latitude 62°, it was 45°. 53. In the second year, (1858,) 

 in latitude 74°.4, it was 36°. 60. In 1859, latitude 72° it was 40°'.12. 

 The lowest monthly mean temperatures Avere in January. That in 

 1858, in latitude 73°. 2, was — 24°.87. The second, that in 1859, in 

 latitude 72°, Avas — 33°.57. 



The mean annual temperature of the year, as deduced from all the 

 observations at Port Kennedy, in latitude 72°, Avas 1°.85. The 

 highest temperature observed at the same place Avas 55°, Avhich oc- 

 curred July 29, 1859, and the loAvest — 49°. 8, January 21 and Feb- 

 ruary 15 of the same year. 



The extreme range, therefore, Avas 104°. 8. 



The highest temperature observed by Kane at Van Rensselaer 

 harbor, in latitude 78°. 37, Avas 51°, and the minimum — 66°.4, givino- 

 a range of 117°.4 



The greatest daily range of temperature Avas in June, amounting 

 to 9°. 60; the least in December, Avhen it was 0°.84. At Van Rens- 

 selaer harbor, the greatest range occurred in April, and Avas 9°. 09 ; and 

 the least in November, amounting to 1°, shoAving a correspondence 

 in amount of variation, but not in time. 



On an average, the maximum temperature is reached between noon 

 and 1 p. m., and the minimum between 2 and 3 a. m. ; whereas at 

 Van Rensselaer harbor, as observed by Dr. Kane, these hours Avere 

 respectively 2 p. m. and 1 a. m. 



One of the most interesting points of the discussion of the obser- 

 vations is that of the connexion of the phases of the moon Avith the 

 loAv Avinter temperatures. This connexion Avas early suggested by 

 arctic explorers, and Avas independently deduced from tlie observa- 

 tions by Kane. In order to investigate the question of this connexion 

 the mean daily temperatures Avere divided into periods of five days, 

 thus forming a table in Avhich the dates of the occurrence of full and 

 new moon are given with the corresponding mean temperatures. By 

 taking the differences of the mean temperatures it is seen that the 

 mean temperature is lower at full moon than at new moon by about 



