GENETIC STRUCTURE OF POPULATIONS. IV. SUMMER ENVIRONMENTAL VARIABLES AND LETHAL AND SEMILETHAL FREQUENCIES IN A NATURAL POPULATION OF DROSOPHILA MELANOGASTER.

Abstract

Band and Ives (1961) have presented evidence of a significant relationship between the immediate environment and lethal and semilethal frequencies observed among second chromosomes in the natural population of D. melanogaster of South Amherst. This relationship is manifest in the negative correlation between temperature range prior to sample collections and the lethal + semilethal (le + sle) frequency recovered in the samples. The significant association has been found to persist when data for collections made in 1960 (Band and Ives, 1963) and 1961, 1962 (Band, 1964) are included with the previous observations spanning 1945 to 1959. Other investigators (Dobzhansky, 1952, 1956; Strickberger and Wills, 1966) have reported correlated changes between specific inversion frequencies and rainfall in populations of other Drosophila species. The present study was undertaken to determine if correlated changes between le + sle frequencies and rainfall also existed for the population of D. melanogaster of South Amherst. Rainfall is defined as summer rainfall, the total amount of rain during June, July and August prior to population sampling, usually in September. When samples were collected twice during the Fall, only data from the first Fall collection have been used. The possible relationship between average summer temperature and le + sle frequency recovered in the first Fall collection has also been investigated. It is also shown that temperature range prior to collection and le + sle frequency remain significantly correlated when only first Fall collections are included in the analysis. MATERIALS AND METHODS The weather data.-The data from which average summer temperature and total summer rainfall have been computed come from the weather records of the University of Massachusetts Weather Station at Amherst, Mass. When the instrument shelter site for this station was moved in December, 1960, temperatures recorded were lower than at the old site. For December, 1960, and thereafter, monthly means have been adjusted by one of us (PTI). The total temperature data provide evidence of a long-term gradual warming trend, noted especially in the increase in average summer and spring temperature. As shown from decade means in Table 1, there is an increase of over 10 F in the 70-year period from 1889 to 1955. In agreement with Dorf (1960) the latest short term warming cycle within the overall warming trend appears to end in 1955; this is especially marked in Amherst when the years after the sixth decade are subdivided into 1949-1955 and 1956-1965. For summer, July means were 70 F or higher from 1933 to 1955, 23 years in a row. It was under 70 F in 17 of the 44 years between 1889 and 1932 and in 4 years of the decade since 1955. The three hottest Julys on record were, in order, 1955, 1949 and 1952 and these were also the hottest summers from 1889 to 1966. Eleven summers had mean temperatures of 70 F or above in the period from 1933 to 1955, only two in the period from 1889 to 1933. Trend changes are less regular for the winter season. Nevertheless, the partdecade mean for the winter quarter in 1949-1955 is over 3.5 F higher than in the