by Carl Strang
Earlier in the season I described my block counts of singing insects, in which I walk around the block where I live, counting the singing insects I hear. In these counts I heard 13 species in 2009, compared to 15 in 2008. The two missing species were the dog day and scissor-grinder cicadas, both of which I heard on my block at times other than sampling sessions.
Only 3 species have been present in large enough numbers to allow comparisons between years. Most abundant is the striped ground cricket, with median counts of 13, 14.5, and 12 for 2007, 2008 and 2009, respectively*. Statistical tests found no differences among years in the striped ground cricket.
Median counts of Carolina ground crickets were 3, 3, and 7 for the three years. The change from 2008 to 2009 was statistically significant. The third species, the greater angle-winged katydid, produced median counts of 7, 1, and 0. The drop between 2007 and 2008 was statistically significant, but that from 2008 to 2009 was not.
The other species I have noted in my neighborhood, in rough descending order of abundance, are two-spotted tree cricket, snowy tree cricket, fall field cricket, narrow-winged tree cricket, Linne’s cicada, common true katydid, Allard’s ground cricket, and a few species I have heard very infrequently: jumping bush cricket, Say’s trig, Davis’s tree cricket and possibly fork-tailed bush katydid.
*For those interested in the technical details, I included only counts in August-October, the months when these species appeared consistently. Total numbers of counts were 49, 46 and 35 in the three years, respectively. I am very conservative in my use of statistical tests. Here I used the large-sample version of the Mann-Whitney U-test, a nonparametric test which produces (ironically?) a z test statistic. Non-parametric tests are needed for count data, and I seldom see anything like a normal distribution in the aspects of nature I study, anyway. Test statistic values for striped ground cricket counts were z = -0.73 for 2007 vs. 2008, and -1.58 for 2008 vs. 2009, both P > 0.01. Again, being conservative, I use the 1% rather than the 5% level of error. Carolina ground cricket counts showed no change between 2007 and 2008 (z = -0.68, P > 0.01), and an increase from 2008 to 2009 (z = -3.32, P < 0.01). Greater angle-wing counts showed a decrease from 2007 to 2008 (z = -3.33, P < 0.01), but no change from 2008 to 2009 (z = 0.32, P > 0.01).
Nature Inquiries 