graph. Typically, females in the family Eriococcidae have two instars in addition to the
adult (Stehr 1991). There are a few possible explanations for the potential presence of a
fourth cluster. There may be a wide range of size variation in the size of the adult of T
ovatus. In this case, the weak fourth cluster could be the result of outlying data from
measurements of extremely large adults. Outlying data are more difficult to identify
when comparing data sets with several dimensions such as this (15 observations per
specimen). The presence of a fourth cluster could also be due to variation in the initial
measurements. The measurements taken from the Auto-montage image were in two
dimensions; slight errors in the measurements could have occurred if the leg segment of
the slide mounted specimen was not perfectly horizontal. Both of these problems could
be solved by taking a larger sample size (n > 62). Another explanation is that there is a
supernumerary instar; this could also be elucidated by analyzing a greater sample size.
The occurrence of supernumerary molts in laboratory reared colonies is not an
uncommon observation (Chapman 1998).
Correlation of Gall Size to Nymphal and Adult Stages of Tectococcus ovatus
By comparing r2 values for multiple regressions of gall size vs. insect
measurements, the length of the prothoracic trochanter/femur segment was determined to
have the closest relationship with the gall width (Table 3-1). Figure 3-5 illustrates the
close correlation between these two variables; the best fit equation is y = 44.603x +
16.233 (r2 = 0.7126; df = 1, 65; p < 0.001).