1967] 
Pinter — Widow Spiders 
297 
Fig. 25. Measurements. Ratios of Latrodectus Nos. 1, 2, 3, and 4. Modi- 
fied Dice-Leraas diagrams of the ratios of Femur I/Sternum width (A), 
Metatarsus IV/Sternum width (B), and Femur I/Patella-tibia I. The 
horizontal line denotes the mean, the rectangle indicates the 95 per cent 
confidence intervals for the mean, and the vertical line represents the 
observed range. N equals 30 for Latrodectus Nos. 1, 2 and 3 ; N equals 4 
for Latrodectus No. 4. 
sus IV/Sternum width, and Femur I/Patella-tibia I) and the 
individual ratios for the four specimens from population No. 4 are 
illustrated in Figure 25. 
The t-test was applied to the differences between the means of 
each of the ratios for each of the three populations. With a com- 
parison of any of the three ratios, one finds a statistically significant 
difference (.95 confidence) between populations No. 1 and No. 2, 
No. 1 and No. 3, and No. 2 and No. 3. 
CONCLUSION 
By failing to consider geographic variation, Gerschman de Pikelin 
and Schiapelli confused a number of species. Their conclusion may 
be rejected on the grounds that there is reasonable evidence to con- 
clude that at least three, and probably four, species besides L. geo- 
rnetricus are present in Santiago del Estero, Argentina. Because this 
study was limited to sympatric species, it was possible to show that 
a morphological discontinuity exists between the populations, hence 
reproductive isolation. These results support the findings of Abalos 
and Baez. 
