Invertebrate Cave Fauna 
105 
reacted to dead amphipods and isopods, but when a live amphipod or 
isopod was put in the water, the larva would raise itself on its front legs 
and usually its hind legs as well. It then remained motionless until the 
prey came within 2 to 4 cm of its snout. Then, with rapid sucking 
action, the salamander ate the prey item. Larvae will also attack the 
ends of small brushes moved slowly through the water, indicating that 
mechano-reception is the primary method of prey detection. 
The behavior of the larvae toward prey is remarkably uncompli- 
cated. The functional response is linear over a wide range of prey densi- 
ties (Fig. 36), a wider range than is normally encountered in the field. 
This linearity is apparently due to the very short handling time and the 
larvae’s prodigious appetites. One larva ate eight Caecidotea recurvata 
in 30 minutes in the laboratory. 
Although G. prophyriticus attempts to capture any Caecidotea 
recurvata or Crangonyx antennatus that comes close to its snout, it is 
about three times as successful at capturing C. recurvata, apparently 
because C. antennatus often avoids predation by swimming out of 
range. Actual predation rates also depend on the fraction of the popula- 
tion accessible to predators. Because nearly all G. porphyriticus larvae 
are in pools rather than riffles, and relatively still water aids prey detec- 
tion, individuals in riffles and on flowstone are inaccessible to preda- 
tors. In McClure Cave in Lee County, it was found that the actual 
predation rates (proportional to success rate times the proportion 
accessible to predation) did not significantly differ between the two 
species (Culver 1975). This is probably coincidental, but it does facilitate 
modeling the system. Using standard competition and predation 
equations, which is justified in part by the linear functional response, it 
was predicted that both the density and frequency of C. antennatus 
should increase. Even though it is preyed upon, the reduction in density 
of its competitor C. recurvata more than compensates for predation 
losses. Field data confirmed these predictions. Frequency of C. antennatus 
increased from 0.09 away from larvae, to 0.44 in the immedidate vicinity 
of larvae. 
Salamander predation also resulted in habitat shifts of the prey. 
With salamanders nearby, a greater frequency of both prey species was 
found in riffle and flowstone “refuges.” Consequently, when salaman- 
ders first invade a cave, a greater frequency of the prey population is 
available. Potential predation rates are nearly double those at equili- 
brium. Thus, invasion should be easy, but the establishment of a repro- 
ducing population should be difficult. This prediction is also confirmed 
by the data. Most “populations” are fewer than five individuals. 
Competition 
Competition is the most extensively studied and probably the most 
important interaction in caves. Barr (1967b) and McKinney (1975) have 
