96 Ray E. Ashton, Jr. 



The primary habitat used by juveniles consisted of granite boulders, 0.5 

 m diameter, with underlying sand-gravel substrate. Movements into leaf 

 beds were rare (only two sightings) until early spring, when 17 percent 

 of all sightings were made in this habitat. The beds that salamanders 

 used were formed over a mud bank on the low-energy side of the river. 

 During this period, leaves were intact or only slightly decomposed, and 

 many potential food invertebrates were present in the leaf litter. 



Captive Behavior 



Retreats. — As did tagged animals in the field, captive N. lewisi 

 used rocks for cover or hid under artificial plexiglass plates when rocks 

 were not provided. Development of a retreat began within 24 hr after 

 the animal was placed in an aquarium. Retreats were made under cover 

 by moving the underlying sand and gravel to either side of the develop- 

 ing cavity, or by forcing it from under the rock. Substrate was removed 

 by shoveling or pushing with the snout and head. Excess sand and 

 gravel was moved through an opening that eventually became the main 

 opening to the cavity. This excess material formed an elliptical shelf 

 directly in front of the opening. The cavity itself usually was oval in 

 shape, and had a diameter measuring approximately two-thirds the total 

 length of the animal. 



Once developed, the retreat area was maintained by the attending 

 animal. Gravel and sand were occasionally moved from the cavity by 

 mouth or by shoveling with the snout. Oral intake of gravel was 

 observed in three females and one male. The pieces, approximately 3 

 mm diameter, were moved to the shelf in front of the main entrance and 

 expelled from the mouth with some force. A female was observed 

 repeatedly picking up the same piece of gravel in the mouth, then plac- 

 ing it in a different location each time (like a bird placing sticks in a 

 nest). The shelf areas of all captives were maintained devoid of algae 

 and any other debris that commonly formed on surrounding gravel. The 

 rock surface directly over the cavity was similarity maintained. 



In order to test cleaning behavior (?) by N. lewisi, flat granite rocks 

 with algae encrusted upper surfaces were collected from the Litte River. 

 They were then placed in a tank, on sand and gravel substrate, with 

 their encrusted surfaces facing down. Three adult N. lewisi were released 

 into the tank and established residence under the rocks. Within 48 hr, 

 the area of rock undersurface that covered each retreat was clear of 

 algae. 



In microhabitats used by "tagged" animals in the field we on sev- 

 eral occasions observed similar retreat area maintenance. When animals 

 were under broken granite rocks the opening was always on the down- 

 stream side of the rock. The shelf area in front of the opening was 



