206 
Psyche 
[June 
finally shrank back from the sticky silk. After about 30 minutes 
the ants began to close the entrance from within. No more ants 
emerged. Up to this point the spider remained in one corner of its 
web. As soon as the ants withdrew into the nest, the spider wrapped 
its captives with more silk and then began to suck them out, one 
after another (Fig. 4). 
The entrance of the ants’ nest remained closed. After three days 
a new entrance was found about 1 m from the previous one. It 
would seem therefore that the spider can prey only once at a given 
entrance. Further observations on other harvester ant hunting 
Steatoda fulva were essentially the same. 
As mentioned before, we found Steatoda fulva with Pogono- 
myrmex badius on 10 occasions. Once a S. fulva was seen at the 
entrance of the fungus growing ant Trachymyrmex septentrionalis 
(McCook). However this exception seemed to be an accident as 
this very small Trachymyrmex nest opened within a Pogonomyrmex 
nest crater and the entrances were only 15 cm apart. Thus it 
might be that the spider, attracted by the Pogonomyrmex nest area, 
chose the wrong entrance. 
Clearly, it would be most interesting to investigate the signals by 
which the spider locates its prey and the nest entrance. Pilot experi- 
ments in an olfactometer arena have shown that hungry spiders orient 
towards a slight air current which carries the odor of Pogonomyrmex 
workers rather than an odorless air current. It is not yet known 
whether the spider reacts to only specific ant odors, but it is remark- 
able that the distribution of Steatoda fulva (Levi, 1957) coincides 
very well with the range of the genus Pogonomyrmex (Cole, 1968). 
The related Steatoda albomaculata (De Geer) [= Lithypantes 
albomaculatus\ also often feeds on ants (Levi, 1957). 
Acknowledgements 
I thank Dr. and Mrs. H. W. Levi, Dr. E. O. Wilson and Mr. 
D. S. Woodruff for critically reading the manuscript. I am grateful 
to Dr. Levi for the determination of the spiders, to Dr. A. J. 
Meyerriecks for his hospitality at the University of South Florida, 
to my wife, Turid, for her collaboration in the field work, and to 
Miss Kathy Horton for her help in preparation of the manuscript. 
This work was supported by a grant from the Max Kade Foundation 
and the U. S. National Science Foundation (Grant No. GB-7734, 
E. O. Wilson, Sponsor). 
