ridfe believes to be the case with spiders making a double door branched 

 nest, that when an enemy succeeds in getting entrance to the main tube 

 the spider will run into the branch.^ This nest is represented in fig. 15, 

 Plate V. Though the tube is represented nearly natural size the length 

 is much fore-shortened, The space represented by the dotted lines a 

 and (5 was 10 inches for each. The spider was found at (^; the cork 

 door can be seen at c. 



Another specimen was found, about which, unfortunately, I have no 

 record nor recollection, except that it was collected sometime during the 

 spring. This, though I have not given it a careful examination I believe 

 to belong to the same species. I left it in the possession of Dr. Geo. 

 Marx, at Washington. 



The specimen collected Apr. 6, I placed in a jar of earth to see 

 what kind of a nest it made in captivity. For several days it showed no 

 signs of working. Finally I found it buried in the earth much as Avis'/- 

 valvata marxii is, when having just completed the dome over the tube, 

 and from the appearance of things I think the burrow was started, and 

 the dome made in the same way as in the case of that species. The 

 next day in place of the dome, in which there was no moss, was a perfect 

 cork door with an abundance of moss in it. The spider had evidently 

 cleared away the dome, which was made of earth and silk, and made 

 the door of new material. 



When I lifted up this door the spider would catch hold of it. In 

 a few days I found this door fastened down and that end of the tube 

 filled with earth for a distance of 5 cm. The tube extended in a circui- 

 tous direction for 10 cm or more where it came to the surface and was 

 closed by another nicely fitting trap-door. I think the spider was alarm- 

 ed at the discovery of her nest, and attempted by this strategy to deceive 

 her enemy. If the jar of earth had been more capacious, and a longer 

 time given the spider she might have made a branch and second door. 



' Food habits. I have not yet had an opportunity of making very ex- 

 tensive observations on the food habits of these species, and cannot say 

 that each one is confined to the insect for which I am certain it has a 

 special fondness as an article of diet. Of the species o f Pa f4>'/o/«(?/'«srny 

 observations have been confined to P. caribivorus. I have already stated 

 the fondness which this species has forcarabid beetles,^ and though kill- 

 ing several ants and flies did not use them for food, During July, while 

 at Ithaca N. Y. I fed several carabid beetles to a P. caribivorus which I 



1 Harvesting Ants and Trap-Door Spiders. 

 •-' Am. Nat. July 1886. Vol. 20, p. 592. 



