INTRODUCTION xci 



On the coast and lowlands they visit exactly the same plants, as I have mentioned 

 in the case of the endemic bees {Nesoprosopis). In places where they breed some may 

 be observed in the greatest abundance flying about rocks, especially where these are 

 piled up to form rough walls, up and down which they are continually coursing. 



Their choice of nesting places is very varied. Some (like O. sociabilis) burrow in 

 the ground, especially in dry sandy lava. Others always form their burrows in dead 

 tree trunks, and many (e.g. O. obscure-pimctatus) are indifferent, breeding freely in posts 

 and dead tree trunks, or in the porous cavities of lava rock, which latter appear to be the 

 sole or chief nesting place of others. O. nigripenms and the allied O. 7'adula occupy 

 all sorts of nesting places, such as all those mentioned, and often the empty mud-cells of 

 Sceliphron. These and others (e.g. Nesodynerus rudolphi, O. montamis, hiloensis, etc.) 

 freely enter the verandahs and even the rooms of houses, where they make their cells 

 in cracks or holes in furniture, in the cavities of bamboo, and other situations. 

 O. nigripennis is so common about many houses, that it cannot fail to be often carried 

 on ships with furniture, from one island to another. It has been taken as single 

 specimens on Kauai, but whether it is really established there, I do not know. A few 

 species, Odynerus oahuensis being one of these, have abnormal nesting habits, forming 

 complete mud-cells, which are attached to the leaves of trees or to other objects. They 

 are generally placed in sheltered situations, sometimes for example in a curled up leaf 

 of Pelea, wherein a spider has made its nest ; and as various other creatures, including 

 young moUusca, seek the same shelter, we have several times found quite a collection 

 of animals around the wasp's cell. On one occasion this wasp was watched bringing 

 caterpillars to store an open cell, which was (with several already completed) suspended 

 from the rootlets of a large fallen Ohia tree, bridging a small mountain stream. On 

 other of the islands, where O. oahuensis does not occur, similar nests have been found, 

 but the makers have not been identified. 



The prey of Hawaiian Eumenidae, so far as is known, consists entirely of 

 caterpillars. The number of these contained in a single cell varies according to the 

 size of the species of wasp and according to that of the size of the caterpillars. A large 

 variety of caterpillars are employed, and a single species of wasp is not only not constant 

 to the use of one kind of caterpillar, but sometimes even mixes the species in a 

 single cell. 



On the whole it may be said that Pyralid and Microlepidopterous caterpillars are 

 the favourite prey and that Geometridae are rarely utilized. It is most remarkable, 

 seeing that the latter are occasionally taken (e.g. by O. montanus, eucharis, etc.), that 

 this should occur so rarely, for the Geometrid caterpillars are so very numerous that 

 they could be often obtained in any quantity. It is amusing to watch the common 

 O. nigripennis hunting the caterpillars of Omiodes, which shelter themselves by spinning 

 together the leaf or leaves of their food-plants. On Molokai a large ' wiliwili ' tree 

 {Erythrina) was the home of hundreds of the caterpillars of O, monogona and was 



m 2 



