6 



and for this purpose perhaps a tank breeding mosquitoes would be as good 

 as anything. The best time to undertake it would be about the 1st of June. 

 I will select some one as soon as I hear from you. 



Very truly yours, 



DAVID S. JORDAN. 



On receipt of yours of July n, 1905, with the advance of 

 S500 of the Territorial appropriation covering the expenses of 

 this undertaking, I started from Stanford University to the 

 Southern United States. It was my intention to secure the 

 top-minnows at or near New Orleans, but the rigid quaran- 

 tine in operation in Louisiana prevented my carrying out this 

 plan. Seabrook, near Galveston, Texas, was then selected as 

 the next place most available. At Seabrook I found the family 

 of top-minnows, Poea'hidae, in large numbers. They were 

 swarming in all the stagnant waters at sea-level as well as 

 in various ditches, ponds, and standing pools. Mosquitoes 

 are very plentiful in and about Seabrook, but after a study of 

 the situation I am convinced that their source is not the bodies 

 of water containing these fish but rather temporary and arti- 

 ficial breeding places, such as closed pools, tubs, tin cans, and 

 other refuse which are not accessible to these fish. 



As per his letter to you, Dr. Jordan recommended the fol- 

 lowing genera: Mollienesia, Adinia, Gambusia and Funduhts. 

 These are all members of the single family Pceciliidae or top- 

 minnows. I first made a careful examination of a number of 

 the stomachs freshly taken from members of the above genera. 

 The stomach-contents was found to consist largely of larvae 

 of various insects, including those of mosquitoes; egg-masses 

 of mosquitoes ; minute crustacean and some vegetation. The 

 results showed that Gambusia were the best insect feeders. 

 Of 100 stomachs of this genus examined all contained many 

 insect larvae and eggs among which I noticed especially num- 

 erous egg-masses of the mosquito. However, Mollienesia, 

 Fundulus and Gambusia differed slightly in regard to their 

 capacity for the various insect larvae and the difference was 

 probably due to the different food localities. The temperature 

 of the water in and about Seabrook in which these fish were 

 found ranged from 74 0 to 87 0 . 



It now remained to determine under what conditions the 

 fishes could be most successfully transported to Hawaii. Six 

 ordinary ten-gallon milk cans were prepared by puncturing 

 the covers with numerous holes and placing the cans in bran 

 sacks, the intervening space being tightly packed with Span- 

 ish moss. This served to keep the water at an even tempera- 



