1907.] Wheeler, Fungus-groiving Ants of North America. 793 



me by Professors K. Escherich and F. W. Neger it would seem that in the 

 case of the ambrosia beetle Tri/podendron Uneatus the fungus is found only 

 in the mycelial and conidial stages when the insects are present, but that 

 when these have been removed stromata with globular or flattened heads, 

 similar to those figured by Hedgcock for Graphiuvi atrovirens and amhrosii- 

 geruni, are produced on the walls of the galleries. 



The constant association of certain species of ambrosia beetles -s^ith 

 certain species of fungi, irrespective of the kind of wood on which they 

 grow, indicates that the mother beetles must be instrinncntal in transferring 

 the plant from colony to colony and from tree to tree in some manner analo- 

 gous to the fungus transfer of the Atta queen when establishing her for- 

 micary. Hedgcock seems to have found evidence of some such transference 

 of Ceraiostomella conidia. He says: "These are readily disseminated by 

 the wind and are probably carried by insects which penetrate the wood 

 and bark of trees, like most of the ambrosia and bark beetles. At the stage 

 in which the conidia form a mucilaginous mass, they adhere readily to any 

 insect that may pass over them. In the laboratory a number of species of 

 mites which feed on fungi carried spores on their bodies from colony to colony 

 in an agar plate to a sterile portion of the surface of the medium and started 

 new colonies of the fungus. Bark beetles were placed in a dish with the 

 conidial stage of CeraiostomcUa and after allowing them to remain a short 

 time were transferred to sterile agar plates Avhich were inoculated with spores 

 from the insects. It is probable that some species of insects feed on the 

 conidial stage of Ceratostomella, especially one or more species of ambrosia 

 beetles and a number of mites infesting their channels in the wood; but 

 proof is yet lacking on this point. The constant occurrence of this fungus 

 in the channels of a number of wood boring beetles indicates that the co- 

 nidia or the ascospores must be carried in some manner by these insects." 



Interesting as are the observations on the fungicolous ants, termites and 

 beetles collated in the preceding pages, we must admit that they are still frag- 

 mentary and leave many fundamental questions unanswered. It will be 

 seen that our knowledge of the fungi cultivated by all three of these insect 

 groups is very unsatisfactory and that many more investigations must be 

 undertaken before we shall be able to determine the precise taxonomic 

 affinities of the plants and to estimate the extent of the modifications induced 

 in their growth by the symbiotic insects. Equally fragmentary is our knowl- 

 edge of the phylogenetic origin and development of the fungus-growing 

 habit. Indeed, this problem in the termites and ambrosia beetles has 

 scarcely been recognized a yet. The views that have been entertained in 

 regard to the phylogeny of the Attii and their habits are perhaps, of sufficient 

 interest to command attention till further observations are forthcoming. 



