1983] 
Haskins & Haskins — Rhytidoponera me tallica 
173 
been in direct contact during their ontogeny, when those workers 
are pair-tested in fingerbowls on an individual basis. This compati- 
bility is not universal, however. Incompatibility was observed in a 
few cases even between workers of two halves of a population 
separated for a year or more but maintained under identical envir- 
onmental conditions, including diet, whether tested in pairs or in 
groups of ten. When the diet had consistently differed markedly 
throughout the period of separation, the numbers of workers exhib- 
iting incompatibility appeared somewhat increased, but was still a 
minor proportion. It is possible that such individuals eliciting attack 
were in fact laying workers, as found by Holldobler in Novomessor. 
(2) When earth-containing Lubbock nests occupied by one fraction 
of the divided population throughout the periods of separation were 
introduced into the arena of another, the situation was dramatically 
altered. Mass hositility and mass raiding of the introduced nest by 
the recipient moiety regularly followed, regardless of whether the 
preceding dietary history was the same or different. We conclude 
that, as reported by other investigators in a number of higher ant 
genera {Pogonomyrmex, Oecophylla, Lasius) and in the Ponerine 
genus Hypoponera) colony-specific nest site marking is important 
also in Rhytidoponera metallica, despite its relative primitiveness 
and the typical diffuseness and vagility of its colonies. Typical ran- 
dom markings of the floors of arenas about earth-containing Lub- 
bock nests long occupied by colonies of metallica, as illustrated, 
indicate that, as with at least some higher ants, and in several Pone- 
rine genera including Paltothyreus, Leptogenys and Hypoponera, 
fecal contents are the characteristic marking “vehicle”, perhaps 
including, as in the higher ants, colony-specific pheromones. If this 
is true of R, metallica, as suggested in the experiments reported, it 
becomes interesting to consider the factors involved in mediating 
this specific reaction between two halves of a single population 
separated for less than two years and maintained on identical diets 
and in identical arenas placed side by side on the same laboratory 
bench during that period. No evidence has been found of trail mark- 
ing, or indeed of trail laying, in R. metallica. 
Acknowledgments 
We would like to express particular appreciation to Professor 
Bert Holldobler of Harvard University for his invaluable assistance 
and suggestions in this program. 
