1979] Haskins & Haskins — Rhytidoponera metallica 301 
Mode of Testing: 
As mentioned, reactions of workers on meeting were usually 
immediate and quite clear cut, although some anomalies were 
observed. The procedure used was simply to introduce the individ¬ 
ual to be tested into the colony tested against, or, alternatively, to 
place two workers together in a confined space, and to observe their 
behavior over periods ranging from less than a minute (when incom¬ 
patibility was unquestionably evident, as in immediate and violent 
attack) up to an hour if necessary to be sure of compatibility. Reac¬ 
tions were arbitrarily (but, it is believed, reasonably accurately) 
scored in four “envelope” categories: 
A. Extreme incompatibility, accompanied by attack, vigorous 
attempts to sting, and frequently tight body “closure”. 
B. Evident incompatibility, usually with seizure of antennae or 
legs, but without the violent “closure” of A. Both A and B were 
taken to mean marked incompatibility. 
C. Initial “starting back” or alarm on first encounter, and some¬ 
times brief running, but no threat or attack. The situation is 
commonly followed by fraternization and complete com- 
patability. 
D. Complete compatibility from the outset, lasting permanently, 
and frequently accompanied by vigorous antennal cleaning and 
sometimes mutual body licking. C and D may be justifiably inter¬ 
preted, we believe, as ultimate compatibility, although obvious 
distinctions initially recognized in the C-type reactions were 
sometimes interesting. 
Categories A and D were very clear cut. Inevitably, some observer 
subjectivity entered into assignments of B and C categories, but 
every effort was made to estimate them accurately. Very occasion¬ 
ally a reaction was sufficiently indefinite, or changed sufficiently 
with time, so that it had to be judged a “border case” and was so 
designated — as, for example “C/D”. 
The tests summarized below were designed to test several aspects 
of the whole question, as indicated. 
Controls 
Two sets of controls were run where the results to be expected 
were obvious, primarily to test the validity of the experimental con¬ 
ditions. One was between workers drawn from within a small con¬ 
tiguous area, the second between two fairly well-separated popula¬ 
tions within a single collection location. 
