Abstract:
On the basis of their association with ants, neotropical species of the genus Acacia may be
grouped into two broad categories. 'Ant-acacias', comprising less than 10% of the species
in Central America depend in varying degree on a mutualistic association with ants of
the genus Pseudomyrmex. The plants provide their ants with shelter in swollen stipular
spines and with nourishment from foliar nectaries and nutritive structures (Beltian bodies)
at the leaf tips. The ants in turn provide for the plants' protection against mammalian
and insect herbivores and against neighbouring plant competitors (Belt 1874; Brown
1960; Janzen 1966, 1967). Janzen (1967) has demonstrated that Acacia cornigera (L.)
Willd. plants cannot survive after experimental removal of their associated ant colonies.
The remaining species of Acacia in Central America comprise the second group, the
'non-ant-acacias'. These plants do not harbour mutualistic ant colonies, nor do they
possess the various morphological features of acacias with such colonies. Survival of
non-ant-acacias is presumably dependent, therefore, on other means of defence against
herbivores.
After noting that the foliage of non-ant-acacias was markedly bitter to human taste,
whereas that of ant-acacias was mild-tasting, Janzen (1966) proposed the following
hypotheses: (i) non-ant-acacias are protected from herbivores by the presence in their
foliage of toxic or repellant chemicals; (ii) symbiosis with ants has been evolved by antacacias
as an alternative means of protection; and (iii) chemical defence has subsequently
been lost in the ant-acacias, possibly because maintenance of both ant and chemical
defence places an unnecessary metabolic burden on the plant. In the present paper we
present evidence in favour of these hypotheses.
For experiments in the laboratory we selected A. cornigera, an ant-acacia, A.farnesiana
(L.) Willd., a non-ant acacia, and A. chiapensis Saff., a species showing intermediate
characteristics. A. cornigera and A. farnesiana are widely distributed through Central
America, but A. chiapensis is very limited in distribution. Plants of A. chiapensis are
normally found in nature in association with ant colonies, yet they can survive in the
absence of ants; therefore, this species will be regarded as a non-a nt-acacia for the
purposes of this study.
Since insects known to attack unoccupied ant-acacias in Central America were unavailable
for our work, the southern armyworm, Prodenia eridania (Cramer) (Noctuidae), was
selected for bioassay of the acacias. The caterpillar of this species is highly polyphagous
with an outstanding ability to detoxify insecticides (Krieger 1970; Krieger & Wilkinson
1969) and, presumably, toxic compounds encountered in its wide array of food plants (Gordon 1961; Krieger, Feeny & Wilkinson 1971). Any demonstrated toxicity of acacia
leaves to the southern armyworm should thus be of more general interest than would be
similar results obtained with a more specialized phytophagous insect.