138 LIFE : OUTLINES OF GENERAL BIOLOGY 



SHELTER-ASSOCIATIONS.— WTiile the contrast between a 

 parasitic relation and an cpizoic or epiphytic one is in most cases 

 clear, there may be some difficulty with what are called "shelter 

 associations". Thus the little pea-crab Pinnotheres pisum is often 

 found off English coasts sheltering in the mantle cavity of the 

 Norway Cockle ; and other members of the Pinnotherid family occur 

 in other bivalve molluscs, as well as in worm-tubes and corals. A 

 small bivalve is commonly embedded in the cellulose tunic of 

 Ascidians. Tlie slender fish called Fierasfer insinuates itself tail- 

 foremost into the end of the food-canal of sea-cucumbers, and it 

 also finds its way into some large bivalves and starfishes. It feeds 

 independently like any other fish, but it seems to enjoy the shelter 

 of animals in which there are active currents of water. WTien the 

 Holothurian is placed in water with insufficient aeration, the 

 Fierasfer comes out and rises to the surface taking gulps of air. 

 A small fish, Amphiprion, with resplendent coloiu-s, lives inside 

 a large sea-anemone, hiding itself deeply when disturbed. It does not 

 seem to do either good or harm to the sea-anemone, but it is said to 

 die when it is dissociated from its "host". This may serve to illustrate 

 how these relations shade into one another. 



Similarly some insects find shelter in plants on which they do not 

 feed. A spider may frequent a particular flower; another makes its 

 web below the insect-attracting margin of the pitcher of a pitcher- 

 plant. Many ants live in hollow stems and thorns, but feed elsewhere; 

 so they get nothing but shelter from the plant and confer no benefit. 

 Kut, in oth(T cases, the sheltering ants feed on secretions exuded 

 by glands on the leaf base, or leaf-tips, of which the inhabited thorn 

 is a stipule. In this case the ant has good reason to guard its leaf 

 from other ants, sometimes leaf-cutters, and promptly nips of! the 

 intruder's head accordingly. These ants may thus be viewed as a 

 l)odyguard; yet it is prudent to add that some naturalists regard the 

 evidence as unconvincing. Some plant-mites live in little shelters 

 ("domatia") on the plants they frequent, and it is difficult to draw 

 a line between those that puncture the plant and those that simply 

 clean the epidermis. The moist spaces between the leaves of the 

 r])iphytic Hromeliads of tropical forests afford shelter to an astonish- 

 ing number of more or less epiphytic insi^cts; and their inter-rela 

 tions with their habitat are sometimes .so subtle that any sharp 

 classification becomes impossible. Yet after all. is it not pushing 

 logic to pedantry, and non-evolutionary besides, to seek to put 

 every grade of inter-relation into a separate pigeon-hole? There are 

 also puzzling epiphytic shelter-associations between plant and 

 plant; one of the best known being the occurrence of the Alga, 

 Xostoc, in certain parts of the water-fern, Azolla, and in certain 

 liverworts, also. The Alga thus finds a sheltered and imiformly 

 moist habitat, biit it is not known to do either good or harm to its 



