THE COMMUNITY 



486 



this association in Panama are those of the 

 brilliantly colored tree frog Agalychnis 

 dacnicolor, which suspends its egg masses 

 above the water on the trunk of the tree; 

 and the tadpoles of the equally vivid ter- 

 restrial dendrobatid, Dendrohates auratus, 

 which are transported into the water holes 

 by the adult male (Dunn, 1931). 



The tree hole, as a discontinuous ex- 

 tension of the floor, has a population of log 

 mold animals as its walls soften and the 

 cavity begins to fill with litter. Such animals 

 include oribatid and parasitid mites, pre- 

 daceous carabid, staphylinid, and pselaphid 

 beetles, and numerous ants, to mention only 

 a few forms. 



Considering the whole floor, animals of 

 this layer may be divided into patobionts,*' 

 those living all their normal life in this 

 stratum; floor transients (patocoles), those 

 spending a regular portion of their life out- 

 side of the floor; and patoxenes, or ac- 

 cidental visitors. 



The patobionts are numerous in species 

 and in individuals. One of the best word 

 pictures given this layer is that of Jacot 

 (1935, p. 425), which merits quotation: 



"Deer, jumping mice and the oven-bird are 

 denizens of the forest floor by virtue of using 

 it as their substratum, but there is also a host 

 of curious animals which use the forest floor, 

 especially the litter of dead leaves, twigs, 

 branches and fruit parts, as their walls, ceiling 

 and sub-basements. Looked at from the eye 

 level of the cockroach, this litter becomes a 

 several-story edifice of enormous extent. The 

 various floors are separated by twigs, midribs, 

 petioles, fruit husks, samaras, skulls, elytra and 

 feces. The lower one descends, the more com- 

 nact is the structure. The leaves become more 

 fragmentary, the feces of worms, which have 

 come up from the soil, of caterpillars which live 

 in the trees and of the inhabitants themselves, 

 as well as grains of sand brought up by the 

 worms and a heterogeneous assortment of 

 beetle skulls and wing covers, become more 

 abundant. This comolex is rendered more in- 

 tricate by the growth of minute fungus moulds 

 which feed upon the dead leaves and other 

 organic refuse, weaving it all into a compact 

 mat by their myriad white hyphae. Thus is the 

 woof woven into the warp of the woodland 

 rug." 



The floor residents are either primarilv 

 engaged in a ceaseless reduction of floor 

 leaf and log mold to soil humus or sec- 

 ondarily encased in feeding upon these 



• From the Greek patoma, floor. 



mold eaters. Under favorable conditions, 

 complete reduction of a leaf fall is con- 

 summated in two years in the moist virgin 

 forests of warm temperate zones. In north- 

 ern softwood forests the htter reduction is 

 accomplished much more slowly (Jacot, 

 1936a). Whether the criterion of reduc- 

 tion be the total amount of feces, or oxygen 

 consumption as a measure of total activity, 

 or food consumption (Bomebusch, 1930, 

 1930a; Thamdrup, 1932; Ulrich, 1933), this 

 process is a basic industry of the floor 

 stratum and vital to the whole community. 



Mold transformation is a large-scale, co- 

 operative process. Cellulose-splitting bac- 

 teria and fungi (Waksman, 1932); oribatid 

 and hoplodermatid mites (Jacot, 1936a, 

 1940; Williams, 1941); millipeds of such 

 typical genera as Diplohdus (Lyford, 

 1943), Fontaria (Rommell. 1935) and 

 Spiroholus; snails, such as Punctum pifg- 

 maeum, Striatura milium (Jacot, 1935a) 

 and Anguispira; myriads of Collembola: 

 numerous termites in tropical log mold 

 (Allee, 1926a); manv ants; larvae and 

 adults of many elaterid, passalid, and tene- 

 brionid beetles; these are a few representa- 

 tive groups united in this general activity. 



In this reduction complex the response of 

 a resident to the mold may be general or 

 specific. A common forest milliped (Dip- 

 loitdus caendeocincttis) was shown to have 

 a differential feeding response between 

 leaves from the same tree, between leaves 

 of neighboring trees of the same species, 

 and more strongly between leaves of differ- 

 ent tree species. These feedinsf reactions, 

 covering a two year experimental period 

 (Lyford, 1943), were foimd correlated with 

 the percentage of calcium in the leaf. Bass- 

 wood, elm, and hickory leaves vvere eaten 

 more freely than those of beech and oak 



As for coniferous leaves, available in- 

 formation suggests that firm, undecom- 

 posed needles are not attacked by floor 

 arthronods. Such leaves must first be acted 

 upon bv fungi, which reduce the cell con- 

 tents and much of the mesonhyll, leaving 

 the needle a shaft of soft punk. In this state 

 sxich needles as those of spruce, for ex- 

 ample, are attacked bv phthiracarid mites 

 (Jacot. 1936a, 1939). This mav be a partial 

 answer to the relative slowness of litter 

 reduction of coniferous floors previouslv 

 noted. These needle-fungus-mite reaction 

 chains are specific in spruce forests, and 



