22 
form of Corallina officinalis. These plants in their turn 
provide a locus for the attachment of a host of other 
genera (Plate VII). As week succeeds week new arrivals 
attach themselves to the plants already present until in 
the summer the strata of epiphytes may be five deep. 
During the winter the population of these pools suffers 
great denudation. Everything is removed except the 
broken and stunted remains of the Covallina ; even the 
surface of the Lithophyllum looks as if it has been planed 
down to a minimum basal layer. Limpets in these 
pools stand up as minute conical islands and suffer 
denudation less severely than the surrounding plane 
surfaces. Microscopic examination shows that the limpet 
shells are covered by the basal remains of a large number 
of plants and that there are also small fragments of 
thalli of various genera; these have put out rhizoidal 
attachments and anchored themselves in the interstices 
of the shells. During the depth of the winter when 
nothing may be showing in the upper pools except 
debilitated Lithophyllum and Corallina, the limpet islands 
may still bear a comparatively varied though not luxuriant 
flora. It is noticeable that the genera found in such 
situations are all plants provided with a spongy or fibrous _ 
system of attachment organs; just the type that can 
most easily attach itself to the ridges and furrows of the 
limpet shells. Microscopic investigation of the limpets 
also reveals numerous minute sporelings in various 
stages of development, representing genera and species 
not in evidence in the pool flora except on the limpet 
shells. These genera and species may, however, be 
found lurking in pools at deep water level or as stunted 
plants in sheltered situations. The limpet island flora 
therefore represents relics of an autumn migration that 
sets seaward at the onset of unfavourable conditions. 
Development of this relic flora takes place early in the 
