Three replicate artificial canopy gaps of four sizes (ca. 40 m2 ca. 150 m2 ca. 350 m2 and ca. 2500 m2) were made in rain forest on Maracv° Island. Gaps were defined according to Brokaw (1982)and all vegetation (>2 m tall) within them was felled and left in place. The nine most numerous dicotyledonous tree (‚â•10 cm dbh) species of unfelled forest were chosen for studies on size‚Äìclass distribution sprouting and seedling dynamics in the artificial gaps. Seedling dynamics of two pioneer species one small understory tree species and a shrub are also presented as well as the response of the herbaceous vegetation. The artificial gaps were ephemeral in that within a few months of their creation the growth of sprouts and herbaceous vegetation often exceeded 2 m height and hence the strict definition of gap no longer applied. Later there was much young tree and liana growth and the upper branches of the surrounding trees were encroaching into the gaps reducing their area (after 4 yrs) by as much as 85%. The gaps caused enhanced seedling recruitment for at least a year. Cecropia palmata and Jacaranda copaia which are well-known pioneer species were both recruited in all gaps including the smallest ones. Tetragastris panamensis behaved similarly to these two species and had many seedlings recruited in all gap sizes. The survivorship and height growth of all seedlings were recorded in sample quadrats in each gap size. Data for nine species which had reasonably large seedling cohorts are presented separately. The gaps reduced the mortality rate and increased the rate of height growth of the seedlings present at the time of gap creation. Of the nine leading species Brosimum lactescens had the lowest mortality rate and almost all the C. palmata and J. copaia recruited after gap creation died by the end of the study. Mean height of seedlings in the forest plots varied little throughout the 4-yr study while small seedlings of B. lactescens Himatanthus articulatus and Lindackeria paludosa had reached heights up to 7 m in the large and very large gaps. The results provide no support for the notion that rain forest species are adapted to gaps of a certain size. Certainly there are light-demanding species and shade-bearing species but the observed variations within individual species the heterogeneity of the gap environment and the ephemeral nature of gaps all make it unlikely that a species is best adapted to one gap size.