Forest canopies support diverse and abundant mini-ecosystems of vascular and non-vascular plants and arboreal soils that are perched on branches and trunks. These provide experimental arenas to study long-term effects and processes of disturbance and recovery of miniature natural landscapes. I synthesize four studies. 1) Stripping of plants and soils from branches in Monteverde, Costa Rica, resulted in recovery that required over 20 years, with new growth initiating from the branch underside. Similar experiments on Acer macrophylllum trees in Washington State resulted in encroachment into the cleared plot from the side. 2) Experimental delivery of extra propagules to branch substrates that retained intact plant communities retained significantly more propagules than stripped substrates (24% vs. 5%). 3) Clipping aboveground epiphyte parts (with soil intact) resulted in slow recovery of the subset of epiphytes that sprout from roots; however, over 30% of the original flora did not return after 7 yr. 4) Epiphyte mats that were excised and exposed to hotter, drier conditions by placing them on trees at lower elevations where mist is less frequent, resulted in epiphyte decline and death, followed by the germination of terrestrial gap-colonizing species that were stored in the seedbank of the epiphyte mats. Clipping the aboveground parts of mats in the greenhouse evoked a similar response of germinating invasive plants, whereas unclipped mats did not support this growth. These experiments indicate that epiphyte mats have a complex set of recovery mechanisms. Working with them may yield insights on forest restoration at a larger spatial scale.