Summary We compared hydraulic architecture, photosyn-

thesis and growth in Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), a shade-intolerant species, and western hem-lock (Tsuga heterophylla (Raf.) Sarg.), a shade-tolerant spe-cies, to study the temporal pattern of release from suppressive shade. In particular, we sought to determine whether hydraulic architecture or photosynthetic capacity is most important in constraining release. The study was conducted at two sites with mixed stands of 10- to 20-year-old Douglas-fir and western hemlock. At one site, the stand had been thinned allowing re-lease of the understory trees, whereas at the other site, the stand remained unthinned. Douglas-fir had lower height growth (from 1998–2003) and lower relative height growth (height growth from 1998 to 2003/height in 1998) than western hem-lock. However, relative height growth of released versus sup-pressed trees was higher in Douglas-fir (130%) than in western hemlock (65%), indicating that, although absolute height growth was less, Douglas-fir did release from suppression. Re-lease seemed to be constrained initially by a limited photo-synthetic capacity in both species. Five years after release, Douglas-fir trees had 14 times the leaf area and 1.5 times the leaf nitrogen concentration (Narea) of suppressed trees. Needles of released western hemlock trees had about twice the maxi-mum assimilation rate (Amax) at ambient [CO2] as needles of suppressed trees and exhibited no photoinhibition at the high-est irradiances. After release, trees increased in leaf area, leaf N concentration and overall photosynthetic capacity. Subse-quently, hydraulic architecture appeared to constrain release in Douglas-fir and, to a lesser extent, in western hemlock. Re-leased trees had significantly less negative foliar δ13C values than suppressed trees and showed a positive relationship be-tween leaf area:sapwood area ratio (AL/AS) and δ13C, suggest-ing that trees with more leaf area for a given sapwood area ex-perienced a stomatal limitation on carbon gain. Nonetheless, these changes had no significant effects on leaf specific con-ductivities of suppressed versus released trees of either species, but leaf specific root conductance was significantly lower in re-leased Douglas-fir