I report lithium abundances and radial velocities for 15 late F, G and K st
ars in the Coma Berenices open cluster. All the stars are proper motion mem
bers, and the radial velocities enable us to confirm membership for ii sing
le stars and identify three SB1 systems and one short-period SB2 system tha
t are very probable cluster members. The pattern of Li abundances is simila
r to that in the Hyades, with some evidence for higher Li abundances in Com
a Ber for T-eff < 5700 K. I argue that these data do not support standard m
odels for Li depletion involving only convective mixing, because the Coma B
er stars have consistently determined lower metallicities and shallower con
vection zones than the Hyades and should have suffered a great deal less Li
depletion than observed. Instead, it seems that the Li abundances in these
and other young clusters are primarily ranked according to their age rathe
r than their metallicity. I propose a scenario where the extreme metallicit
y dependence of pre-main-sequence Li depletion is tempered by a mechanism t
hat severely inhibits convective mixing, where continued main-sequence depl
etion occurs via non-standard mixing mechanisms, and where any dependence o
f Li depletion on metallicity takes at least several hundred Myr to emerge.