Hysteresis losses W have been calculated as a function of temperature T, bias field ${\mathit{H}}_{\mathit{b}}$, and amplitude ${\mathit{h}}_{0}$ for planar geometry in the critical-state framework with ${\mathit{j}}_{\mathit{c}}$=\ensuremath{\alpha}(T)/${\mathit{B}}^{\mathit{n}}$ for n=0, 1/2, and 1 in the range ${\mathit{H}}_{\mathit{b}}$/${\mathit{h}}_{0}$\ensuremath{\lesssim}2. ${\mathrm{\ensuremath{\chi}}}_{\mathrm{max}}$, the maximum of the ac susceptibility, ${\mathrm{\ensuremath{\chi}}}_{\mathrm{ac}}$=W/\ensuremath{\pi}${\mathrm{\ensuremath{\mu}}}_{0}$${\mathit{h}}_{0}^{2}$ versus T/${\mathit{T}}_{\mathit{c}}$ for various ${\mathit{h}}_{0}$, ${\mathit{H}}_{\mathit{b}}$, and \ensuremath{\alpha}(T) are seen to trace a ``universal'' ac-loss valley for each n with 0n\ensuremath{\le}1 when plotted versus ${\mathit{H}}_{\mathit{b}}$/${\mathit{h}}_{0}$. $T sub max--- corresponding to ${\mathrm{\ensuremath{\chi}}}_{\mathrm{max}}$ graphed versus ${\mathit{H}}_{\mathit{b}}$/${\mathit{h}}_{0}$ has a peak whose position differs significantly from that of the minimum in the ac-loss valley. ${\mathrm{\ensuremath{\chi}}}_{\mathrm{ac}}$ at the onset of full penetration, denoted ${\mathrm{\ensuremath{\chi}}}_{\mathrm{*}}$, and the corresponding T, denoted ${\mathit{T}}_{\mathrm{*}}$, are seen to deviate markedly from ${\mathrm{\ensuremath{\chi}}}_{\mathrm{max}}$ and ${\mathit{T}}_{\mathrm{max}}$ in all these circumstances.