jats:titleAbstract</jats:title> jats:pWe investigate the stellar populations for a sample of 161 massive, mainly quiescent galaxies at 〈jats:italicz</jats:italic> jats:subobs</jats:sub>〉 = 0.8 with deep Keck/DEIMOS rest-frame optical spectroscopy (HALO7D survey). With the fully Bayesian framework jats:monospaceProspector</jats:monospace>, we simultaneously fit the spectroscopic and photometric data with an advanced physical model (including nonparametric star formation histories, emission lines, variable dust attenuation law, and dust and active galactic nucleus emission), together with an uncertainty and outlier model. We show that both spectroscopy and photometry are needed to break the dust–age–metallicity degeneracy. We find a large diversity of star formation histories: although the most massive (jats:italicM</jats:italic> jats:sub⋆</jats:sub> > 2 × 10jats:sup11</jats:sup> jats:italicM</jats:italic> jats:sub⊙</jats:sub>) galaxies formed the earliest (formation redshift of jats:italicz</jats:italic> jats:italic jats:subf</jats:sub> </jats:italic> ≈ 5–10 with a short star formation timescale of jats:italicτ</jats:italic> jats:subSF</jats:sub> ≲ 1 Gyr), lower-mass galaxies have a wide range of formation redshifts, leading to only a weak trend of jats:italicz</jats:italic> jats:italic jats:subf</jats:sub> </jats:italic> with jats:italicM</jats:italic> jats:sub⋆</jats:sub>. Interestingly, several low-mass galaxies have formation redshifts of jats:italicz</jats:italic> jats:italic jats:subf</jats:sub> </jats:italic> ≈ 5–8. Star-forming galaxies evolve about the star-forming main sequence, crossing the ridgeline several times in their past. Quiescent galaxies show a wide range and continuous distribution of quenching timescales (jats:italicτ</jats:italic> jats:subquench</jats:sub> ≈ 0–5 Gyr) with a median of jats:inline-formula jats:tex-math

</jats:tex-math> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mo stretchy="false">〈</mml:mo> mml:msub mml:mrow mml:miτ</mml:mi> </mml:mrow> mml:mrow mml:miquench</mml:mi> </mml:mrow> </mml:msub> <mml:mo stretchy="false">〉</mml:mo> mml:mo=</mml:mo> mml:msubsup mml:mrow mml:mn1.0</mml:mn> </mml:mrow> mml:mrow mml:mo−</mml:mo> mml:mn0.9</mml:mn> </mml:mrow> mml:mrow mml:mo+</mml:mo> mml:mn0.8</mml:mn> </mml:mrow> </mml:msubsup> <mml:mspace width="0.50em" /> mml:miGyr</mml:mi> </mml:math> <jats:inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="apjac449bieqn1.gif" xlink:type="simple" /> </jats:inline-formula> and of quenching epochs of jats:italicz</jats:italic> jats:subquench</jats:sub> ≈ 0.8–5.0 (jats:inline-formula jats:tex-math

</jats:tex-math> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mo stretchy="false">〈</mml:mo> mml:msub mml:mrow mml:miz</mml:mi> </mml:mrow> mml:mrow mml:miquench</mml:mi> </mml:mrow> </mml:msub> <mml:mo stretchy="false">〉</mml:mo> mml:mo=</mml:mo> mml:msubsup mml:mrow mml:mn1.3</mml:mn> </mml:mrow> mml:mrow mml:mo−</mml:mo> mml:mn0.4</mml:mn> </mml:mrow> mml:mrow mml:mo+</mml:mo> mml:mn0.7</mml:mn> </mml:mrow> </mml:msubsup> </mml:math> <jats:inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="apjac449bieqn2.gif" xlink:type="simple" /> </jats:inline-formula>). This large diversity of quenching timescales and epochs points toward a combination of internal and external quenching mechanisms. In our sample, rejuvenation and “late bloomers” are uncommon. In summary, our analysis supports the “grow-and-quench” framework and is consistent with a wide and continuously populated diversity of quenching timescales.</jats:p>