The spatiotemporal dynamics in systems of active self-propelled particles is controlled by the propulsion mechanism in combination with various direct interactions, such as steric repulsion, hydrodynamics, and chemical fields. Yet, these direct interactions are typically anisotropic, and come in different "flavors", such as spherical and elongated particle shapes for steric repulsion, pusher and puller flow fields for hydrodynamics, etc. The combination of the various aspects is expected to lead to new emergent behavior. However, it is a priori not evident whether shape and hydrodynamics act synergistically or antagonistically to generate motility-induced clustering (MIC) and phase separation (MIPS). We employ a model of prolate spheroidal microswimmers-called squirmers-in quasi-two-dimensional confinement to address this issue by mesoscale hydrodynamic simulations. For comparison, non-hydrodynamic active Brownian particles (ABPs) are considered to elucidate the contribution of hydrodynamic interactions on MIC and MIPS. For spherical particles, the comparison between ABP and hydrodynamic-squirmer ensembles reveals a suppression of MIPS due to hydrodynamic interactions. Yet, our analysis shows that dynamic clusters exist, with a broad size distribution. The fundamental difference between ABPs and squirmers is attributed to an increased reorientation of squirmers by hydrodynamic torques during their collisions. In contrast, for elongated squirmers, hydrodynamics interactions enhance MIPS. The transition to a phase-separated state strongly depends on the nature of the swimmer's flow field -with an increased tendency toward MIPS for pullers, a reduced tendency for pushers. Thus, hydrodynamic interactions show opposing effects on MIPS for spherical and elongated microswimmers. Our results imply that details of the propulsion mechanism of biological microswimmers, like pattern and time dependence of the flagellar beat, may be very important to determine their collective behavior. arXiv:1807.01211v1 [cond-mat.soft]