When does secat equal relcat? (Q2636783)

This paper continues and slightly extends work by the same authors in [Topology Appl. 160, No. 5, 766--783 (2013; Zbl 1264.55004)]. In their earlier paper the authors defined and established basic properties for a number of approximating invariants for the so-called \textit{sectional category} of a map. One such approximation, which they call the \textit{relative category} of a map, and its relation to sectional category, is the subject of this article. It should be noted that the term ``relative category'' of a map is used in at least two other (and different) senses in the general Lusternik-Schnirelmann category area. In particular, the ``Fadell'' version of relative category (cf.~ Prop.7.5 of [\textit{O. Cornea} et al., Lusternik-Schnirelmann category. Mathematical Surveys and Monographs. 103. Providence, RI: American Mathematical Society (AMS). (2003; Zbl 1032.55001)]) uses an iterated fibre-cofibre construction for its definition, whereas an iterated fibrewise join construction is used to define the relative category of the present paper. This paper focuses on the question of when sectional and relative category agree. Generally, they may differ by at most \(1\), as was established in their earlier paper, and as follows easily from the definitions (Proposition 12). Knowledge of this could, in principle, be useful for studying sectional category since, for example, relative category is not increased by homotopy pushouts, whereas this is not so in general for sectional category. Much of the paper consists of convenient review of results, and direct consequences of results, from the earlier paper. The main new contribution seems to be Theorem 14, which gives certain dimension and connectivity conditions under which sectional and relative category agree.