Let us take the example of Recessive epistatsis, an epistasis in which a double recessive gene mask the phenotypic expression of alleles of another locus. (adapted from: An Introduction to Genetic Analysis)
An example from the same book:
In plant blue-eyed Mary (Collinsia parviflora), the biochemical pathway is as follows:
The w and m genes are not linked. If homozygous white and magenta plants are crossed, the F1 and F2 are as follows:
Complemention results in a wild-type F1. However, in the F2, a 9:3:4 phenotypic ratio is produced.
This kind of interaction is called epistasis, which literally means “standing on”; in other words, an allele of one gene masks the expression of the alleles of another gene. In this example, the w allele is epistatic on m+ and m$^1$. Conversely, m+ and m can be expressed only in the presence of w+.
Interpretation & Question:
The gene product of m$^+$ is needed to form blue pigment from magenta. On the other hand, in the absence of w$^+$ allele (i.e. w w genotype), no magenta (precursor of blue) is formed so the phenotypic expression of m$^+$ allele(s) is masked. Now, what has m got to do here? $^1$ It is not encoding anything.
On referring other books I found they do not suggest a hypostatic gene to be recessive (as $m$ has been considered here).
My idea of recessive allele is that one that due to loss-of-function mutation either produces a non-functional variant of the protein or the same protein at a lower quantity.
Has recessive allele been considered an hypostatic allele here considering that it might be able to code normal functional protein at a lower rate?
Scope of answer:
Which concept should be followed? Why (ideally with citation)?