Took this from antweb. Take from it what you will. It's just a hypothesis anyways.
"Creighton (1950) ventured the opinion that replete production was a matter of opportunity, rather than an innate, physiological phenomenon. He further offered the idea that replete production is not a response to arid desert conditions but rather that it is the result of a previous xerophile having invaded less rigorous climes. In other words, replete production is the result of the availability of an excessive amount of nectar available in mesophytic environments."
The trademark honeypot repletes that you see are not filled with protein or fat but really only store sugars. Most of the species or variants that have been found to have repletes are also exclusively insectivores. Myrmecocystus is an extremely understudied genus, and from what I've read on antweb, whether honeypots make repletes is determined by the location where they were caught, as there are species of Myrmecocystus where replete production is completely absent in some locations but found in others.
That being said, it seems like the honey pots which are not physiologically used to making repletes do still have some sort of capability to do so.
I'm assuming we are calling Myrmecocystus spp. "honeypots".
All Myrmecocystus spp. utilize and produce repletes. The only thing your quote from Creighton(1950) there says is that their ancestors come from less harsh environments and that they will only produce repletes given the resources to do so. Well..... of course. They don't fill the repletes with air. If they can't collect the resources to fill a replete.... they can't fill a replete.
Here is a quote from AntWiki: "in 1908 Wheeler described melliger mimicus, an ant in which he noted the lack of repletes even though he excavated many nests. However, a few years later he received samples from San Diego County, California, of Myrmecocystus mimicus which included repletes. Rather than abandon or re-evaluate his theory Wheeler (1912) described these ants as a new subspecies (lomaensis) closely allied to mimicus. It is true that he made note of some differences to separate the California specimens, but it is equally true, as Creighton (1950) has shown, that these differences will not hold up. Indeed, they do not even exist, for it is quite possible to find specimens in the type series of melliger mimicus which differ from melliger mimicus in precisely the same manner as does melliger lomaensis."
You are not the first person to suggest that there are species of Mymrecocystus that do not produce repletes. It is scientifically unfounded. Given the opportunity, all species of Myrmecocystus have and will produce repletes. It is likely that this is more a demonstration in our lack of self-awareness when excavating nests and expecting to find repletes than this is that some species do not produce repletes outright.
Currently Keeping:
Trachymyrmex septentrionalis
Pheidole pilifera
Forelius sp. (Monogynous, bicolored) "Midwestern Forelius"
Crematogaster cerasi
Pheidole bicarinata
Aphaenogaster rudis
Camponotus chromaiodes
Formica sp. (microgena species)
Nylanderia cf. arenivega