It's complicated: The role of false brood cells as an anti-parasite strategy in Hymenoptera is not simple despite their ubiquity

Nest collection: Nests of A. morosus were collected from naturally
abscised fronds that skirt the base of the rough fern tree, Cyathea
australis in Kurth Kiln Regional Park, Yarra Ranges National Park,
Cathedral Range State Park, Marysville Bushland Reserve and Grants Picnic
Ground in the Dandenong Ranges region in eastern Victoria, Australia.
Nests were collected across seven separate collections from December 2017
to January 2021. Collections were performed at dawn when temperatures were
below 15°C to ensure all bees were in their nest. Nest entrances were
sealed upon collection and nests placed on ice in insulated boxes and
transported to Flinders University for nest processing. Nest processing:
Nests were opened lengthwise until the first brood cell was reached. All
adult female A. morosus and any adult parasites in the nest entrance were
removed and placed in 99% ethanol for preservation and identification.
Nests of A. morosus can be characterized as either i) social – in which at
least two adult females occupy the nest and provision brood with no false
cells; ii) sequential solitary – solitary female provisions brood with no
false cells; iii) false cell solitary – solitary female provisions brood
cells that are interspersed with false cells (Fig. 1). The brood cell
closest to the nest entrance was checked to determine the maturity of the
brood. The nest was fully opened if the brood had reached pupation or
adulthood, whereas, the nest was left to develop at ambient room
temperature if brood were still developing. Nest architecture (nest
length, brood cell lengths and false cell lengths) and contents (number of
brood, sex of brood, number of false cells, failed cells and parasites)
were recorded. Rates of parasitization (number of nests containing at
least one parasitized brood cell) were designated into suites depending on
their respective mode of entering and parasitising brood (i.e.
Gasteruption, Anthrax, Ephutomorpha; (Hearn et al., 2021)). We categorized
nest productivity using three metrics: as the total number of brood cells
produced, the number of surviving brood cells, and the proportion of
surviving brood cells in a nest. In A. morosus, it is important to make
this distinction between the total number of brood cells and the number of
surviving brood cells due to high rates of brood loss that can arise from
factors other than parasitization (e.g. microbial infection) (Spessa et
al., 2000). Nest structure: Amphylaeus morosus were only ever
found to construct false cells and never vestibular cells (Fig. 1).
Solitary nests of A. morosus are arranged linearly with false cells
separating closed brood cells. False cells are often greater in diameter
compared to brood cells as they are further enlarged by the
nest-constructing adult female who uses the pithy ‘excavations’ as a
barrier in front of the next brood cell (Fig. 1). In contrast, brood cells
are fully lined in a cellophane-like material produced by adult females
(Almeida, 2008).  Statistical analysis: All statistical tests
were performed using R version 4.2.2 (R Core Team, 2024). Some of the
variables analysed were strongly zero or one-truncated, and where data
were not normally distributed, we used non-parametric tests. Orphaned
nests of A. morosus (nests containing brood but no adult female) are
common and this can have implications for brood survival. For some
analyses, we therefore separated nests based on the number of adults
present at the time of collection. We used Kruskal-Wallis tests to compare
the differences in nest productivity variables based on the number of
adult females present. Since social nests never produced false cells (see
below), we restricted analyses of the effect of false cells on nest
productivity to nests with one or zero adult females. We used Mann-Whitney
pairwise comparisons to test the effect of false cells on nest
productivity (total number of brood cells provisioned and number of
surviving brood cells) between nests with one adult female and orphaned
nests. However, these analyses may be confounded by the fact that brood
production in orphaned nests may have been curtailed by death of the adult
female. We therefore also repeated the Mann-Whitney analyses using the
proportion of surviving brood cells rather than the absolute number. To
assess how the number of false cells influenced parasitism rates, we used
a binary logistic regression with a logit link function. We used a
generalized linear model with a Poisson distribution with a log link
function to test the influence of nesting strategy on count data (total
number of brood cells produced and total number of surviving brood cells).
Only nests containing at least one brood cell were used in all analyses.