Adiponectin is well known as an insulin-sensitising hormone secreted exclusively from adipose tissue. Paradoxically, adiponectin is increased during calorie restriction (CR) when typical adipose depots are reduced in volume. Interestingly, bone marrow adipose tissue (BMAT) increases in volume during CR and has been shown to be a major source of circulating adiponectin in this context. CR itself exerts diverse metabolic and skeletal effects: one hypothesis is that increased BMAT contributes to bone loss during CR, whereas many of the metabolic benefits of CR are similar to those ascribed to adiponectin. However, it remains unknown what causes BMAT to increase during CR; if BMAT contributes to circulating adiponectin in other conditions; and if BMAT and/or adiponectin influence skeletal or metabolic adaptations to CR. The goal of my PhD research was to address these critical gaps in knowledge. My first hypothesis was that BMAT is also a source of adiponectin following treatment with thiazolidinediones, a class of anti-diabetic drugs that increase both BMAT and adiponectin. To address this, mice resistant to BMAT expansion were treated with rosiglitazone, a well-studied thiazolidinedione, to investigate the ability of BMAT to secrete adiponectin in other conditions of BMAT expansion and hyperadiponectinaemia. These mice were less resistant to BMAT expansion in response to rosiglitazone compared to CR, which limited the ability to draw strong conclusions; however, there were some indications that BMAT is an important source of adiponectin in this context, and the data provide other insights about how BMAT compares to other adipose subtypes. To begin investigating the roles of BMAT and adiponectin during CR, wild-type (WT) male and female mice were placed on varying durations of CR or ad libitum control diet. The increases in BMAT were then assessed and compared to changes in skeletal architecture, adiponectin and glucose tolerance. I found that BMAT expansion depends not only on CR duration, but also varies by skeletal site and between the sexes. BMAT in tibiae and femurs increased with 6-week CR in both sexes, but only female femurs showed increased BMAT after 2-week CR. BMAT expansion generally coincided with bone loss in femurs but occurred after bone loss in tibiae; notably, bone loss occurred in humeri despite a lack of BMAT in these bones. Together, these findings show that BMAT expansion is not required for CR-induced bone loss. In terms of metabolic and endocrine effects, adiponectin was maximally increased after 3 to 4 weeks of CR, approximately coinciding with the increases in BMAT and adiponectin expression within bone. However, improvements in glucose tolerance occurred after only 1 week of CR, especially in male mice, preceding BMAT expansion and hyperadiponectinaemia. Strikingly, I found that females resist many of the metabolic benefits of CR: unlike males, during CR females did not lose fat mass and had only mild improvements in glucose tolerance. Collectively, it was not clear from these findings if BMAT expansion or adiponectin are playing a role in the metabolic benefits of CR. For this reason, adiponectin knockout mice (KO) and WT controls were subjected to CR to investigate if adiponectin contributes to CR-induced improvements in glucose homeostasis. Male KO mice performed better in glucose tolerance tests than WT mice, a finding going directly against my hypothesis. This indicates that adiponectin may be playing an unexpected role during CR, in which it impairs glucose tolerance. One possibility is that this relates to adiponectin’s role in stimulating fat oxidation; however, further work is required to elucidate the underlying mechanism. Together, these studies reveal several new discoveries. Firstly, CR-induced BMAT expansion is dependent on CR duration and is both sex- and skeletal-site-specific. Secondly, BMAT expansion is not required for CR-induced bone loss in humeri, but may contribute to bone loss in femurs and tibiae. Thirdly, female mice resist fat loss and improvements in glucose tolerance during CR, an unexpected finding that warrants further study. Fourthly, CR-induced improvements in glucose tolerance precede BMAT expansion and hyperadiponectinaemia, suggesting that the latter are not drivers of the former. Finally, lack of adiponectin is associated with improved glucose tolerance during CR, suggesting that the function of adiponectin during CR is distinct to its reported roles in obesity and insulin-resistant states. The latter indicate that adiponectin may impact the therapeutic benefits of CR and may provide insights toward the evolutionary function of adiponectin during periods of food scarcity.