Wood specific gravity (WSG) is an important biometric variable for above-ground biomass calculations in tropical forests. Sampling a sufficient number of trees in remote tropical forests to represent the species and size distribution of a forest to generate information on WSG can be logistically challenging. Several thousands of wood samples exist in xylaria around the world that are easily accessible to researchers. We propose the use of wood samples held in xylaria as a valid and overlooked option. Due to the nature of xylarium samples, determining wood volume to calculate WSG presents several challenges. A description and assessment is provided of five different methods to measure wood sample volume: two solid displacement methods and three liquid displacement methods (hydrostatic methods). Two methods were specifically developed for this paper: the use of laboratory parafilm to wrap the wood samples for the hydrostatic method and two glass microbeads devices for the solid displacement method. We find that the hydrostatic method with samples not wrapped in laboratory parafilm is the most accurate and preferred method. The two methods developed for this study give close agreement with the preferred method (r(2) > 0.95). We show that volume can be estimated accurately for xylarium samples with the proposed methods. Additionally, the WSG for 53 species was measured using the preferred method. Significant differences exist between the WSG means of the measured species and the WSG means in an existing density database. Finally, for 4 genera in our dataset, the genus-level WSG average is representative of the species-level WSG average.