ABSTRACT
Despite progress made in recent decades, gender bias is still present in scientific publication authorship. The underrepresentation of women and overrepresentation of men has already been reported in the medical fields but little is known in the fields of exercise sciences and rehabilitation. This study examines trends in authorship by gender in this field in the last 5 years. All randomized controlled trials published in indexed journals from April 2017 to March 2022 through the widely inclusive Medline dataset using the MeSH term "exercise therapy" were collected, and the gender of the first and last authors was identified through names, pronouns and photographs. Year of publication, country of affiliation of the first author, and ranking of the journal were also collected. A chi-squared test for trends and logistic regression models were performed to analyze the odds of a woman being a first or last author. The analysis was performed on a total of 5259 articles. Overall, 47% had a woman as the first author and 33% had a woman as the last author, with a similar trend over five years. The trend in women's authorship varied by geographical area, with the higher representation of women authors in Oceania (first: 53.1%; last: 38.8%), North-Central America (first: 45.3%; last: 37.2%), and Europe (first: 47.2%; last: 33.3%). The logistic regression models (p < 0.001) indicated that women have lower odds of being authors in prominent authorship positions in higher-ranked journals. In conclusion, over the last five years, in the field of exercise and rehabilitation research, women and men are almost equally represented as first authors, in contrast with other medical areas. However, gender bias, unfavoring women, still exists, especially in the last authorship position, regardless of geographical area and journal ranking.
ABSTRACT
Very large colonies of territorially dominant arboreal ants (TDAAs), whose territories are distributed in a mosaic pattern in the canopies of many tropical rainforests and tree crop plantations, have a generally positive impact on their host trees. We studied the canopy of an old Gabonese rainforest (ca 4.25 ha sampled, corresponding to 206 "large" trees) at a stage just preceding forest maturity (the Caesalpinioideae dominated; the Burseraceae were abundant). The tree crowns sheltered colonies from 13 TDAAs plus a co-dominant species out of the 25 ant species recorded. By mapping the TDAAs' territories and using a null model co-occurrence analysis, we confirmed the existence of an ant mosaic. Thanks to a large sampling set and the use of the self-organizing map algorithm (SOM), we show that the distribution of the trees influences the structure of the ant mosaic, suggesting that each tree taxon attracts certain TDAA species rather than others. The SOM also improved our knowledge of the TDAAs' ecological niches, showing that these ant species are ecologically distinct from each other based on their relationships with their supporting trees. Therefore, TDAAs should not systematically be placed in the same functional group even when they belong to the same genus. We conclude by reiterating that, in addition to the role played by TDAAs' territorial competition, host trees contribute to structuring ant mosaics through multiple factors, including host-plant selection by TDAAs, the age of the trees, the presence of extrafloral nectaries, and the taxa of the associated hemipterans.
