In this study We examine investment in male function in the context of geography and floral morphology in the plant ring species Euphorbia tithymaloides, which colonized the Caribbean from Mexico/Guatemala along two fronts that roughly correspond to the Greater and the Lesser Antilles and that meet in the vicinity of the Anegada Passage.

Our results on investment in pollen relative to ovule production are consistent with E. tithymaloides relying on pollinators for reproduction across its range in the Caribbean. We document a geographic pattern of reduction in anther number and pollen:ovule ratios in populations of E. tithymaloides towards the Anegada Passage, which is consistent with a possible geographic transition towards a mixed breeding system where selfing has increased importance in this area where its two expansion fronts meet. The reduction in pollen production is correlated with inflorescence morphology, which converges to relatively shorter cyathia towards the Anegada Passage along both fronts. We discuss alternative scenarios and potential drivers of the patterns we document.

The linked datasets consist of:

1. Georeferenced pollen counts of Euphorbia tithymaloides.

2. Georeferenced measurements of cyathia of Euphorbia tithymaloides.

3. Images from where cyathial measurements were derived.

The cyathia used for this study were collected in the field (fixed in ethanol 70% or FAA) or sampled from herbarium specimens.

POLLEN COUNTS

• Individual cyathia were dissected to visualize the anthers. The number and state of anthers was recorded. One undehisced anther per cyathium was selected at random and transferred to an Eppendorf tube containing 10mL of distilled water, where it was kept at 4°C until further processing. Anthers were macerated within the tube using a micro pestle, and the contents transferred to a Marienfeld Neubauer counting chamber (Paul Marienfeld GmbH & Co. KG, Lauda-Ko ̈nigshofen, Germany) to be examined under the microscope (Velab Prime VE-T50). The number of pollen grains in each quadrant was recorded separately and then added together to get the number of pollen grains per anther for a given sample. To ensure repro- ducibility, the counting exercise was repeated three times for each anther and averaged.
• The number of pollen grains per cyathium was obtained by multiplying the number of pollen grains per anther by the number of anthers per cyathium.
• Ovule production in E. tithymaloides is fixed to three per cyathium, so that P:O ratios are a direct reflection of pollen production per cyathium.

CYATHIUM MEASUREMENTS

• We contribute new cyathial measurements (length, height) for 56 cyathia representing nine populations (min 5 1; mean 5 6.22; max 5 12 cyathia per population). Images of cyathia were taken in the field with a scale, and measured in Geogebra v. 5 (http:// www.geogebra.org). These new data were added to data from a previous study (Cacho et al. 2019) and together are part of an ongoing morphometric data collection for this system.

OCURRENCE DATA

• Our own collections and images were geo-referenced using GPS; for herbarium samples and images where GPS coordinates not available, georeferencing was performed using Google Earth tools.

FILE: cyathia_measurements_20191221.csv

FILE: DAT_polenWcyathiaNcoordsPOP_STDvals_20200104_english.csv