Bird-mediated dispersal is presumed to be important in the dissemination of many different types of organisms, but concrete evidence remains scarce. This is especially true for biota producing microscopic propagules. Tree-dwelling birds, such as woodpeckers, would seem to represent ideal dispersal vectors for organisms growing on standing tree trunks such as epiphytic lichens and fungi. Here we utilize bird natural history collections as a novel source of data for studying dispersal ecology of plants, fungi, and micro-organisms. We screened freshly preserved specimens of three Finnish woodpecker species for microscopic propagules. Samples were taken from bird feet as well as chest and tail feathers. Propagules were extracted using a sonication-centrifugation protocol and the material obtained was studied using light microscopy. Diverse biological material was recovered from all specimens of all bird species, from all positions sampled. Most abundant categories of discovered biological material included bryophyte fragments, fungal spores and vegetative propagules of lichens. Also freshwater diatoms, bryophyte spores, algal cells, testate amoebae, rotifers, nematodes, pollen, and insect scales were identified. The method developed here is applicable to living specimens as well, making it a versatile tool for further research. Our findings highlight the potential of bird-mediated dispersal for diverse organisms and showcase the use of natural history collections in ecological research.

Bird specimens were obtained from recent accessions to the collections of the Finnish Museum of Natural History (LUOMUS). We sampled altogether 15 specimens of three woodpecker species native to Finland (Dendrocopos major, Dryocopus martius, and Picus canus, five specimens of each species). Bird specimens were selected based on collecting date, cause of death, body condition and post-accession preparation. Only recently acquired specimens that had been removed from their natural setting fast, e.g. due to a window collision or bird ringing accident, were chosen to avoid a risk of post-mortem contamination. All specimens were in good physical condition and had been quickly frozen after collection, wrapped in clean newspaper, shipped to the museum and then incorporated into the collections, labelled, bagged in airtight plastic bags, and frozen for further study. The birds had not been washed or otherwise treated before sampling. Using the specimen-specific links in the data, the digitized specimen metadata can be accessed.

All bird specimens selected for analysis were processed as follows. First, the specimens were removed from their bags and thawed. This was done to avoid condensation of moisture on the thawing birds. Three different cotton swabs were used to sample the chest feathers, tail feathers and feet of each bird, respectively. For feather sampling, the swab was run proximally multiple times across multiple feathers, and in the tail feathers especially any visible dusty material attached to the underside of the pin of the feather was targeted. Feet were sampled in a circular motion on the underside of the digits, including under the claws. In addition, a single top layer chest feather and a 2 cm fragment from the tip of one tail feather (rectrice 1 or 2) were removed for detailed analysis.

Collected cotton swabs and feathers were suspended in 500 μl of Milli-Q water in sterile 1.5 ml Eppendorf tubes. Tubes were vortexed briefly and placed in a sonication bath for 3 minutes to dislodge biological material from the cotton/feathers. After sonication tubes were again briefly vortexed followed by 5 min centrifugation (13 000 rpm). Feathers/swab tips were then carefully removed from the tubes with forceps, avoiding disruption of the potential pellet. This was followed by 3 min centrifugation (13 000 rpm), after which most of the supernatant was removed. The remaining suspension (10-20 μl) was briefly vortexed, spinned down and inspected and photographed under phase contrast microscopy (Leica Microsystems, Wetzlar, Germany). Biological material was identified as accurately as possible, and the abundances of main categories of propagules encountered were roughly estimated. As direct counts were impossible due to the large amount of material, the amount of the main three categories were semi-quantitatively estimated, ranking from no material to few (1-5) units to multiple (5+) units. In the dataset, no material is coded with 0, few units as + and multiple units as ++.

Digitized specimen metadata can be accessed using using the specimen-specific links in the data. As direct counts were impossible due to the large amount of material, the amount of the main three categories were semi-quantitatively estimated, ranking from no material to few (1-5) units to multiple (5+) units. In the dataset, no material is coded with 0, few units as + and multiple units as ++.