Picking Pollen

So, you want to 3D-print some pollen models? Great! That's so good to hear. Here you can find a few considerations on how to start the process off: this is how you can identify what models to use and locate somewhere to download them from.

What to scan?

Let's start by assuming you know what kind of pollen you want to print – let's say it's pollen from a European daisy. (Many of the principles in this section also hold for the kinds of non-pollen samples discussed on the 3D Not-Pollen page.)

The very first thing to do is to find out your plant's scientific name (Bellis perennis, in this case). Two-part (binomial) scientific names in Latin are useful for searching scientific databases as a) that's often how they're structured, and b) scientific names are much less variable between countries/languages, etc. than common names are. It's worth having both on hand!

Where to look?

Next, check available sources of 3D pollen models. One or more of them might have the exact model you're looking for! At the moment this involves looking in a few different places, though I may try and bring the various model lists together at some point. Here are the online sources of 3D pollen models I'm aware of:

The 3D Pollen Project (here!). To check what's available, the two best places are the table on the Available Files page and the search results on MorphoSource. Model files, derived from confocal laser microscope scans, can be downloaded from MorphoSource in .stl format at 2,500x life size, so a pollen grain 20 µm across would produce a model 50 mm/5 cm across if you didn't rescale it. Videos, and in some cases full-resolution image stacks, are also available. The data are available under a CC-BY-NC licence (free non-commercial use with attribution).
The Cardiff University Bioimaging Hub via NIH 3D Print Exchange. This is a large dataset of microscopy-derived 3D pollen models, similar to the ones in the 3D Pollen Project, though there are slight differences in the microscopy approach, and the Cardiff models include pollen grains' cell contents. You can download files in formats designed for 3D printing, VR and AR, as well as microscopy scan images. The models files are at 1,000x life size, so a 20 µm pollen grain would make a model 20 mm/2 cm across. The data are available under a CC-BY-NC licence (free non-commercial use with attribution).
The Australian 3D Pollen Project. This is a set of 13 models of pollen from Australian plants. It was derived from confocal microscopy, but scan images are not available. The scaling of the model files compared to the samples scanned is not stated, but it appears to be around 1,500x life size (almost certainly somewhere between 1,000 and 2,000x life size). The files are shared under a CC-BY-NC-SA licence (free non-commercial use with attribution, and any derivative works must be shared under the same licence).
University of Illinois at Urbana-Champaign. This is an massive but very niche collection of super-resolution 3D scans of pollen from one particular branch of the bean plant family. If the Amherstieae tribe of the Fabaceae subfamily Detarioideae is your jam, then you are very much in luck! Interestingly, this is also currently the only instance I am aware of where pollen grains from extinct plants have been scanned – Striatopollis, in this case. The database is of scan images rather than 3D surface models, but you can make the conversion fairly easily by following the guidance on the Methods page.
There are some additional collections of 3D pollen models, but they don't appear to be online or downloadable at the moment. The Sketchfab sites of 1manscan and rcgear_japan have some examples of designed (rather than scanned) pollen reconstructions. The University of Maine's BEAST lab has also produced some, though it's harder to track images down (but I did it)! Then there's the original 3D models of New Zealand pollen produced by Kat Holt. This effort pioneered the methods used in the three scanning projects listed above, but sadly the models are not currently available online.

What if it's not there?

Searching through those sources of 3D pollen data, you might find a model from the exact species you want – daisy (Bellis perennis), for example, is included in the Cardiff collection. However, for the overwhelming majority of plants on Earth, that won't be the case. For example, let's consider groundsel (Senecio vulgaris), which isn't in any of the collections above. If you can't find a model for your exact species, try and find a lookalike.

For this, you need to know what your target pollen looks like. A simple online image search can be really useful, but there are some excellent, dedicated resources for pollen images. PalDat might be the best and most comprehensive, but there's also the Global Pollen Project, the Australasian Pollen and Spore Atlas, and the various ID keys of RCPol, among others. These resources can also help you figure out when you've found a good match. This is what groundsel pollen looks like.

In pollen as in people, you might find a good lookalike in a close relative. Pollen generally tends not to vary much at species level. Instead, closely related plants tend to have very similar – and frequently indistinguishable – pollen. That means that pollen from a different species in the same genus (same first part of the Latin name) might well be identical to the one you want – especially once it's been 3D-printed. You can see this by comparing, for example, sessile (Quercus petraea) and pedunculate (Quercus robur) oak pollen.

The relationship between pollen similarity and species relatedness isn't constant across the plant tree of life – passionflower pollen varies quite a lot within the Passiflora genus, whereas the 12,000-odd members of the grass family (Poaceae) all produce virtually identical pollen. While there's no other groundsel or Senecio pollen in the current 3D pollen datasets, there are several members of the daisy family, Asteraceae, which all have the same type of apertures, same overall shape, and same sharp spines on their walls. Oxeye daisy (Leucanthemum vulgare) pollen, for example, might be a good alternative, and that can be downloaded here.

What comes next?

These steps should give you the best chance to find a suitable pollen model file – whether it's the exact one you want or a lookalike. There are still gaps in the databases, of course – types of pollen for which no 3D files exist yet – but we're trying to fix that. Our aim is that by late 2025 the 3D Pollen Project dataset will cover a substantial part of the world's pollen diversity – from across the plant tree of life, the range of pollen shapes and structures, and the world's ecosystems. To help build the most useful and representative collection possible, we're asking about what types of pollen are important for different people around the world. So, please fill in the survey to tell us what pollen would be useful for you, and help make it easier for everyone to find the pollen models they need!

Picking Pollen

What to scan?

​