Identifying things

One of the most fundamental things we do is to work out what things are – it’s a very human interest. We like putting a name on something and then using that to learn more. 

Books

This is the traditional way, and how I started as a kid – using Collins field guides (by Chinery) to the insects of Britain and Northern / Western Europe. Then for moths I started with South’s Moths of the British Isles (a lovely old book but with terrible plates), and then moved on to Skinner’s guide to moths, which was a revelation and probably did more to get people buying moth traps. Then came the newer insect field guides by Paul Brock and the moth guide by Waring, Townsend and Lewington, which raised the guides to another level and fixed a lot of the issues with earlier books.

It does take a bit of time and determination to get into using guide books and they can be a bit expensive for a child to just buy. While the moth books are usually fairly comprehensive, insect field guides can only show a tiny selection of the >25,000 UK species and so they come with big limitations. This is made more confusing because they will often show a pitcure with a species name written under it and in the small print say “similar species exist”, where they should have just given it a generic name and said “many identical species exist” or “cannot be identified to species without microscopic examination”. There are thousands of incorrect records in circulation for “Sarcophaga carnaria” or “Pollenia rudis“, which might be correct, but probably aren’t because they need microscopic examination to confirm the ID. 

Apps

iNaturalist
iRecord
ObsIdentify
Merlin Bird ID

The rise of the internet and mobile computing over the last 30 years has brought with it a raft of new apps that promise to identify whatever you can take a photo of. The seem to work like magic, but are they? Can you trust them?

First you need to understand what they are doing and how they have trained themselves. In most cases they are using a simplified version of Artificial Intelligence called Image Recognition, which seeks to isolate the subject in an image and find other images that look similar and then use the name that the photo has next to it. It doesn’t know how to identify anything – it can’t work out which features in the image are important and which are just variation, but it relies on testing thousands of photos. It also relies on those photos having the correct name associated with them, and these assumptions reveal the weakness of the apps. If many species look virtually identical from photos then it will often just pick the commonest name in the database for all of them; and if the photos are wrongly labelled then it will just repeat the error. 

Social media

One of the best ways to confirm your ID is simply to join one of the very many identification groups and ask the experts. In that way you can get a really authoratative opinion and often some information about whatever it is that can make the ID more interesting. But the groups are also full of well-meaning people who will guess the ID based on their limited knowledge so you have to be a bit careful who you believe. 

Watch out for comments like “Looks like X” … or “Compare to Y” – they seem like an ID but are they? If you are starting from a position of ignorance and you don’t know what the thing is or how complex the group is then if someone suggests a name then you’re very likely to go off and Google it, and if it looks convincingly like your thing then you’re very likely to think it is … but that’s called Confirmation Bias – you are likely to choose something that confirms what you’ve been told, or your own biases. But if you don’t know that there are dozens of other near identical species then you might be wrong. Also, just like with the apps, you’re not equiped with the knowledge to know what to look for on the species you’re trying to identify. 

I use apps myself, to check the moths that I occasionally trap i nthe garden, and it’s a great way to process dozens of photos and get IDs rapidly. The apps have also spotted rare species amongst the catch that I had missed and so they have really helped me to learn more. But I am also quite experienced in identification of insects so I pretty much always know when the app has chosen the wrong thing – I know the species that it can’t ID and I know when to downgrade an ID to genus or a species complex. I don’t blindly pick the first ID the app gives me if it feels wrong. 

Conclusions

The main take away is that there is no magic behind getting a good ID – you can use apps to take shortcuts but at the end of the day you do need to check things with experts. Whether you use a book or an app I’d always double check with an expert in social media or in your local natural history society or museum. 

2 Replies to “Identifying things”

  1. I just read this entry on identification. How true. I recently bought a book on solitary wasps which, while it promised much, delivered very little. Actually I’m beginning to suspect the thing I found was a wasp-like sawfly, and don’t even get me started on identifying them. The easy ones are easy, it’s the others………

  2. Yes, I think it’s fair to say that I learned most by going on free ID workshops in a lot of groups and then chatting to other experts on social media about which keys they used. Also it helps to get out in the field with good people and I’ve always found moost entomologists to be very friendly and willing to show you the ropes.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.