Print Page | Close Window

Individual Adder ID

Printed From: Reptiles and Amphibians of the UK
Category: Herpetofauna Native to the UK
Forum Name: Adder
Forum Description: Forum for all issues concerning Vipera berus
URL: http://www.herpetofauna.co.uk/forum/forum_posts.asp?TID=1365
Printed Date: 28 Mar 2024 at 2:26pm
Software Version: Web Wiz Forums 11.06 - http://www.webwizforums.com


Topic: Individual Adder ID
Posted By: administrator
Subject: Individual Adder ID
Date Posted: 08 Mar 2006 at 12:04pm

Hi All,

Carrying on from another thread where I requested information regarding techniques for adder photo ID I've found that

Benson, P.A. (1999): Identifying individual adders, Vipera berus, within an isolated colony in east Yorkshire. - British Herpetologcial Society Bull. 67:21-27

is published on the web at:

http://www.allerthorpe.org.uk/pages/identifyingadders3.htm - http://www.allerthorpe.org.uk/pages/identifyingadders3.htm

I'm also looking for a copy of the following paper if anyone can help out:

Sheldon, S & Bradley, C. (1989): Identification of Individual Adders (Vipera Berus) by their Head Markings. British Journal of Herpetology Vol. 1 pp. 392-396.

Ta




Replies:
Posted By: herpetologic2
Date Posted: 08 Mar 2006 at 5:25pm

 

Hi Gemma

I have a copy of that paper - it looks at the dead markings and divides them into three - apex of zig zag, H or V and eye spots -

When you get good close up photos of the adders head I tend to look at the individual scale markings - you can then match these up in different photos - this allows for positive id with photographs of animals before sloughing, or light conditions etc also you would hopefully be able to id skins from this method

Skins if collected can be cut along the ventral surface so that they can be pressed inside out on a white background - I remember reading about this in an English Research report on reptile survey methods

Here are two examples of male adder skins (you can sex the skin by the number of sub caudal scales, if it is complete

 



Posted By: administrator
Date Posted: 08 Mar 2006 at 6:00pm

Hi Jon,

I'm beginning to get the idea, I've been looking more closely at some of my images and can see now how the head scalation can be mapped from a number of images and scale groups counted, see below:

It has come as a bit of revelation just how much variation and individuality there is in the scale patterns, this along with the head markings is getting towards a 'finger print' and can also see from your images just how easy it will be to identify a known animal from a slough. Tony mentioned this a while ago, though it hadn't really dawned on me how to go about it all.



Posted By: administrator
Date Posted: 09 Mar 2006 at 3:41am

Jon,

For subcaudals Smith gives 32 to 40 males, 27 to 35 in females, (with annoying overlap) is this about right, or have later workers improved on this to your knowledge?



Posted By: herpetologic2
Date Posted: 09 Mar 2006 at 3:56am

 

Another thing that people may want to look at is the actual scale arrangements on the head - to add to the markings - all these pieces can be used to determine an individual - so when you dont have all the pieces - say the animal hasnt sloughed yet - markings would be harders to determine but the scale pattern may help determine the individual. NB - you can use scale patterns to id female lizards!

I also meant to say head markings in my last post. I am putting together all the individual adders head markings I have photographed at a study site from the last three years - I know that I have seen the some individuals over successive years - I have also put together all the grass snakes - I have 40 animals here - it will take a bit of time i think

Jon

 



Posted By: Vicar
Date Posted: 09 Mar 2006 at 4:12am

Gemma,

I definitely find the variation in head scales an excellent discriminating factor, but, as you get a large collection of head scales, it can be a real pain comparing each one to see if you have a match.

One solution which works for me, is to compile an ID file which has various images of the individual looking at different aspects. In addition I highlight the particularly unique features of the head scale pattern. How you choose these features is, at the moment, subjective, its what draws my eye as unusual.

Here's an example: Several obvious improvements can be made, but it does work :P

 



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 09 Mar 2006 at 4:26am

Benson coded individuals by counting scale groups to form part of the code for each animal.

the 5331052 bit relates to the number of scales in groups 1,2,3,4,5,6,7, respectively as shown above. I'm not convinced yet by my labelling and identification of the groups, but practice makes perfect and all that. Though some animals will have the same code for scalation pattern, it should cut down the number of images you have to check to find if you have encountered the animal before.

I've known some individuals for years, I know for sure I see them quite regularly, it is now getting to the point though where I'm visiting more sites and the number of visits is more restricted due to work etc. So I'm really aiming to get an ID system that can be relied on from a single encounter.



Posted By: administrator
Date Posted: 09 Mar 2006 at 4:34am

Hi Steve, posted at same time there, this looks a great idea, perhaps incorporating the above system of scale group counts, some of the subjectivity can be taken out? Though I think there is value in having some subjective things that just catch the eye, particularly when trying to ID in the field.

What I'm hoping is that the scalation coding and other coding (see the sand lizard thread) will help pull a small selection of record cards quickly from a large set. Then it will be easier to compare just a small group of ID cards.

Biggie, is a whopper eh! I can see it just from the head shape



Posted By: herpetologic2
Date Posted: 09 Mar 2006 at 5:47am

 

Brillant

I am going to look into this technique for my photos

Jon



Posted By: Vicar
Date Posted: 09 Mar 2006 at 7:47am

I like the Benson approach, but am having just a little trouble applying it. How would you codify 'Biggie' above ? I'd like to compare results :P

 

(BTW, I might be having a 'blonde' moment, but is there a group-5 on your example?) 



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 09 Mar 2006 at 8:26am

Hi Steve, you know I was just contemplating this, it would be really useful to exchange some examples and try and work out the labelling.

Follow the link in my first post if you haven't yet seen Bensons' examples, which will shed some light. 

I think I'm going to need to do a bit research regarding the naming of head scales, before I can group them. I'm sure I've labelled my diagram wrongly, the top "2" looks OK as an internasal, the other "2"s look like they should have been labelled as prefrontals and therefore should have been "3"s - group 6 I haven't a clue what you would classify them as.

The only reference I have at the moment is Smith, The British Reptiles and Amphibians, where scalation of a grass snake is given, so I'm struggling a bit at the moment! If anyone has any other sources for preferably V.b head scale nomenclature it would be of great help.

Just for fun I make Biggie

7281?(I'm not sure yet which is group 5)92

On the bright side I did realise whilst sitting at the dentist this morning there is a catch all, if the groups are consistently labelled the odd labelling mistake isn't the end of the world as you could still test for n, total number of scales for labelled groups, i.e. from my reckoning Biggie would have n = 7 + 2 + 8 + 1 + 9 + 2 = 29 though the matching set will probably be larger particularly for numbers close to the mean, it does allow for some further checking if there no immediate match.

I'm off to scrutinise Benson's examples a bit more, though the web images are a bit blurry. If anyone has the original paper to hand.. beg beg beg beg



Posted By: Vicar
Date Posted: 09 Mar 2006 at 8:40am

ROFL!

I made it 6441082, and I have some head shots here, where I don't even know where to start!...OK  I'm researching too :P



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: Vicar
Date Posted: 09 Mar 2006 at 8:47am

this webPage looks familiar , and could help some.

http://www.onewildworld.co.uk/reptiles/viperaberus.htm - http://www.onewildworld.co.uk/reptiles/viperaberus.htm



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 09 Mar 2006 at 9:12am

Well it wasn't a bad start at least we agreed on the number of parietal scales ,

Thanks for that link, I had a sneaking suspicion Lee's site might be holding some gems and 'twas my next port of call.



Posted By: administrator
Date Posted: 09 Mar 2006 at 10:37am

Steve, I've taken a look at Benson's paper again. It appears to give two examples for the method of coding the scale groups, though it also appears one of these is spurious.

Taking the one good example I've produced the following Figure, (named After Benson as though it is a complete rework of the given information I don't want to run into a copyright issue)

My next step is to recode my original example following this, then I'll have another go at 'Biggie'. At least I now know why I was confused by group 5 and group 6.

 

It seems at this stage that it is more important to define the numbered groups by their ease of grouping from a dorsal head photograph, than worry too much about nomenclature for individual shields or scales.

I think if we could produce a dozen or so standard reference figures like the one above it will all start to get a lot easier.

 



Posted By: administrator
Date Posted: 09 Mar 2006 at 11:37am

Another, this time adapted from a third, graphical explanation, of the method from Benson's paper.



Posted By: administrator
Date Posted: 09 Mar 2006 at 12:06pm

OK my original example reworked.. with some different numbers

This one threw me a bit, I had a lot of trouble deciding which scales belonged in group 3 and group 2. I used the logic that the ones marked in red (group 3) qualified by either

a) They were directly adjacent to the frontal (group4), or

b) They were directly adjacent to group 5 or group 6

If anyone knows Paul Benson and could direct him towards this thread. I think we would all be forever grateful.

 



Posted By: administrator
Date Posted: 09 Mar 2006 at 1:34pm

And here is Biggie,

To the above rules I've added (I almost typed addered then, time to stop working me thinks) the further rule:

c) To qualify for group 5 and group 6 the scale must be directly adjacent to either the left or right supraocular

This gives following each of the three rules stated so far:

 

Of course this only my interpretation of the method, but by forming some guiding rules I've found it is easier to form the groups without too much subjectivity.



Posted By: Vicar
Date Posted: 09 Mar 2006 at 1:54pm

I was going to suggest that maybe the frontal count on 'Biggie' was 2 ? Just been having this conversation here, I have seen examples where there is no obvious frontal at all, just split into many smaller scales ?

Are you thinking its OK to have blanks ?, or do we need to account for all the major head-scales ?

Doesn't really matter what was originally intended, although it would be better to have universal agreement, but spot on Gemma, we do need a good set of definitions for each group.

Where's David Bird when you need him ? :P



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 09 Mar 2006 at 2:36pm

The blanks were my first stumbling block, at first it didn't seem rigorous to leave blanks.

The trouble is it starts to get complicated to form the rules if all scales are to be counted. It is also easier and quicker not to have to classify all the scales that are present (don't forget I've an engineering background and am inherently lazy) Also idealy for the code we want the scale group counts to be no larger than 9 else more problems start to occur.

The primary concern is to be able to form the groups consistently i.e. to have a foolproof method that different workers can reproduce over and over and that can be easily learnt.

I really like the idea that someone could go to a site I visit regularly (which happens) then call me up and say, "I had a 6231442 today, do you have any record match". I then find 3 possible matches and we compare photographs for a final ID.. that would be really great!

Take for example one with dozens of broken up scales, it won't happen that often. So if you code using the rules only a few of the scales need to be looked at saving a lot of work. OK there will be lots of blanks, but at the end of the day we just want to grab some likely fits from a large record set, the final ID will be scrutiny of photographs, scale group counts, head patterns and also the way the scale groups are arranged and the way individual scales are marked. All will be added together for a final ID.

I'm considering for example dropping rule (b). Is it needed? OK you get more blanks, but rule (a) alone would still give reproducible results, i.e. all workers would have recorded 3 scales in group 3 for my own example based on rule (a).

Regarding 2 frontals for Biggie, my logic says that frontals are most often one large scale, maybe 2 or 3.. the small fragmented scale is an anomaly that is best ignored when forming the scale groups but is very useful if highlighted on the ID card as an identifying feature in the way that you have Steve. (I hope that is what you ment and I've not missed a joint line in the frontal)

I have for one of my working codes P! which simply draws my attention to the parietal scales of the individual as they have distinctive markings. Biggie could become F!6251242 for example?

We need more examples come on all those watching this thread, your dying to give it a go to, post up your results



Posted By: herpetologic2
Date Posted: 09 Mar 2006 at 6:07pm

I did this without seeing the previous posts - look at the shape of scale 4!

This one is hard to see the numbers but at least I am having a go eh

Jon



Posted By: herpetologic2
Date Posted: 09 Mar 2006 at 6:19pm

My graphic programme is rather limited - as I only have the scanner programme at the moment - but this is a male adder from one of my sites in Hants

Jon



Posted By: administrator
Date Posted: 09 Mar 2006 at 6:38pm

Great stuff Jon, take a look at this one that came from my Kent records..

 

Groups 5 and 6 absent.

I'm fairly sure now that really small scales need to be ignored. Hence  I've not scored a 1 for group 5 on the ID above.

Benson mentions being able to spot the scales from 1-2 meters distance for field ID. Something we need to think about also is that not everyone has a digital camera so they might want to use a quick field sketch to form the code. So I'm leaving very small scales as blanks.

Keep at it Jon, the more examples we get the easier it will be to get a working system that we can all use.



Posted By: administrator
Date Posted: 09 Mar 2006 at 6:56pm

This is a totally irrelevant post of a piccy of F_vb_wbp_ID_03_08_2005_6211002 just because she is pretty.

It does show though that you wouldn't see the little scale in the field and that groups 5 and 6 are therefore for our purposes absent.



Posted By: administrator
Date Posted: 10 Mar 2006 at 12:07pm
appologies to anyone who was trying to follow the codes, I put up a few errors, this should be corrected now.


Posted By: herpetologic2
Date Posted: 10 Mar 2006 at 3:09pm

 

Hi Gemma

heres another one

Code m-Vb-EB-2005-ID6321222



Posted By: administrator
Date Posted: 10 Mar 2006 at 6:43pm

A few more, first one is a repeat of the last example:

 

 

 

mug shots for last three examples, be good if someone else could have a go at these and see if they get the same code?



Posted By: administrator
Date Posted: 11 Mar 2006 at 7:50am

I've got my first match using the code  

I had a suspicion during a mitigation I worked on last year that it was possible for adder to return into the works area after they had been moved - so called for additional exclusion fencing at the site.

You'll notice from the examples above that head markings are not particularly variable for this population, so I had not had much success trying to use them to decide whether or not I had rescued any of the individuals twice.

Now I know. The first positive match using scale group counts pulled from dozens of photographs. Once the match came up I was able to check other photographs and confirm it was the same animal. I'll be adopting this system from now on in all my adder work. Paul Benson's work has really made a huge difference to how I can now approach keeping my records.

 

 

I'm really pleased 5241332 was moved out to a safe area before works began, even if she had other ideas

 

 

 



Posted By: Vicar
Date Posted: 11 Mar 2006 at 10:33am

Gemma,

That's excellent ! I particularly like the colour-coding to emphasise the match.

When you get a good set of examples, I'd be very interested to know how many code-matches you get in practice (that are in fact different individuals).

It would be nice to do some stats and to determine 'how unique' to an individual this code is.

BTW, The Sheldon/Bradley paper looks at patternation and does not appear to address scale-counts, which is fair enough as it was a field system. We can probably do better now we are in the age of the affordable digicam.



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: Vicar
Date Posted: 11 Mar 2006 at 12:02pm

Just been thinking again (sorry!)...

A numerical code system obviously rocks, as it facilitates database searches, something very hard to do with graphical information.

But as far as field work goes, occasionally, if you suspect an abnormality, its sometimes too late to check, once you are back at base.

I'm pretty sure I couldn't do a head-scale-group count on every adder I see at a particular site I may be studying. So maybe some graphical information, which could be carried in a folder would be useful to me.

Reading through the Sheldon/Bradley paper again, they suggest that Dorsal apex, eyeline and 'V' are all good discriminators individually, and when all three aspects are taken into consideration together, give a unique result.

I think I would like to develop a diagram system, which would highlight both patternation and head-scale groups, So I could use either.

I'm no artist....but this is along the lines I have been considering...Thoughts ?

Rightmost pic is end result, other images show how I got there, and for comparison.



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 11 Mar 2006 at 12:44pm

Steve, have you got Photoshop? I'm creating the images as Photoshop psd files, they contain the head markings, the original photographic image, then the scale groups, thus I can generate any combination by switching the layers on and off or changing the opacity settings for any layer. This lets me have anything from a full composite, to just the mug shot, just the markings, or just the scale groups.

 

 

 

 



Posted By: Vicar
Date Posted: 11 Mar 2006 at 5:44pm

Here's an alternative layout:

Aim is to include detail of head scales, dorsal marking apex, 'V' and eyelines. Oh and sex, can use black markings for male, brown for female.

I prefer working with diagrams, rather than photos, as light and pose can make the same snake look somewhat different in two photos, somehow the eye compensates for this when using a diagram.  Thoughts ?



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: administrator
Date Posted: 12 Mar 2006 at 3:58am

Steve excellent , that is a much clearer layout, the first version seemed to have too much information (probably just colour information, but it was stretching my IQ a bit to look at it)

This new layout is really clear and I would imagine very 'field friendly' 

I was wondering how much of this work is a familiarity exercise? Not a bad thing if it aids field ID, just an observation.

What I'm saying is that spending the time doing diagrams or scale counts allow the observer to absorb a lot of information about the individual animals they are studying, which may well lead to instant recognition from a fleeting field encounter... very usefull indeedy, particularly on a warm day when they seem  to rely on rocket propulsion to get around



Posted By: herpetologic2
Date Posted: 12 Mar 2006 at 9:33am

 

Using individual scales and their colouration can help to determine matches very quickly - plus you have a system of ID which is determined in the field - head scales using binoculars - you dont have to capture the animal to do the ID once you have managed to get a good head shot etc.

Also I have a yearling here which I will try and get some photographs of the head to see if the system would work on neonates?

Jon



Posted By: Vicar
Date Posted: 12 Mar 2006 at 10:20am

Jon,

According to the Sheldon/Bradley paper, the basic system works very well with neonates. From memory...out of 80? specimens studied (over 8 years?), only one developed an (one) extra scale in adulthood. I can check exact numbers if needed, but those are the right orders of magnitude.



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: Vicar
Date Posted: 14 Mar 2006 at 12:10pm

Here's an interesting issue...

I've been drawing up some diagrams from digital photos of a small study site, and came across these two, drawn up independently from separate pictures (taken a few days apart).

Now, I'm pretty certain these are the same snake, but my interpretation of the scale pattern, from digi photo, is slightly different in each case, which would lead to a slightly different Benson code (Group 6? in this case).

If I can muck it up sat at home in my lounge, you bet I'll make a sloppy mess of it using binoculars in the field :P

Anyhow, this has convinced me that I need diagrams in addition to codification, and am now back to looking at image processing to derive a unique ID (or a range of numbers close enough to be the same individual).

Anybody had similar issues using codes ?



-------------
Steve Langham - Chairman     mailto:steve@surrey-arg.org.uk">
Surrey Amphibian & Reptile Group


Posted By: herpetologic2
Date Posted: 14 Mar 2006 at 1:25pm

 

I would say that photos are needed in this case - I would say that it is the same snake as several patterns do match -

I make the Benson code as follows

1. 6301012 or 6211012

2. 5301012 or 5211012 - you need to look at the photos to look at that fifth/sixth number 1 scale

It shows that diagrams may reflect the artist's skill in drawing through a computer -

JC



-------------
Report your sightings to the Record Pool http://recordpool.org.uk" rel="nofollow - http://arguk.org/recording


Posted By: administrator
Date Posted: 17 Mar 2006 at 6:16am

I've not found slightly different numbers a problem. If I can't see from a digital image how many scales are in a goup I've coded the group as ! Ignoring very small scales helps.

In practical terms going through a set of say 30 different potential matches, it is easy enough to match 62313!2 with 6231321 as the majority of groups match up, then it is a case of spending a little time looking at other images for the animals to see if it was indeed the same one.

The Benson code is never going to give a unique ID for every individual, it doesn't really need to and it simply can't as some animals will have the same code, and sometimes the code produced from a digital image will be slightly wrong depending on the quality of the image.

In practice though it is a very useful system. You take a stored head shot, you spend a few minutes grouping the scales and you file the image in your site folder.

It is when you file the image that you think.. hmmm, that code looks a bit like the third one I did a month ago, you then compare the two head shots, your scale pattern diagrams and any other associated images from the two separate encounters and decide if they represent the same animal. A far better method than having to check all 30 records for a potential match




Print Page | Close Window

Forum Software by Web Wiz Forums® version 11.06 - http://www.webwizforums.com
Copyright ©2001-2016 Web Wiz Ltd. - https://www.webwiz.co.uk