Saturday, 30 January 2016

Field Marks - Bold and Bland Remiges and Retrices

Recent rare gull activity in Ireland and Britain has prompted a closer look at this question.  While trying to ascertain the grey scale value of the pale wing-tips of Ireland's first putative Glaucous-winged Gull found by Fionn Moore (HERE) it was quite evident that debate continued to rage based on the variable appearance from different sets of photographs. In any sample of online images there didn't seem to be any consistency in the tonal levels.  Compare for instance THESE and THESE images.  Not surprisingly, the question over the tone of this bird's primary feathers has remained in flux.

Then along came a striking Thayer's-type gull from Aberdeenshire, Scotland.  The observer Chris Gibbins was in an unenviable position of having to leave the bird before clinching all it's features to his satisfaction.   Luckily others were able to capture a suite of images to assist with the bird's identification.  Later, as the first flight images of the bird emerged they appeared to show a dark contrasting secondary bar and everyone seemed happy the bird was a classic Thayer's (see finder's account HERE).  Then as more images emerged it appeared the secondaries were more concolorous with the wing (see HERE).  Once again lighting seemed to be playing with the tones of remiges and retrices.

Structure Of the Remiges and Retrices
Flight feathers including the remiges (wing feathers) and retrices (tail feathers) are more stiff than body contour feathers but still flexible enough to assist birds with masterfully subtle and graceful controlled flight.  They play a vital role in flight by both capturing and directing airflow to give a bird lift and accurate, agile manoeuvrability in the air.  While the feathers overlap and often work in relative unison, similar to the control surfaces of an aircraft, during flight they occasionally part or may have gaps due to moult or damage.  This can alter a bird's flight characteristics and efficiency.  Wing strokes are dynamic and variable depending on such factors as air resistance and lift requirements and the type of manoeuvre a bird is undertaking.  Tracts and individual feathers therefore at any given moment may be straight and level, bowed laterally or longitudinally, or even twisted about the feather centre.  The same applies to the feathers of the tail.  We often think of the tail as a straight, level structure.  But, in flight the tail may be just as dynamic as the wing.  Very often the tail is somewhat bowl-shaped for stable level flight, but may be splayed and/or twisted during complex aerobatics.

Here are a series of pencil sketches based on a number of published photographs of the Aberdeenshire Thayer's/Kumlien's Type Gull including images by Hywel Maggs, Jonn Nadin and Stewart Whooley.

The Bold Versus The Bland 
I have spent a great deal of time on the blog analysing the distinction between bold and bland field marks.  Bland field marks are far more challenging to work with.  For more please see the posting HERE.

Lighting Factors
Allied to the complexity of angles with which birds can hold their flight feathers, light and shade on the retrices and remiges can pose an extremely complex challenge.  Angles play a major role in how tones are presented.  As outlined in the posting on Lambert's Cosine Law (HERE) the brightness of any point on the surface of a subject is proportional to the angle of the surface to the incident light direction.  Feathers which are bowed or twisted can have points on them that are angled 180 degrees or more from other points, revealing simultaneously to the camera both the upper and under surfaces of the feather and all angles in between.

Translucency in another oft ignored factor which is particularly relevant to any analysis of flight feathers.  While most feathers of a bird cover underlying body tissue and bones resulting in a surface that is more or less opaque to light, the retrices and remiges stick out from the body.  Light can readily pass through these feathers, and in doing so can influence the appearance of subtle tones on the feathers themselves.  Great care should be taken to establish the qualities and angle of the light.  The translucency of a feather is directed related to it's pigmentation.  Paler feathers are influenced by light transmission to a far greater extent than dark feathers.  For more on translucency see HERE.

To the outline sketches above I have attempted to overlay the key shadow and lighting detail in the published images.   The shadows are relatively easy to interpret but I cant be so certain of the influence of light transmission on translucent feathers.  It was Stewart Whooley's stunning image (HERE) which sparked the debate about this bird and got me thinking about writing this blog post.  The combination of the fact the primary pattern of the right wing seems so strikingly subdued in this image and the rather bright marginal covert area both lead me to suspect that the primaries and outer secondaries are being lit from below, resulting in a dilution of contrast in these areas.  Note also the primaries of the left wing are well illuminated below (less so in Hywel Maggs's image HERE).

The Role of Flight Feather Tonal Patterns In Identification
For a birding community well versed in subtle bird identification it has been interesting to observe some of the recent debate around issues like the primary tip pattern of the Irish Glaucous-winged Gull and the Scottish Thayer's-type gull.  If you think about it, there aren't too many identification challenges that rely on such a critical analysis of subtle tones in the flight feathers of a bird.  Primary pattern is important in the identification of many adult gulls but generally we are dealing with bold (black and white) patterning and not the subtle contrast created between midtones, or an exact measure of tones along a standard scale.  Perhaps we are somewhat unprepared for the kinds of challenges posed by the Aberdeenshire gull.

To the rough sketches I have overlayed what I think may be the natural pigmentation level of the remiges and retrices of the Aberdeenshire gull.  This is how the gull might have looked under perfect, neutral light.  Of course I cannot be totally certain of this as all the published images have their own lighting characteristics.  This is a matter of trying to piece together a puzzle without all the pieces available.

Top 10 Tips For Judging Tones In Remiges and Retrices
Having been pondering this question for the last couple of weeks I have come up with a list of top tips for assessing subtle tones in remiges and retrices.  There is no 'magic bullet'.  These tips only help to reveal clues to the true nature of a bird's plumage field marks.  As always it is best to have as wide a range of images as possible to work from.

(1) Always assume that lighting is variable across the remiges and retrices and that a proper forensic analysis of tones is fraught with difficulties.
(2) For the best results the light must be diffuse and not too bright.
(3) Be especially conscious of the translucency of flight feathers and the potential for transmitted light to influence the appearance of tones on flight feathers.  Note that translucency can sometimes make feathers appear darker than they actually are - a white feather viewed against a dark background can appear grey under certain light.
(4) Ideally the background should not contrast too strongly with the subject.  Images captured against the sky (or water reflecting sky) can be problematic as they often challenge the dynamic range of the camera.  The result may be an underexposed subject and/or inaccurate contrast between tones on the subject.
(5) Individual tones cannot be assessed in isolation.  As the posts on Greyscales and Gulls revealed (eg. HERE) totally accurate exposure is very difficult to achieve.  Despite this, comparisons can often be made between tones recorded within the same image.  Tones can also occasionally be compared across images provided that the brightness and contrast settings are matched.  But this can be a risky exercise.
(6) Tones within the same image should be judged by comparing surfaces which are totally parallel to one another, and therefore lit in the same way.
(7) Always be mindful of where shadows are falling, especially when trying to compare tones on different parts of the subject.  Remember a shadow can be created by an object outside the frame of the image.
(8) One should never attempt to compare tones in different parts of an image by eye alone.  Human eyesight is not good at global tonal comparisons (see for example my posting on brightness illusions HERE).  I would recommend the use of image editing software to make objective tonal comparisons.  The simplest method is to use something like MS Paint to grab a sample patch and drag it across an image to directly compare alongside another patch.  Human eyesight is very good at making local tonal comparisons.  For a more objective analysis Color Quantizer is an excellent free software programme which I have found very useful for mapping the tonal distribution across an image.
(9) It often helps to compare surfaces which are more or less parallel to the camera plane.  But be especially mindful of translucency, particularly where lighter-pigmented feathers are concerned.
(10) Working in greyscale often helps.  Note the human eye is more sensitive to green than it is to red or blue.  People also have individual varying sensitivities to different colours.  If the colour hue or saturation isn't particularly relevant to your analysis it makes sense to transform the image to greyscale (luminance alone).  Greyscale reduces the comparison of tones to one of tonal luminance (more accurately termed luminosity).  For an explanation of the important distinction between luminance and luminosity see HERE.

The Scottish Thayer's/Kumlien's Gull
I think the published images of this intriguing gull highlight just how potentially misleading initial impressions can be.

Here I have combined my analysis of lighting in various images with my analysis of the overall pigmentation by overlapping the two sketches.  The results seem to match the available online images quite well.  Whether 100% accurate or not the purpose of the exercise is to highlight the complex interplay between variable structure, variable lighting and subtle pigmentation.

The Irish Glaucous-Winged Gull
While the debate around the exact tonal level of the primary-tip pattern of the Irish Glaucous-winged Gull may continue I think it is worth putting my 10 Tips to the test.  I think the analysis below stands up to scrutiny.  The lighting at the time the image was taken was dull and diffuse as the morning was overcast.  Though the image is taken against water the backdrop is not overly contrasting or bright.  There is a rich range of tones in the image which strongly suggests the exposure control was quite good.  The grey scale method relies on a tonal comparison either with a grey card or between tones within the same image.  In this case the comparison is with the grey of the wing.  The wing is reasonably in parallel with the camera plane and shading is minimal.  The tonal comparisons were made by directly sampling different regions of the primary tips and directly comparing with a standard grey scale.  For more see HERE.

Tuesday, 19 January 2016

Field Marks - Grey Scales and Gulls (Part 3)

Vega Gull, Duncannon, Co. Wexford, 10th January, 2016 (c) Killian Mullarney.

News Flash!  Global Upsurge In Interest In Grey Scales
Doubtless, the recent rise in interest in the blog was sparked more by a certain interesting rare bird than an interest in obscure image quality and analysis tools!  Perhaps I should add more rarity spice to my postings from here on.  In any case hopefully this topic has struck a chord.

In the earlier posting I promised to provide an update on the comments received from North America regarding the Irish Glaucous-winged Gull.  Here is a summary.  While the comments certainly point more towards Glaucous-winged Gull parentage there has been a lot of concern from those familiar with the hybrid zone of the North American Pacific Northwest suggesting the bird might be an American Herring Gull X Glaucous-winged Gull hybrid or back cross.  This is not just on the basis of the darkness of the primary tips but also due to the bird's relatively slight bill, among other subtle points.  Having said that the Irish bird may be a perfectly good match for many Glaucous-winged Gulls wintering in Asia.  Few North American commentators would claim a thorough knowledge of the birds breeding to the far northwest of the breeding range and wintering on the other side of the Pacific.

What I have found perhaps most intriguing from the discussion is the apparent disparity between the quoted grey scale range for the primary tips of Glaucous-winged Gull (eg. 6-8 per both Howell & Dunn and Olsen & Larsson) and those, clearly darker-tipped birds wintering in East Asia.  There is a suggestion that darker wing-tipped birds may represent an older hybrid swarm that is now stable.  This in turn might account for their apparent exclusion from the quoted grey scales in the literature.  But that is just speculation on my part - I have not been able to establish how the range 6-8 was first established and whether it is representative of the whole breeding range of GWG or just a sample, 'pure' population say from Homer, Alaska.

Hopefully the grey scale tool as presented here will provide at least a means to objectively analyse grey scales from photo collections and help develop this debate a bit further.

Vega Baby!
No sooner had discussion of the GWG begun to simmer down here, Killian Mullarney revealed he had found, in his home county of Wexford, what just might be a first for the Western Palearctic - a Vega Gull (Larus vegae or L. argentatus vegae or L. smithsonianus vegae depending on what taxonomy you choose to follow).  After two failed attempts to catch up with this marvelous beast Killian has kindly allowed me to carry out some analysis of his own images using the grey scale method.

While the image at the top of this post is beautifully exposed, the problem with the comparative method is it relies on a reference point.  So, in the absence of a grey card or some other reasonably reliably grey scale value it is not really possible for me to confirm the accuracy of the exposure of this image and therefore to gauge the mantle grey scale of this bird from this image.  

Killian also however provided another image showing the Vega Gull accompanied by a group of Herring Gulls and this is all that is required to verify the Vega's grey scale value.  Vega Gull is slightly darker than American Herring Gull (L. smithsonianus) and more in line with Northern Herring Gull (L. a. argentatus) here in Europe.  The birds alongside the Vega in this case are all of our local race, L. a. argenteus which are within the same tonal range as AHG (4-5).  By adjusting the image brightness to bring the mantle shade of the HGs in line with this on the digital grey scale its then just a matter of sampling and reading the VG's mantle shade.  And the Vega is right where it should be (7-8).  QED.

Wednesday, 6 January 2016

Field Marks - Grey Scales and Gulls (Part 2)

In the posting Grey Scales and Gulls (Part 1) I found an effective way to replicate the famous Kodak Grey Scale in our digital space (sRGB colour space).  The recent exciting discovery of what appears to be Ireland's first Glaucous-winged Gull (Larus glaucescens) in Co. Cork afforded an opportunity to test out this process in the field.  

I had two methods in mind.
(1) Calibrate scene lighting and exposure using a grey card (the preferable method)
(2) Hope to capture the GWG in the company of another gull species and benchmark the GWG's grey scale against a known standard (less ideal, as potentially less exact).

I arrived at Castletownbere Harbour before dawn on January 3rd.  Together with many others, and in great anticipation we began a search for the bird, found the previous evening by Fionn Moore.  After a couple of fruitless hours searching, and just as hopes were beginning to fade, news emerged of the bird's rediscovery at the main pier...and it was coming to bread!

Glaucous-winged Gull, potentially Ireland's 1st, Castletownbere, Co. Cork, 03rd January 2016.

Standard references (eg. Table 3 of Howell and Dunn, Gulls of the America's, 2007) place GWG in the Grey Scale range of 5 to 6.  It is slightly darker-mantled than Herring Gull (both European L. a. argenteus or American L. smithsonianus) and, certainly initial impressions in the field, in direct comparison with our HGs this bird looked right where it needed to be in terms of it's greyness!

Despite the excitement of seeing such a superb bird at close quarters I set up the grey card and tried to calibrate lighting and exposure in order to calibrate the bird's mantle to my sRGB grey scale.  Though I have managed this technique under less stressful circumstances in the past, unfortunately in this case the results were less than perfect.  In reality, torn between trying to observe the bird, photograph it, video it and now analyse it's grey scale, it was probably a bit much to expect a perfect result!

Resorting to method two I have selected an image containing both a HG and the GWG.  Having first confirmed the exposure of the image is reasonably accurate (by testing that the HG grey scale is within the expected range of 4 to 5 I have been able go test the GWG close by and confirm that its mantle shade is also within it's expected grey scale range for this species as illustrated below.

Shortly after I left the area another related species, an adult Yellow-Legged Gull (L. m. michahellis) rolled up beside the GWG on the same patch of rock and Killian Kelly was able to obtain an image of the two together with an L. a. argenteus HG  There is enough in this image to add further confirmation that the grey scale value for the GWG is within the right ball park of approximately grey scale 6.  L. m. michahellis falls in the grey scale range 6 to 7 as shown in the illustration below.

Another useful comparison available between Yellow-legged Gull (upper 2nd from left hand bird) with Herring Gull (left hand bird) and Great black-backed Gull and the Glaucous-winged Gull (right hand bird).  Thanks to Killian Kelly for allowing me to post his great image.

There are a couple of important caveats to consider when applying this relative comparison method.

Sample Point Selection
Lighting on three-dimensional subjects paints a range of tonal levels as we all know, so how do we know where on each subject to take comparable samples from?  As discussed in my earlier posting on grey scales and gulls, the upper mantle seems to be the most consistent place to sample grey from.  Killian's image above is a good example of why this is the case.  Typically, under normal natural lighting conditions where gulls are at rest, lighting on the upperparts is at it's most consistent on the upper mantle area.  This is also coincidentally where many large gulls first develop a large uniform patch of grey scale plumage during their 2nd or 3rd calendar year making it possible to use this method for different gull age cohorts.

Lighting Varying With Perspective
I have spent a good deal of time pondering the question of lighting and perspective in postings HERE and HERE.  There are various reasons as discussed in those postings why the lighting and therefore the pattern of tones will vary between identical subjects positioned in different places across the same image.  Good sample point selection can help with the problem but I think we must allow for a small margin for error in a comparative analysis of samples.  I don't believe this margin is more than 1 grey scale in this kind of analysis.  Some of the factors that influence the result include direct versus diffuse lighting, lens focal length and therefore field of view and vignetting, the subject's posture and relative distances of subjects from the camera etc,

We are helped in this particular case by personal observations in the field which confirmed that this GWG was only slightly darker than nearby HGs and, from experience somewhere in the YLG or argentatus HG sphere of greyness.

Last Word - Primary Grey Scale
There has been a certain amount of disquiet expressed by some who have seen the bird in life that the primaries seem excessively dark in some cases including from some photographs.  Among some of the photos that have appeared of the bird online I sense an overuse of contrast which certainly will make the primaries appear darker relative to the mantle.  This for me raises an interesting question.  Just what is the maximum allowable grey scale value for the primary pattern in GWG?  At what point should we begin to suspect a 1st generation hybrid?  I am going to put this question to the Frontier's of Bird Identification forum and will report back on any useful findings.  This last image shows the left wing primary pattern in reasonably diffuse, low contrast lighting (note P10, the outer primary is missing).  From my own observations of GWG in Vancouver in July 2010 I certainly saw birds that matched the pattern of the Irish bird, but how many of those were 1st generation hybrid and hybrid back-crosses with pure GWG?  Will a DNA sample even provide us with the answer?

Based on a direct comparison between the grey of the wing and the darker primary tip pattern I have arrived at a grey scale value of 11 to 12 for the primary tips of this individual.  The question is where does this sit within the normal maximum pigmentation range for GWG?

Here is a brief video of a flock of GWG's which I took while visiting Stanley Park, Vancouver in July 2010.  Note especially the primary pattern of the incessantly calling adult bird which shows roughly the same level of primary grey as the Irish bird.  Note also, out of interest, the rather bleached immature bird.