A Genetics Puzzle – Coat Color in Border Collies

Some many years ago, back when I was an instructor of first and second year biology at a British Columbian college, I used to give my students genetic problems.  These are really exercises in logic, brain teasers, if you will.  I’ve always loved them, although I understand that some of my students weren’t quite so fond of genetics.  Recently, I have been inspired (more about that in another post) to take an interest in the genetics behind coat colors in dogs, specifically border collies.  So here’s the puzzle …

The seven pups shown in the image above were born as a result of a cross between the two parents shown below:

Here is some more information:


Mahogany sable with amber eyes; parents were both black white tri-colored.


Black white tri-colored with wall eyes – one blue and one marble; nose and front legs have a small amount of ticking; parents were both black white tri-colored.


Three of the pups are black white tri-colored.  All three have black noses.  One of these pups has wall eyes (one blue and one amber) and limited ticking.  The other two have brown eyes and no ticking.

Four of the pups are brown and white.  All of them have noses which look to be either black/pink or brown/pink.  One of the four is a chocolate white tri-color with limited ticking and amber eyes.   The other three are much lighter in color, show a bit of black tipping on their fur, and are likely sables.  One of these sables has light patches around the eyes.  All three of the sables have wall eyes (one blue and one amber).

What is the genetic explanation for the colors of both parents and pups?

Let’s take a quick look at the genes that affect coat color in border collies.  These are:

A locus (agouti): 

Ay – sable
aw – agouti/wolf grey
at – tri-color or tan points
asa – saddle
a – recessive black

B locus (black/chocolate):

B – black
b  – chocolate

D locus (dilution):

D – no dilution
d – dilution of eumelanin to blue or lilac

E locus (extension):

E – eumelanin present
e – eumelanin suppressed to produce recessive red

H locus (harlequin merle modifier):

H – harlequin
h – non-harlequin

K locus (dominant black):

K – solid black
kbr – brindle
k – non-solid black

M locus (merle):

M – merle
Mc – cryptic merle
Ma – abnormal merle (maltese blue)
m – non-merle

S locus (spotting):

S – no white spotting
sp – piebald
sw – extreme white
si – irish spotting (may not be on S locus)

T locus (ticking):

T – ticking
t – no ticking

Tw locus (tweed merle modifier):

Tw – tweed
tw – non-tweed

A locus alleles will only be expressed when a dog does not have a dominant black (K) gene.  S locus alleles (white spotting) appear on top of anything else a dog has and are not masked.  D and B locus alleles will override all alleles for black (dominant or recessive) and change all eumelanin on the dog.  E locus alleles are not overriden by anything except the S (white spotting) locus, so recessive red will be expressed even on a dominant black dog.  The H and Tw locii will only be expressed on a dog with the merle allele (M locus) and the T locus will only be expressed on a dog with white spotting (S locus).

Wow!  I don’t think I ever handed out a problem that complex to my students!  If you find that this is already blowing your mind, a really good site to check out is Dog Coat Colour Genetics, which has lots of definitions, explanations, and good pictures.

In researching this topic myself, I found a wonderful infographic called All Border Collie Colors made by Suzanne Van Der Kraan.  I have taken the liberty of modifying this graphic by adding the genotypes for each color, which helped me wrap my head around the complex interactions between ten gene locii.

A couple of great web pages for images of actual collies with particular coat colors are Collie Colors and Looks and Diversity in the Border Collie.

OK, so there’s all the information necessary to solve the puzzle.  Let’s look at each piece in turn.


The male is a very nice mahogany sable [(kk)(Ayat)], like this rough collie:

What’s interesting here is that the male’s parents were both tri-colored [(kk)(atat)].  Sable is dominant to tri-color, and you can’t generate a sable by crossing two pure tri-colors.  So, something else must be happening here. My guess is that one of the male’s parents was actually a seal.  Seals were often classified as tri-colors, but they have a different genotype [(K-)(Ay-)].  Here is an example of a border collie with a seal coat:

Note that the coat is very similar to a black white tri-color, such as the the one shown below, which has quite limited amounts of brown.

Another clue to the male’s genotype is that there is at least one chocolate puppy in the litter.  To get chocolate, you need to have (bb).  Since neither of the parents are chocolate, they must each be carrying a single recessive chocolate allele, and the chocolate pup received a recessive allele from each parent.

My guess at the cross that produced the male is:

[(kk)(atat)(Bb)(D-)(mm)(sisi)]   x   [(K-)(AY-)(B-)(D-)(mm)(sisi)]

black white tri-color   x   seal

and the male’s genotype is:



On first glance, the female seems to be a fairly typical black white tri-color.  However, two things should be noted: (1) she has light ticking on her nose and forelegs, so she must have a (T-) genotype; and (2) she has blue eyes.

There are three ways in which border collies can have blue eyes:

        1. As a side effect of the merle gene (M-).  Since the merle gene causes random pigment loss, marble eyes (e.g., mixed blue and brown in the same eye) are a common characteristic of this gene.
        2. When a dog has large amounts of white around its eyes.  This is clearly not the case in my example.
        3. As a completely separate gene, unaffected by coat color.  This gene is, however, rare, although it occurs occasionally in border collies.  Characteristics of this gene pattern are full black nose pigment and black around the eyes.  This might be a possible solution to our puzzle of blue eyes.

A clue to the blue eye mystery is the color of several of the pups’ noses.  They appear to have some pink, mixed with either brown or black.  This is likely a mosaic pattern, which indicates that they may have the merle gene (M-).  However, except for the blue eyes, the female shows little indication of being a merle.  My guess is that she is a cryptic merle (Mc-), like this Australian shepherd:

My guess at the cross that produced the female is:

[(kk)(atat)(Bb)(D-)(Mc-)(T-)(sisi)]   x   [(kk)(atat)(Bb)(D-)(mm)(tt)(sisi)]

black white tri-color cryptic merle with ticking   x   black white tri-color without ticking

Note that both parents could be cryptic merles and ticked, or that one parent could be ticked and one could be a cryptic merle.  Also, one parent could be (BB).

In any case, I’m guessing that the female’s genotype is:



Based on the above rationalization, the pups are produced by the cross:

[(kk)(AYat)(Bb)(D-)(mm)(tt)(sisi)]   x   [(kk)(atat)(Bb)(D-)(Mcm)(Tt)(sisi)]

mahogany sable   x  black white tri-colored cryptic merle with ticking

My guesses at their genotypes are:

      1. [(kk)(atat)(Bb)(D-)(Mcm)(Tt)(sisi)]
        Black white tri-colored cryptic merle with ticking (one pup like the female)
      2. [(kk)(atat)(Bb)(D-)(mm)(tt)(sisi)]
        Black white tricolored (two pups)
      3. [(kk)(atat)(bb)(D-)(mm)(Tt)(sisi)]
        Chocolate white tri-colored with ticking (one pup)
        Example of a chocolate white tri-colored cryptic merle Australian shepherd
      4. [(kk)(AYat)(Bb)(D-)(Mcm)(tt)(sisi)] or [(kk)(AYat)(bb)(D-)(Mcm)(tt)(sisi)]
        Mahogany sable or chocolate sable cryptic merle (three pups)
        Example of a sable merle border collie

So there you have it!  Other guesses welcome!