Breeding For Color

By Mike Safley

The Spaniard, Cieza de Leon, made the first written record of alpaca and llama colors in 1553. The Indian herds of Chile and Bolivia that graze the altiplano still exhibit all the colors found in that original list. Today, the herds owned by Julio Barreda and the big Peruvian cooperatives are primarily white; the smaller Quechua herds of Peru still contain all the colors, but tend to be populated by light-colored alpacas.

Genetically, there are two basic alpaca colors: red and black. The original South American camelids, vicuñas and guanacos were reddish-fawn. Since alpacas are the descendants of these two species, the base color of alpacas is most likely reddish-fawn. Guanacos have both gray and black colors around their heads and this is probably the source of the black gene in alpacas. White is simply the absence of any of these colors.

In addition, there are an almost infinite variation of shades, which are caused by dilution and extension genes that modify the color genes. The theories about which colors are dominant or co-dominant, and which are recessive are often in conflict. The frequency of certain colors is manipulated by breeders and this creates the impression that certain colors are dominant, when they may be simply more apparent.

Fleece color is generally thought to be inherited according to Mendelian principles, but there may be an exception operating. The actual method of coat color inheritance is in question and issues such as how many color or modifier genes exist or which colors are dominant or recessive are not settled. A number of researchers suggest the alleles of each gene pair interact with one another in a dominant or recessive fashion to determine the color of an individual fleece. One researcher suggests that color is inherited in a more complex fashion, as the result of gene linkage. C. Renieri, a member of the faculty of the University of Camenino in Italy, in his 1993 paper, "The Genetic Basis of Pigment in South American Camelids,” wrote that "At present a modern and complete theory over coat color genetic determination in South American camelids lacks completely." Here we will cover some of the more prominent theories of how coat color in alpacas is determined.

HOW COAT COLOR IS DETERMINED

Color inheritance patterns for laboratory animals, such as mice, and some larger domestic animals such as dogs, cats, cattle, and particularly horses have been intensively studied. Coat color in mammals is almost entirely dependent on the presence or absence of the pigment, melanin, in the skin and hair. Melanin is produced by cells called melanocytes, which are concentrated in the hair follicles, the skin epidermis and the retina of the eye. Color in these areas is determined by the size and shape, as well as by the type, number and distribution of the cells or granules of melanocytes. There are two distinct forms of melanin: eumelanin (brown/black) and phaeomelanin (red/yellow). White coat color is the result of: 1) either no pigment; 2) the extreme dilution of red pigment; or 3) a large spot of white superimposed over the entire animal.

In the life of a Peruvian alpaca, white often means survival since white fleece is the color of choice. As a result, in Peruvian alpacas, white genes in various combinations are frequent. The anecdotal evidence is that the white gene is, at least in some alleles, dominant due to the fact that they are passed on to white cria from colored parents. Alternatively, this may be explained by the theory that color is inherited through a process of gene linkage.

Dr. Philip Sponnenberg had this to say about determining color in alpacas in his paper entitled "Jiggling Genes:”

One of the challenges in understanding color in alpacas is to understand that every alpaca has genetic machinery to produce color. On many, though, the whiteness has been superimposed either completely or partially. Therefore, white animals hold lots of hidden surprises for the alpaca breeder. These surprises can be used to good advantage by astute breeders.

The exact genetic control of color in alpacas has never been elucidated. Part of the reason for the lack of information is that most research has focused on fleece color. Fleece color alone does not reveal the genetic intricacies relating to the overall alpaca. As an example, imagine that bay and chestnut horses were alpacas--both would grow red-brown fleece, but the genetic control leading to that final color is distinct, and each will behave very differently in a breeding program. The lesson here is that it is important to look at the entire animal to evaluate the color phenotype, which can then be used to estimate the underlying color genotype.

My basic approach to understanding the color of any animal is to first try to remove the white. This is clearly impossible for white or nearly white individuals. Looking at color is important, and the important questions to be answered include deciding which pigments are present, their locations on the animal, and their relative intensity. My experience with alpacas is not as vast as mine with sheep and goats, but my experiences so far indicate that the following are the basic options for colors:

• black that does not fade or sunburn
  
• black that fades or sunburns to a reddish brown at the tips
  
• black with a light belly (these are tough to spot in fully fleeced animals)
  
• red/brown with no black trim
  
• red with black trim (nose, eyes, lower legs but maybe only toes)
  
• brown or chocolate with black trim
  
• shaded colors with pale lower areas, darker top areas (usually on a reddish tan)

These appear to be the basic patterns available, with other colors derived from these basic ones. The control of these is going to be complicated, and that is because several different loci (or genetic addresses, each with a few choices at that address) can control the final outcome. The several different loci can be imagined to be a series of switches. The switch choices are the different alleles at each locus, and the sum of these choices gives the final outcome. The beauty of this system is that a relatively few loci, with few choices at each location, can give a whole wide range of final colors. That, alas, makes predictions somewhat difficult.

The skin pigmentation of sires and dams could play a role in predicting the color of an alpaca cria, says Julio Barreda. His observations have led him to believe that the skin color of white alpacas can help predict the color of offspring. "Animals with pink outer and inner lips, eyelids, and toenails will produce white progeny when mated to similar phenotypes," he says. "Black-lipped white alpacas will often produce colored progeny if mated to a colored alpaca and may produce fawns when mated to one another."

HOW COLOR IS INHERITED

When Mendel studied his peas, he got lucky. The genes affecting the traits he observed in his pea plants all occurred on different loci on different chromosomes. Chromosomes assort independently (i.e., there is no tendency for certain chromosomes to stick together in germ cell--egg or sperm--formation), so the genes on those chromosomes assort independently too. Because all the genes Mendel was studying did, in fact, assort independently, he believed all genes assort independently, hence his law of independent assortment.

Today geneticists know that there are exceptions to the law. Exceptions to Mendel's second law are caused by linkage. One of these exceptions may apply to alpaca coat color inheritance.

Two gene loci are linked if they occur on the same chromosome. Because entire homologous chromosomes--and the genes they carry--are separated at meiosis (the process by which chromosomes are reduced to half their original number during gamete formation), genes on the same chromosome tend to end up in the same gamete (germ cell). This is only a tendency, however, because of a phenomenon known as crossing over. Crossing over involves a reciprocal exchange of chromosome segments between homologous chromosomes and occurs during meiosis prior to the time the chromosomes are separated to form gametes.

Crossover events are common, and the probability of recombination of genes at any two linked loci depends on the distance between the loci. Loci that are far apart are likely to recombine often. For practical purposes, the genes at these loci will assort independently, just as they would if they had been on different chromosomes altogether. Recombination is much less

Group of colored Peruvian alpacas, 1960.
Photo: Julio Barreda

likely for loci that are very close together because the probability of a break occurring between them is much less. These closely linked loci create exceptions to Mendel's second law.

Color inheritance patterns vary considerably from one species to the next. For instance, mating horses of the same color does not generally produce the same color progeny, except for maybe sorrels or chestnuts. Alpacas appear to breed true much of the time, i.e., white x white often equals white.

These variations occur because genetic traits can be simply inherited at one locus or polygenically inherited at several loci. This means that one gene or set of genes at one specific location on the chromosome may be responsible for a trait, such as color, or the trait may be caused by several different genes located at different spots on one or more chromosomes.

There are many species of livestock in which color is simply inherited. For instance, black or red coat color in cattle is simply inherited. The black allele B is dominant and the red allele b is recessive. Producing red cattle is very easy: Keep only the red cattle. After one generation all the progeny would be red. Black is a little more complex because the red recessive gene could be present in a bull with a black phenotype. (Please note that in dominant-recessive gene action, B-black b-red, there are three possible gene combinations, but only two possible colors, BB and Bb equal black and bb equals red.) By only using bulls that were homozygous for black, the red gene could almost be eliminated over time and all of the progeny would be black.

Complete dominance, as in the black cattle example, will produce the dominant color when the dominant allele is paired with a recessive allele of another color. Complete dominance is the one form of dominance in which heterozygous and homozygous dominant genotypes have the same phenotypic expression. Co-dominance occurs when the recessive gene expresses itself equally with the dominant gene. An example of co-dominance occurs in the coat color of shorthorn cattle in which each genotype (RR, Rr, or rr) is associated with a distinct phenotype, red (RR), roan (Rr), or white (rr). When there is co-dominant gene action, there can be two genes and three phenotypes, as in shorthorn cattle. The co-dominance in shorthorn cattle that produces the roan color from a mix of both red and white hair could also explain gray alpacas, both silver and rose, which are the result of the combination of black and white, or red and white fiber.

Breeding for alpacas of a specific color is much more complex than breeding for coat color in cattle, because most researchers think coat color in alpacas is inherited polygenically. (In horses, as many as 12 loci are thought to affect coat color.)

Most theories of color inheritance in alpacas rely on Mendel's rules of dominance and random independent assortment. Everyone agrees that color in alpacas is controlled polygenically in the form of a) color genes, b) modifiers, and c) extenders. These three types of genes are universally thought to reside on separate chromosomes. Modifier genes in the form of multi, spotting, or diluter genes are thought to sort independently according to Mendel's laws; each of these would be a switch station in Dr. Sponnenberg’s analogy. Another theory of color inheritance in alpacas is that all colors are linked on the same chromosome and color is determined at meiosis, the process by which chromosomes are reduced to half their original number, by a recombination of the color genes.

There is far less agreement on just how many color genes and loci there are. Toledo and San Martin reported in 1948 that there were three series of genes; in 1968 Bustinza reported four series of genes. And there are several more color inheritance models, which contradict one another.

There are three leading theories of coat color inheritance by assortment and dominance: those of Humberto Gundarillas, Dr. Julie Koenig, and Dr. Philip Sponnenberg.

A 1983 article by J. Tillman entitled "Coat Color Inheritance in Llamas and Alpacas," published in Llama World, presented Gundarillas' theory that there are four genes controlling coat color. Those four genes are:

1. C locus wild gene with cc producing white (white is recessive)
   
2. V for brown and vv for black (black is recessive)
   
3. S for solid color and s for spotted
   
4. Lw which controls the extension of spotting for pigmented animals and Lw/Lw for full white animals

Gundarillas also concluded that solid color is dominant over multicolor.

Koenig presents a more complex scheme of inheritance involving eight genes. Three color genes determine the base color of the animal as follows:

1. White: W gene. Two alleles, W and w. WW or Ww produces white (white is dominant), ww produces color which results from the A gene.
   
2. Vicuña and guanaco color. A gene. Four alleles: A+, A, a+, a. Various combinations produce alpacas with light bellies and white inside legs, red-brown bodies and necks, or black bodies with brown underbellies.
   
3. Brown and black. B gene. Two alleles, B and b. BB and Bb produce black (black is dominant), bb produces brown (brown is recessive).

Koenig also theorizes that there are five genes which define the intensity or pattern of color: C and D (affect dilution of color); R (affects roans); S (affects spotting pattern); and P (affects solid versus piebald patterns).

Sponnenberg, an acknowledged expert and author of numerous equine color studies, speculates that primary color is controlled at two separate loci: the extension loci, and the agouti loci. He acknowledges that both of these loci may not be present in alpacas and that the interaction between them is very complex. He proposes that the basic colors may (or may not) be controlled as follows:

Extension:

• black - dominant (called dominant black)
  
• wild type - allows expression of agouti
  
• red - recessive or chestnut red, no black hair

Agouti

• red with black trim dominant
  
• red with extensive black trim
  
• black and tan (tan belly on black)
  
• black - recessive black

COLOR INHERITANCE BY GENE LINKAGE

One theory of alpaca coat color inheritance stands apart from all the others. Researchers William L. Wall and Ron G. Cole, of Australia, who both own alpacas, propose that Mendel's rules of dominance and independent assortment do not entirely explain the inheritance of coat color in alpacas.

Wall's area of interest is agricultural sciences, especially genetics; Cole comes from a mathematics background. They propose a model of inheritance based on gene linkage.

The Wall/Cole theory of inheritance grew from their statistical analysis of matings that were registered by the Australian Alpaca Association's registry. In all, they studied the color of more than 10,000 cria from registered parents whose coat color was known. The results of these matings were compiled in two sets of coat color tables (presented in their entirety in appendix 2): Version 1, which compiled the coat colors of over 7,000 cria, and Version 2 which included the coat colors of an additional 3,000 cria.

Wall and Cole's theory of coat color inheritance in alpacas formed as result of analyzing Version 1 of the tables. They then used their theory to predict the color distribution among the additional cria. These are the figures charted in Version 2. The accuracy of their predictions lends considerable credibility to their ideas.

The goal of the Wall/Cole research was to:

1. determine the minimum number of genes necessary to explain the range of colors found in alpacas
   
2. map the genes on the chromosomes
   
3. explain the action of modifier genes.
   
4. explain the action of the multi gene

In the process, they concluded that coat color inheritance was determined by the process of gene linkage and not by dominance and simple assortment. They further concluded that there were five genes total: three primary color genes--black, red, and white--which are linked on the same chromosome; a modifier gene which determines the amount of color; and a multi gene which determines the distribution of color. Wall/Cole hypothesize that the chromosomes carrying the three linked color genes resemble the above diagram.

Once Wall and Cole settled on the gene linkage method of inheritance, and determined from their coat color tables the relative distance apart of the linked genes, they were ready to predict the outcome of the additional matings that were included in Version 2 of the coat color tables. Their predictions were more than 90 percent accurate.

Because the B, R, and W genes are linked, this allows for 64 possible genotypes (4 alleles X 4 alleles X 4 alleles = 64) which are expressed as 27 phenotypes. This conclusion is reached by taking the B (black) gene, its alleles are B and b, where BB, Bb, bB, or bb represent four possibilities, and making the same assumption for R and W, therefore 4 X 4 X 4 = 64. However, as Bb and bB are indistinguishable, there are three phenotypes (BB, Bb, and bb). The same is true for R and W. Therefore 3 X 3 X 3 = 27 phenotypes.

In similar fashion, Wall/Cole theorized that the diluter gene has four genotypes and three phenotypes: DD, Dd, dD, and dd. When you take the 27 color phenotypes available and multiply them by the three diluter gene phenotypes, the result is a potential for 81 different phenotypes. This range of possible color shades explains every conceivable alpaca color. These colors would occur on a continuous variation from light to dark, red to brown, to fawn and white, etc.

The research derived from the color tables also led Wall/Cole to theorize that there are three alleles of the multi gene: O, o, and ø with solid (O) dominant. The multicolored coat in alpacas is expressed in many forms. These forms include:

1. A small white blaze on the face of an otherwise totally black animal;
   
2. Boots (i.e. feet and lower leg colors different from the coat color expressed over the rest of the animal)
   
3. White on white or black on black (i.e. white spots on a white coated animal or black patches on a black coated animal which, because of the base color of the animal's coat, are unseen as spots or patches).

All grays in this genetic context are considered multis, with the possible exception of "true solid gray."

Calculating the various possible phenotypes that would occur from specific matings under this theory establishes that a two-to-one ratio of solid to multicolored animals would result from matings of multicolored parents. This conclusion is also consistent with the data found in the tables. Finally, their research confirmed that all grays were multis with the black, red, and white genes operating.

Wall and Cole's research was verified independently by examining published data presented by Rigoberto Calle Escobar, who, in his book Animal Breeding and Production of American Camelids reported the following results of a color mating study conducted at La Raya Ranch:

From observations made at La Raya Ranch 1,000 white females mated with white sires produced 50 to 60 percent white offspring; 19 percent were light fawn; 17 percent were patched. In decreasing order came cinnamon, light coffee, dark coffee and black.

It was also verified that from every 300 offspring of the white with white cross, only one completely black offspring was produced. Similarly from the crossing of white sires with other colored females (with exception of light spotted fawn) a predominance of the mothers' color was noted. In the case of females with light fawn and spotted, forty percent of the offspring are white. These results of color crosses which have been verified, reinforce the thesis that color inheritance is complex and is based on many pairs of genes which, because of a not very intense selection in the herds, are maintained in a pool of genes of the population, conserving color variability.

It is interesting to note how Wall/Cole's study's predictive value holds up in explaining the results of the La Raya color mating study. Escobar's La Raya observations and Wall and Cole's calculations from the Australian herd when white was mated to white follows in Figure 12.

MODIFIER GENES

Basic alpaca colors are thought to be diluted or presented in several different shades by the action of a dilution or extension, modifier gene. Modifier genes do not control a trait, but they can determine variations in the phenotype of animals which have the same genotype, for instance, the difference between light brown and dark brown. These genes most likely occur at different loci than the primary color genes. An example of these genes would be Koenig's C, D, R, S, and P gene; Gundarillas' S and Lw genes; and Wall and Cole's fourth gene, a diluter, and fifth gene, a multi gene that controls the distribution of color.

The exact genetic mechanics of the interaction of primary color genes and modifier genes has not been scientifically established. It is possible the same result, for instance a certain shade of fawn, could be the result of several different mechanisms. Sponnenberg says:

The usual rule appears to be that red pigment is diluted, but black is not. Red can be diluted to a wide range of shades of tans and fawns, all the way to ivory or white. If black were diluted, the expectation would be solid and uniform blue-grays, which if present in alpacas are quite rare.

BREEDING FOR COLOR IN PRACTICE

What happens as a practical matter when you breed white to white, black to black, one color to a different color or solid color to multicolor? Alpaca breeders are fortunate to have two studies to draw from. The first is Wall and Cole's exhaustive study of coat color inheritance which is intended to be an easy reference for breeders (see Tables 1-12 in the appendix). The study is based on the phenotypic color of the parents and their progeny; it is not intended to suggest the alpaca's genotype.

The base data for the Wall and Cole work was derived from the Australian Alpaca Association registration database which records alpaca registrations with designated colors. The tables were created from registrations as of March 1996 and included 10,849 alpacas.

There are two types of tables:

1. The solid color cross tables, which present the progeny from crosses of sires and dams of the same color. Numbers of crosses and sex of progeny are listed, together with numbers of cria for each solid and each multiple color registered (Tables 1-8 in the appendix).

   
   A typical Indian herd of mixed color alpacas.
   Photo: Mike Safley

2. The individual color cross tables which list number of matings and sex of progeny, together with results of analysis of each color of male crossed with each color of female and vice versa for each of the colors. There are four of these tables (Tables 9-12 in the appendix).

A second well-documented study useful to alpaca breeders is that done by George Davis, MS, of Ag Research in New Zealand. The alpacas in the herd studied to create Table 13 (in the appendix) were imported from Chile and were owned by the research center. The parents of the progeny who were the subject of the study were pen bred to help assure the accuracy of their pedigrees. The New Zealand study was a much smaller sample group than the Australian study. The color of the alpacas used in the study was based on the main body and not on the extremities. The New Zealand study used different color definitions than the Australian study.

It should be understood that the color tables can not be used to predict the outcome of a specific cross between two animals. The data presented is an analysis of the combination of all available data. It is meant to present the results of past experience.

An alpaca breeder might choose to study the various tables to determine what has transpired in the Australian National Herd as a guide to the likelihood of various possible color outcomes from specific breedings. Wall and Cole suggest that readers of their coat color tables pay attention to the "white space" in the tables. They point out that the absence of offspring of particular colors, as evidenced by "white space," is as informative as the offspring recorded in the tables.

OBSERVATIONS ON COLOR MATINGS

In the Australian color mating tables (Tables 1-12 in appendix), the color of the alpacas were grouped as follows:

1. fawn and roan alpacas were assigned to red;
   
2. silver grays and blacks were assigned to black;
   
3. browns were assigned to brown;
   
4. whites were assigned to white;
   
5. multi-coloreds were assigned according to the mix of colors listed, for example, a dark fawn/light fawn/white alpaca was assigned to red; a dark fawn/medium gray alpaca (roan) was assigned to brown.

Understanding this, you can use the charts to make the following observations:

1. When breeding white to white, the progeny were 60 percent white; 18 percent red; 17 percent brown; and five percent black.
   
2. When breeding white to brown, the progeny were 43 percent brown; 10 percent black; 27 percent red; and 20 percent white.
   
3. When breeding black to black, the progeny were 85 percent black; 11 percent brown; one percent red; and three percent white.|
   
4. When breeding white to black the progeny were 24 percent white; 14 percent red; 30 percent black; and 32 percent brown.
   
5. When breeding brown to black the progeny were 52 percent brown; 40 percent black; three percent red; and five percent white.

The New Zealand study produced results similar to the Australian study, although the colors were simplified to white, brown, black, gray (mixed white and black fibers), and roan (mixed white and brown). The multi-colors were described as piebald (white and black patches) or skewbald (white and brown patches). This approach ignored the subtle shadings of brown and fawn, but it ensured consistency in assigning an animal to a particular color group. Coat color was determined at skin level to avoid mistakes in identifying color changes caused by weathering effects. The following observations can be made from studying Table 13 in the appendix.

1. Mating two white parents, producing 81 progeny resulted in 63 percent white and 25 percent multicolored.
   
2. Where only one parent was white, and there were 159 progeny, there were 32 percent white and 25 percent multicolored. In 132 matings in which the parents were either black or brown, there were only two percent white cria.
   
3. Where both parents were black, producing 26 progeny, 73 percent were black and eight percent were brown.
   
4. Where both parents were brown, producing 76 progeny), 68 percent were brown and 18 percent black.

The fact that only two percent of the cria from colored parents were white supports the theory that white is dominant. If white were recessive, many black and brown alpacas would probably carry one white copy of the gene and when mated together, white progeny would occur in about 25 percent of births.

But the small number of white cria also supports Wall and Cole's theory that the distance between black and white on the linked chromosome map is such that white will result from this breeding infrequently. If brown were completely dominant over black, no brown progeny would be produced where both parents were black, because their color would be the result of double recessive black genes. If black were completely dominant over brown, there would be no black progeny where both parents were brown for the same reason.

The New Zealand color tables do not fit either the black dominant or brown dominant model, although they are closer to the dominant brown model. The Wall/Cole study explains these statistical outcomes by using an inheritance model based on gene linkage.

THE BOTTOM LINE

The color of the progeny can often be predicted with accuracy if the breeder is familiar with the stud being used, particularly if he has sired a large number of offspring. A famous alpaca stud, Hemingway, is a good example. He has been bred to more than 30 black females. All the offspring, 100 percent, have been fawn, mostly dark fawn. When Hemingway is bred to solid-colored females, such as brown or fawn, he almost always produces a lighter colored cria in the same basic color of the mother; when bred to white females, he produces white cria. Accoyo's El Moustachio (white) and Accoyo's Victor (fawn) often produce a cria the color of the mother, especially Victor, who has thrown a lot of black crias when mated to black females.

The highest likelihood for creating a certain color occurs when mating two alpacas of the same color. Alpacas seem to carry a variety of color genes, especially white alpacas. If Cole and Wall are correct, every alpaca carries every color. When crossing a white alpaca with a colored alpaca, the progeny are more likely to be colored than white by a considerable margin. Two colored alpacas almost always result in colored progeny. Pintos can pop up almost anywhere or, as Barreda says, "pintos are hard to get rid of."

Alpaca breeders need to form their own goals as to colors. If they want to produce unique colors for the pet market, they can mix up solids with multi-colors, black with white, and so on. If their goals involve eventually producing commercially valuable fiber, they can breed solid to solid, preferably white.