Blue, blue type2, turquoise, *sapphire*, *teal*,…. and so much more ….

This article has been published in April 2019, in Agapornis.Info, the BVA-International magazine. In the mean while we continued our research and we have following updates:

Update 16/02/2021

Update 02/12/2020
Blue Agapornis fischeri and Agapornis personatus are considered not producing any psittacofulvins (psittacines) in the feathers. However, we could not ignore the fact that there are breeders who argue that their – pure bred – blue2 birds sometimes show traces of psittacofulvins in their feathers (to be clear, blue2 not in combination with *sapphire*).

Others are 100% sure and state that even in budgerigars [Melopsittacus undulatus] and ring-necked parakeets [Psittacula krameri], blue (2 ???) birds always contain psittacofulvins remains in their feathers. If we consider the possible genetic causes behind this mutation, then we can not completely exclude this possibility.
So, we believed it was time to try to find a clear answer and not to continue to rely on assumptions. There fore, we have collected feathers from (100% pure bred) blue2 Agapornis fischeri and Agapornis personatus. We have scan them with the Raman technology and found is NO evident or traces of psittacofulvins in the feathers of blue1 or blue2 birds.

Update 08/09/2020
Teal is considered now as being a basic mutation

Blue, blue type2, turquoise, *sapphire*, *teal*,…. and so much more ….
Article published in BVA International Magazine April 2019.

By Dirk Van den Abeele
Ornitho-Genetics VZW
MUTAVI, Research & Advice Group

Everyone knows by now that in recent years there has been a lot of talk about blue birds within the genus Agapornis. For starters, we are now certain about the existence of blue 2 in Agapornis fischeri, there is the possibility of several new mutations and a breeder in Spain has bred green young from a pair of blue Agapornis personatus. It goes without saying that a lot of people are left with questions. A lot of the questions we currently get are about this topic and the internet and social media is buzzing with rumour, gossip, theories and speculation. Each person has his own opinion. Hence it is difficult to get a clear view on the matter.

For a novice aviculturist it might look simple at first, but the more experienced breeder does realize it is not as simple as that. In addition to the “normal” blue birds there are also the “nearly blue” birds, such as aqua and turquoise. These two mutations result in an intermediate type when they are paired with each other or with blue and are therefore considered to be alleles of the bl-locus [1, p. 251].

When we received the first reports in 2012 about the presence of the turquoise phenotypes in Agapornis fischeri and Agapornis personatus they posed of course new challenges for the aviculturists. A lot of breeders took up the challenge and several mutation combinations with turquoise were set up. The sky was the limit.

Those who looked at it more scientifically soon realized that again a lot of questions need answers. With the knowledge about the various alleles of the bl-locus in mind and the experiences with other bird species we know that there is indeed interaction between turquoise and blue, but not everything is sufficiently clear yet. There are still a lot of unanswered questions.

The breeding results we received from America about this turquoise Agapornis fischeri soon taught us that we are mainly dealing with heterozygotic birds (TurquoiseBlue, combinations of a turquoise and blue allele). The main question was what the homozygotic (purebred) turquoise looks like.

In my first article about this turquoise Agapornis fischeri [2] in the BVA-International journal of April 2012 I already mentioned this. I quote from the article: “Quote …One of the possible problems with this mutation could be that the visual colour difference between turquoise and TurquoiseBlue might not always be clearly distinguishable. ……………… Therefore, for me it is very important that the combination of turquoise and blue is avoided as much as possible. Since it will be the purebred (homozygotic) turquoise type which will be accepted at the various shows. Also, a proper administration and a correct record of the breeding results are of great importance to avoid any misunderstanding. There are already too many misunderstandings and stories floating around about this mutation. …..” unquote.

During the various seminars worldwide, we kept insisting on this. Yet most breeders prefer to pair these turquoise birds only with blue. Which makes sense because due to the combination with blue phenotypically turquoise young are already possible and the desire to have the first turquoise young in the nest usually prevails. We cannot really fault anyone for this.

It has always been custom within Ornitho-Genetics VZW to buy these new mutations ourselves to set up the needed test matings. But because the cost of everything keeps increasing and the research budget is not increased at the same rate, we will need to stop doing this. Hence, we need the help and information from other aviculturists. As you can understand this is not as straightforward as it might seem. Not everyone keeps track of all data and can or will not share this data. The few breeding results we were able to obtain all pointed in the direction of ‘normal’ turquoise. A mutation which we know in a lot of parakeets, among other also in Agapornis roseicollis.

Two years later, in 2014, we received the first reports that turquoise young were born out of a blue bird. Unfortunately, little was known about the origin of the birds, so we did not get the answers we needed. There were still too many questions and doubt to build from this. Yet at that point we were already considering the possibility of the presence of blue type2. Some time later a few more similar reports came in from other aviculturists, but they all had to admit that the ‘blue’ birds concerned, actually still had psittacine hue and were probably plain turquoise. In the meantime, we are still looking for breeding results. The little we were given still pointed towards turquoise as we know it in most species.

In 2017 we again received a number of reports of blue birds which has turquoise young. Everything was recorded carefully by our staff and the idea about the presence of blue type2 was getting more and more firm, but again hard evidence was still lacking. These aviculturists were repeatedly asked to send their breeding results, but they never arrived.

While gathering this information we usually asked very general questions, without revealing any details. Hence, we usually ask for the breeding results in general. Of course, we could ask explicitly whether there are aviculturists who have specific results but we try to avoid this as much as possible. Past experience has taught us that we will then immediately get a number of responses from aviculturists who will confirm this but unfortunately this is not always correct. Some do it solely for the attention but others, as it turns out later, ‘because they misunderstood’ or there is still some confusion regarding mutations, background of the birds, etc… That is the reason why we try to keep our questions as general as possible, since each detail matters. For some this might be difficult to understand, but for us it is an absolute necessity that the research is conducted objectively. We also never look for an indication to confirm something, not at all, because those are easy to find. We always look for indications that prove the opposite. If we do not find those and all possible questions/possibilities have been answered/checked, then we can assume that it is indeed the case. This is also what happened with this blue type2.

Of course, we can build on the statements of one person or accept and confirm these, but we never do that. Scientifically speaking this is not even acceptable. All reports must be confirmed at least twice, through independent sources and there must be no more open questions [3]. I understand that this must be frustrating for people dealing with these mutants, but we do not make any exceptions for anyone. We again noticed that some people do blame us for this and assume we want to ‘sabotage’ them. Others feel that we are being pressured and comment about that. Believe me, this is the story of my life.
Fortunately not everyone is so small minded. Half-truths and assumptions have never benefited anyone. I do not think anyone is waiting for amateurism. Besides, you will notice further on in this article that things are not as simple as they appear at first sight. A lot of items need to be taken into account before we can reach the correct conclusions.

2018 was almost completely taken up by a number of translations of my book and ongoing research. In May 2018, a South-African aviculturist, published an article where he stated that the homozygotic type of turquoise was a blue bird in his opinion. He based his conclusion on the breeding results of his birds. He did not make the link himself with blue type2 in his article, but his article did bring it to the foreground. He spread it through social media, which in turn provoked a lot of reactions towards us. Most however denied the content of the article and claimed to have completely different breeding results. In other words, it was quite an endeavour to judge and verify everything objectively. They were all requested to send us their breeding results and to set up extra pairings if needed. And now, probably because of this article, luckily more information came our way. So, we could keep on working.

Once the ongoing research and translations were finished in November 2018, we had received about a hundred emails on this subject. There did not appear to be a main thread. One thing is for sure, there exists a lot of confusion and misunderstanding about the bl-locus. Most had just always paired their birds with birds from the blue series.
We were forced to directly launch the question through facebook whether there were breeders who could demonstrate using their breeding results which phenotype a homozygotic turquoise has, in other words who could confirm the article from South-Africa. Believe me, a lot of breeders again stated in no uncertain terms that it definitely was not blue. A few were hesitant, one or two people had concrete proof that this was indeed a blue bird.
All breeding results were again reviewed thoroughly by us and from the reports of two aviculturists it was clear that there are indeed homozygotic ‘turquoise’ birds which have a blue phenotype (amusing detail: proof came from a breeder who actually wanted to prove the opposite using his breeding results…).

Unfortunately, one clear indication was still missing and that is when someone can state that he or she had all turquoise young from the combination of two blue birds, one of which comes from the ‘turquoise line’. This is actually for us the last unanswered question. It was a Philippine aviculturist who provided us with the proof. As a bonus, it was a BVA-International judge, who set up targeted trial pairings with turquoise and who could provide is, on exactly the same day, with an extra confirmation. He had a blue young from the combination green/turquoise x green/turquoise. Everything seemed to be moving along rapidly and we could finally substantiate and confirm the existence of a blue type2.

Of course, we do not exclude the possibility that in addition to this blue type2 there is also a real turquoise. This must be investigated further, taking into account the fact that we now know for certain that there is a blue type2. It might look simple, but it definitely is not. The best way forward would be to investigate with as many aviculturists as possible. Hopefully this article will help you in gaining a better understanding of this subject matter and help us to get the answers to so many questions.


It seems logical that we must first have a look at the blue birds. Just to be clear, there is no blue pigment in the feathers of lovebirds. It is the structure of the keratin, this is the material from which the feather is made up of, which is responsible for this blue colour [4, p. 341]. The reflection of light on/in these keratin structures causes the blue colour. This is nothing new. In 1665 Robert Hooke already indicated in his book “Micrographia” that the feathers of a peacock [Genus Pavo] reflected a blue colour [5]. In Agapornis roseicollis this was investigated and described scientifically for the first time by Dyck in 1971 [6].

Because most feathers in lovebirds are green, we need a mutation to end up with a blue bird. To understand this properly we must first understand why a green feather is ‘green’.

This green feather consists of a number of parts. The most striking parts of a feather are the shaft in the middle of the feather, and on both sides of this shaft there is ‘a vane’. This vane consists of barbs, which look like tubes, and these (tubes) barbs, all next to each other, make up the vane. This is the part of the feather which displays most of the colour.

If we cut through the barb of a green feather horizontally and look at it under a microscope, we will actually be able to distinguish three different rings. The outer ring, called the cortex, contains in the green feathers of lovebirds a bright yellow pigment. This pigment is known under numerous names: psittacofulvins, psittacine or parrodienes [7]. In Europe it is customary to use the term psittacine [8, p. 70] similar to budgerigars [Melopsittacus undulatus]. All these names have actually been derived from Psittaciformes (parrots) since this pigment is known only in parakeets as far as we know.

The middle ring is called the spongy zone. This is colourless keratin containing a very fine tube-like structure, a crisscross of narrow passages, just like in a sponge (hence the name spongy zone). The inner ring, the medulla, contains in lovebirds, black eumelanin and medullar cells. The black eumelanin is grouped around the medullar cells (also called vacuoles).

Simply put we can state that the eumelanin present in the medulla absorbs the daylight. This light is partially reflected back through the eumelanin. This reflection causes blue light to be created in the spongy zone. This blue light again passes through the cortex containing yellow psittacine [9]. The combination of blue light and yellow pigment causes us to view these feathers as green.

In a lot of species there is an autosomal recessive mutation present which blocks the psittacine in the feathers. Because of this there is no longer any yellow psittacine present in the cortex of these normal green feathers. As a result, we see only the blue light created in the spongy zone. In other words: the blue mutant.

The first birds of which we are certain that they have a blue mutation are budgerigars. In 1878 the first reports came in about a blue bird [8]. Yet it is only recently that scientists mapped the gene responsible for this blue mutation, in budgerigars [10]. Thomas F. Cooke investigated the DNA from 234 budgerigars, 105 of which were blue birds. In addition, he also scanned the DNA of 15 specimens from the wild. This is the data he started with. He managed to identify a gene which he called MuPKS (Melopsittacus undulatus PolyKetide Synthase) and which is involved in the creation of psittacine. He discovered that only one single mutated base in this gene causes the absence of the psittacine in blue birds. In other words, a simple SNP mutation (single nucleotide polymorphism).
Thanks to these findings we seized the opportunity to also look at this gene in lovebirds when scanning the genome of lovebirds for our ‘Agapornis Genome Study’ [11], [12]. We are not fully finished yet but the research is progressing slowly. Hopefully this will soon yield more information about the bl-locus in lovebirds.

As stated before, the blue mutation inherits autosomal recessive. With these combinations the male is always indicated first, but in an autosomal recessive mutation it does not matter whether the male or the female is the mutation. The results are the same.
I will provide a number of base pairings:

Green x blue:

100% green/blue

In this combination all young are phenotypically (visually) green but genotypically heterozygotic for blue or are green split for blue.

Green/blue x green:
50% probability of green
50% probability of green/blue

This pairing should be avoided. All offspring are green and only trial pairings will make it clear which birds received the blue factor and which did not.

Green/blue x green/blue:
25% probability of green
50% probability of green/blue
25% probability of blue

In this combination 25% is purebred green and 50% green, split (heterozygotic) for blue. Visually these green birds will not display characteristics which point to them being split. These birds are called ‘chance split’ in the aviculturist jargon.

Green/blue x blue:
50% probability of green/blue
50% probability of blue

This is a combination which yields clearer results than the previous. With these young you can be certain that all green offspring are split blue.

Finally blue x blue:
100% blue

There are not a lot of questions about these blue birds, although…. We see that in blue Indian Ringnecks [Psittacula krameri] the red colour of the beak is not affected. In blue Agapornis fischeri and the rest of the eye ring group this does happen. Hence some people wonder whether this is really blue in Indian Ringnecks? But the first question which appears to us is whether this beak colour in this species is indeed formed by psittacine? One should not forget that in a lot of species the colour of the beak is formed as a structural colour in the keratin layer on the beak. Maybe that is also the case for this red beak colour? This needs to be investigated first. Another question is whether the beak colour in these species might be controlled by a separate gene? As you can see a lot of factors play a part. Luckily the international agreements are clear: blue is the total absence of psittacine in (only) the feathers.


In 1998 the Australian Peter Bergman introduced in the article ‘Gene function in Yellowface Budgerigars’ the idea that more than likely there is more than one allele of the bl-locus which can produce ‘real’ blue birds [13]. He based his theory on his experiences with “yellow face” budgerigars. Yellow face budgerigars are the equivalent to what is known as the PPR mutations – partial psittacine reduction – aqua and turquoise in lovebirds. Within MUTAVI, Research & Advice Group we were one of the first to support his idea. This article can still be found on the MUTAVI web page [14]. This theory is usually called the ‘Bergman theory’. Just to be perfectly clear, not all budgerigar breeders agree with these assumptions.

I can imagine that when younger aviculturists, without any experience with budgerigars, read this article, it will be difficult for them to comprehend. Especially because the author, in accordance to the then prevailing line of thinking, still considered turquoise or yellow as dominant over blue and hence the use of the terms SF and DF. Nowadays, with the knowledge about alleles, this has been corrected and we use the terms heterozygotic and homozygotic.

Blue2, just like ‘normal blue’ in a homozygotic condition, present as a mutation on both alleles of the chromosome pair would therefore result in blue birds. Combined with the other blue allele, hence blue1, results in turquoise phenotypes. Of course, this is just conjecture, which is actually the original blue gene. Both mutated alleles/genes result in a blue bird in a homozygotic condition and we need both mutations/alleles to prove this. So, without a thorough understanding of the genetic background, it is difficult to determine which is type1 or type2.

As mentioned before it is generally assumed that both genes, both blue type1 and blue type2, are located on the bl-locus. Yet there are some indications that we might be dealing with epistatic genes. With an epistatic gene, the expression of a gene is activated, triggered or even oppressed by a completely different gene. Some of these genes have no effect on the phenotype, except when they are present combined with other specific genes [15], [16]. Because the results in pairings, both with alleles of the same gene and epistatic genes, usually provide an intermediate type and we cannot sufficiently substantiate this scientifically in lovebirds, we will, for now, in this article assume that we are dealing with alleles of the same gene. The rest is left up to the genetic nerds.

It might seem complicated but as you will soon notice it will become a lot clearer with a number of practical examples.

Practical examples of blue2

The breeding results of blue type2 combined with green are identical to green x blue1. We do not need to repeat them. The difference is noticeable in Agapornis fischeri when we pair blue2 with blue1. Whereas blue1 x blue1 always results in blue young, blue1 x blue2 is slightly different. Here we get a (visual) turquoise bird.

Blue2 x blue1: (or vice versa)
100% Blue1Blue2 (turquoise phenotype)

So, although we pair two visually blue birds, we still get turquoise young. This is due to the fact that both these blue types act as alleles of each other. Genetically speaking they are heterozygotic Blue1 and heterozygotic Blue2.

According to the international agreements we cannot use new names for these phenotypes [1, p. 255]. In this case we simply mention both mutations, in one word, concatenated (each name starting with a capital so they are easily recognizable). So, in this case, Blue1Blue2 or BlueBlue2. Because we are not actually certain which is blue1 and blue2 (both result in an intermediate type) it could theoretically also be BlueBlue, but I still prefer BlueBlue2 or Blue1Blue2. Genetic symbol is bl1/bl2.

Blue1Blue2 (turquoise phenotype) x Blue1Blue2 (turquoise phenotype):
25% probability of Blue1 (blue phenotype)
50% probability of Blue1Blue2 (looks turquoise)
25% probability of Blue2 (blue phenotype)

Blue1Blue2 (turquoise phenotype) x blue1:
50% probability of Blue1Blue2 (turquoise phenotype)
50% probability of blue1 (blue phenotype)

Blue1Blue2 (turquoise phenotype) x blue2:
50% probability of Blue1Blue2 (turquoise phenotype)
50% probability of blue2 (blue phenotype)

Blue1Blue2 (turquoise phenotype) x green:
50% probability of green/blue1
50% probability of green/blue2

Of course, there is no visual difference between green split blue1 and green split blue2. We can figure this out through trial pairings, but only if we are certain which of our blue birds are type1 or type2. In other words, another example of the importance of keeping accurate records of all breeding results and data.

Green/blue1 x blue1:
50% probability of green/blue1
50% probability of blue1 (blue phenotype)

Green/blue2 x blue2:
50% probability of green/blue2
50% probability of blue2 (blue phenotype)

Green/blue1 x green/blue1:
25% probability of green
50% probability of green/blue1
25% probability of blue1 (blue phenotype)

Green/blue2 x green/blue2:
25% probability of green
50% probability of green/blue2
25% probability of blue2 (blue phenotype)

If we are not certain about the genome of the birds, then we might expect some surprises:

Green/blue1 x blue2:
50% probability of green/blue2
50% probability of Blue1Blue2 (turquoise phenotype)

Green/blue2 x blue1:
50% probability of green/blue1
50% probability of Blue1Blue2 (turquoise phenotype)

Green/blue1 x green/blue2:
25% probability of green
25% probability of green/blue1
25% probability of green/blue2
25% probability of Blue1Blue2 (turquoise phenotype)

As you can see this can lead to a lot of confusion. Hence, I can understand the surprise of an aviculturist who recently wrote me that he had two turquoise young from two green birds split ‘blue’. From this he concluded that these were turquoise “DF blue” and … actually he was not entirely wrong ….

First a personal reflection: because the genetic background of this blue type2 is not yet sufficiently known, and we do strongly suspect that an epistatic gene could form the basis, it would not surprise me that with blue2 there are sometimes young birds which do display a minimum of psittacine in certain feather areas at certain points in their life. We regularly see with these epistatic genes that the expressiveness of a gene can vary slightly. That psittacine usually disappears with age. Again: this is only a suspicion!!! Just to be certain, everyone claims that their blue2 birds are completely blue, but scientifically speaking, I do not exclude this possibility.

Of course, we could ask the question, whether this is indeed possible/happens, whether the name blue2 would still be appropriate? This is indeed a valid question. To meet the requirements for the name blue, a mutation must have a complete psittacine reduction in the feathers. But because we do not have sufficient answers for this, we will use blue2 in accordance with the budgerigars. It is advisable to keep this in mind.


As we indicated at the start of this article a lot of people are convinced that their birds are real turquoise. According to them their turquoise Agapornis fischeri meets perfectly what we would expect from a ‘pure turquoise’ with regard to breeding results. Admittedly, the breeding results which I have seen do point in that direction but, with the knowledge that blue type2 also exists, there are still too many missing links to be 100% sure. Hence further research and test matings are needed.

In turquoise we see that contrary to blue there is still a part of psittacine present in the feathers. Whether this is a reduction of the amount of psittacine in the feathers or whether it is a different chemical composition of the pigment, remains unclear. We know that the red psittacofulvins in the feathers of parakeet species consists of tetradecahexenal, hexadecaheptenal, octadecaoctenal, eicosanonenal and a fifth unknown component [17], [18], the composition of yellow or orange psittacofulvins is unknown. We can assume that we have a different chemical composition, in turquoise and aqua we do not know for now.

The turquoise birds are easily recognizable. On the body they have a mix of green, blue green and blue feathers. This psittacine reduction is spread variably across the plumage in species where the turquoise mutation can be found. In Agapornis fischeri the turquoise phenotypes have a light orange/yellow mask and a clear green hue and spots on the blue wing coverts. The feathers on the body tend to be bluer.

I know about the stories that violet and/or the dark factor can mask this turquoise factor in Agapornis fischeri. Because of this genotypical turquoise birds might unjustly be considered as blue. Yet I can state with certainty that this is not possible considering the composition of the feather. Both violet and the dark factor modify the spongy zone of the feather and this has no effect on the cortex containing the psittacine. In addition, there is the mask colour, in (phenotypic) turquoise Agapornis fischeri there is still psittacine present. These feathers are of the ornamental type and are not influenced by violet or dark factor. Hence the psittacine present is always visible.

Some aviculturists recommend to check the birds for the presence of psittacine with a black light / ultraviolet light, a part of the yellow psittacine in the feathers. In several parrot species [19], [20] some yellow psittacine present is then clearly visible. All this is related to either the composition of the psittacine and/or the composition of the keratin of the feather. This could in theory prove the presence of a limited amount of psittacine in these turquoise birds. But for none of the domesticated lovebird species this reflection occurs, turning the test unusable for lovebirds.
Then of course there is the possibility that genetically speaking these factors (epistatic) block the psittacine, but since we already have gorgeous violet D turquoise phenotypes in Agapornis fischeri this can also be ruled out.

In any case I will list the breeding results with ‘real’ turquoise.

Turquoise x green:
100% green/turquoise

Green/ turquoise x green/turquoise:
25% probability of green
50% probability of green/turquoise
25% probability of turquoise

Green/turquoise x turquoise:
50% probability of green/turquoise
50% probability of turquoise

Turquoise x turquoise:
100% turquoise

Up until then a normal recessive mutation. The difference can be seen, just like with blue2, when this turquoise is paired with blue or another allele of blue. FYI, I SUSPECT that the results with both blue1 and blue2 with turquoise will be similar with regard to the phenotype and will therefore mention only blue in this paragraph to not complicate matters. You can simply replace it with blue1 or blue2.

Blue x turquoise:
100% TurquoiseBlue (turquoise phenotype)

TurquoiseBlue (turquoise phenotype) x green:
50% probability of green/blue
50% probability of green/turquoise

TurquoiseBlue (turquoise phenotype) x blue:
50% probability of blue
50% probability of TurquoiseBlue (turquoise phenotype)

TurquoiseBlue (turquoise phenotype) x turquoise:
50% probability of turquoise (purebred)
50% probability of TurquoiseBlue (turquoise phenotype)

TurquoiseBlue x green/turquoise:
25% probability of green/turquoise
25% probability of turquoise (purebred)
25% probability of TurquoiseBlue (turquoise phenotype)
25% probability of green/blue

TurquoiseBlue x green/blue:
25% probability of green/turquoise
25% probability of blue
25% probability of TurquoiseBlue (turquoise phenotype)
25% probability of green/blue

Green/blue x turquoise:
50% probability of green/turquoise
50% probability of TurquoiseBlue (turquoise phenotype)

Green/turquoise x blue:
50% probability of green/blue
50% probability of TurquoiseBlue (turquoise phenotype)

Green/blue x green/turquoise:
25% probability of green
25% probability of green/turquoise
25% probability of green/blue
25% probability of TurquoiseBlue (turquoise phenotype)

TurquoiseBlue x TurquoiseBlue (both turquoise phenotype):
25% probability of blue
50% probability of TurquoiseBlue (turquoise phenotype)
25% probability of turquoise (purebred)

It becomes a bit more complicated if we pair TurquoiseBlue1 x TurquoiseBlue2.

Then we get

TurquoiseBlue1 x TurquoiseBlue2:
25% probability of turquoise (purebred, but turquoise phenotype)
25% probability of TurquoiseBlue1 (turquoise phenotype)
25% probability of Blue1Blue2 (turquoise phenotype)
25% probability of TurquoiseBlue2 (turquoise phenotype)

In other words, genotypically there are four types possible if in addition to blue type1 and blue type2 there is also real turquoise. Visually all look turquoise (as far as we know). As you can see it is not so obvious to reach a conclusion by sight and to be able to distinguish the real turquoise from the combination types. The question remains how we can objectively determine that turquoise does indeed exist as an independent allele in Agapornis fischeri?

How to determine what is TurquoiseBlue1, TurquoiseBlue2, Blue1Blue2 or turquoise?

To figure this out we must pair the turquoise phenotypes with PUREBRED green. Note: purebred, meaning all split factors must be avoided. Most reason that it is already sufficient that the bird is not split blue (or one of its alleles), and that split pastel, dec or even split dilute is not so bad. Yet these must all be avoided. The more we learn about epistatic genes, the more we realize that the presence of certain genes can still have a (limited) effect on the phenotype. So, avoidance is the way to go.

We will list these possible outcomes one by one.

Possibility 1 the bird is indeed turquoise:
Turquoise x green:
50% probability of green
50% probability of green/turquoise

Possibility 2 the bird is TurquoiseBlue1
TurquoiseBlue1 (turquoise phenotype) x green:
50% probability of green/blue1
50% probability of green/turquoise

Possibility 3 the bird is TurquoiseBlue2
TurquoiseBlue2 (turquoise phenotype) x green:
50% probability of green/blue2
50% probability of green/turquoise

Possibility 4 the bird is Blue1Blue2
Blue1Blue2 (also turquoise phenotype) x green:
50% probability of green/blue1
50% probability of green/blue2

Now we must determine which birds are actually split turquoise. The only feasible possibility is to pair these green split birds with blue. The aim is first and foremost to extract the birds which are definitely split blue. Therefore, it is important that you know for certain whether your blue birds are blue type1 or blue type2 (stemming from a turquoise line). In these first combinations we will assume that this is known.

Is the bird green/blue1, we will pair it with blue1. Then we get
Green/blue1 x blue1:
50% probability of blue1 young
50% probability of green/blue1

Because of the presence of blue young in the nest we are certain that these birds are split blue.

Is the bird green/blue2 then we must of course pair it with blue2 and we get
Green/blue2 x blue2:
50% probability of blue2 young
50% probability of green/blue2

It goes without saying that these green birds split blue (both blue type1 and blue type2) must be excluded. We should only continue with green birds split turquoise.

AGAIN: It is of the utmost importance that you are 100% certain whether your blue birds are type1 or type2. If you are not certain you might reach the wrong conclusion. For example:

If the birds are green/ blue1 and we pair them with blue2, we can get:

50% probability of Blue1Blue2 young (turquoise phenotype)
50% probability of green/blue2

Or green/blue2 x blue1:
50% probability of Blue1Blue2 young (turquoise phenotype)
50% probability of green/blue1

We might incorrectly conclude that het green bird is split turquoise and will therefore produce TurquoiseBlue young. Of course, this must be avoided. So do pay attention here.

If the bird is green/turquoise and we pair it with blue (both type1 and type2) then we get

50% probability of TurquoiseBlue young – these will have a turquoise phenotype
50% probability of green/blue

With this last pairing, at least, we know that the green parent bird is split turquoise. These can then be paired together.

Green/turquoise x green/turquoise:
25% probability of green
50% probability of green/turquoise
25% probability of turquoise

If, after you have correctly followed the preceding steps (and you are certain that the green bird is split turquoise), turquoise Agapornis fischeri are born instead of blue ones, then you can be certain that there exists a real turquoise type. The aviculturists who think that they have the real turquoise, should investigate this thoroughly. We would love to hear your results.

Shows – BVA-International nomenclature

For the nomenclature within BVA-International (and other organizations) there is a rule that combinations of alleles, so for instance PastelIno, PallidIno, etc are not accepted for the competition class. Hence TurquoiseBlue is also not allowed. We were 100% convinced that most turquoise birds at shows were heterozygotic (combinations with blue), but because we could not determine conclusively whether a bird was TurquoiseBlue or turquoise all turquoise phenotypes were accepted during the Masters.

We now know for certain that within Agapornis fischeri blue type2 does exist and that Blue1Blue2 also results in turquoise phenotypically. In addition the presence of turquoise in Agapornis fischeri is no longer a fact but an open question. So we have a problem: if it cannot be determined that real turquoise does exist, then we must conclude that all turquoise phenotypes in Agapornis fischeri are Blue1Blue2 and hence they cannot be accepted in the competition class. I can imagine that a lot of aviculturists will object and I can understand them.

I propose to give this turquoise Agapornis fischeri the benefit of doubt and to continue with the assumption that real turquoise does exist, for now. Because of this we will indicate turquoise as *turquoise* in the BVA-International nomenclature for the eye ring species, i.e. Agapornis fischeri, Agapornis personatus, Agapornis nigrigenis and Agapornis lilianae. The presence of this asterisk (*) in front of and behind the name indicates according to international agreements that the phenotype might not fit the genotype. This way they can be judged in the show class as a phenotype. Hopefully the future will provide more certainty about the existence of turquoise. Then we will be able to set up final agreements with our partners.

Comparisons with turquoise Agapornis roseicollis and Agapornis personatus

I am aware that aviculturists of Agapornis roseicollis are now probably wondering whether in addition to turquoise, more accurately the selection type (*blue*), there might also be a blue type2. In Agapornis roseicollis we are certain that there is a real turquoise. We have all also seen, step by step, this evolution whereby, probably aided by some anticipation, the turquoise birds became ever bluer. This phenomenon has also been observed in various Australian parakeets and is therefore not unique to Agapornis roseicollis. If we compare this to what happens for Agapornis fischeri we do see some differences.

If we start from the hypothesis that in Agapornis roseicollis there is also blue type2, then with the combination of blue2 x turquoise we would probably get more clearly delineated ‘yellow mask’ types. As far as I know these do not exist yet. The spectrum of turquoise ranges from normal turquoise to nearly entirely blue, with all possible intermediate types. In all types we see a clear increase of the psittacine through the years. This would be the opposite to what I have heard about blue type2.
But anyway, it is important to review everything with an open mind and to leave nothing up to chance. “Multi multa; nemo omnia novit. Nemo est omniscius
Not in the least the turquoise Agapornis personatus. Considering the fact that the first turquoise birds came from a communal aviary containing both turquoise Agapornis personatus and turquoise Agapornis fischeri (and the resulting mixes) we can state with certainty that these turquoise types have the same genetic cause in both species. What is applicable to turquoise Agapornis fischeri will also be applicate to Agapornis personatus.

In January 2017 we received an email from Wicus van der Merwe with a number of pictures of a nearly completely blue Agapornis fischeri, but with a clear orange forehead and a number of small green spots on the wings and surrounding the cloaca. For the remainder the bird was completely blue.

I quote from his email: “I bred seven of these birds from three different pairs of proven Green/turquoise. I say these are proven green /turquoise because in the previous year each of these birds bred TurquoiseBlue chicks with blue series partners. Unfortunately all breeding was done in large cages with multiple pairs. The results of the breeding is obviously green fischeri and green /possible Turquoise and then the blue birds (25%) with orange forehead as per the pictures included. Some of these green offspring have already been bred with blue series and produce visual Green or TurquoiseBlue offspring and no blue offspring. The blue (Orange head) offspring have not bred yet.”.

As always, we advised him to pair these birds with pure green and to continue with the young. One of the first problems we came across is that we were not certain about the parents. If these birds are housed in groups, it is not unforeseeable that a female will mate with multiple males. So, it is very interesting to know which other birds (and possible partners) were also present in the cage.

Another obstacle; the real genome of the green birds. It was concluded that they are green/turquoise because past pairings with birds from the blue series always resulted in turquoise phenotypes. Phenotypes which were considered to be TurquoiseBlue. But is this really the case?
If you have read the previous section, then it is also possible that these green birds are for instance green/blue2. Combinations with blue also result in turquoise phenotypes. Therefore, I had to also consider this train of thought, so hopefully you can see the problem we were facing then. Ideas which occurred to me and which were also reported to him where that this might be the homozygotic type of turquoise? Or an AquaTurquoise phenotype, but then there had to be aqua birds or at the very least green birds split aqua about.

Unfortunately for us, this info and photos had also been spread through social media and as you can already guess: eight persons contacted us almost immediately. Some even sent the pictures Wicus had posted and claimed that this was their bird and asked for information. Others told us that they have/had these birds. Yet others send us various combinations which had resulted in these phenotypes in the past. Over the next few days we received similar reports. As you can understand this did not simplify matters for us. One can question the correctness of these data, but each of the statements had to be taken seriously and will be part of the research. We cannot leave anything to chance.

Checks in our archives brought to light that in 2004 a, in the meantime deceased, Hungarian breeder once reported that he had a blue Agapornis fischeri with an orange forehead. He had bought the bird in a specialty store. The origin of the bird could not be traced back. There was no offspring because the bird was found dead in its cage two weeks later.

In October 2017 I was in South-Africa for a seminar and Wicus was also present. He had brought a pair of his blue birds with an orange forehead band. He indicated that the birds develop this orange forehead band only after a few weeks. He told me that in 2013 he had bought the collection of Agapornis fischeri of a deceased aviculturist. One of the birds from this collection stood out because of his more or less turquoise appearance. Unfortunately, this bird died. With his knowledge of turquoise and blue, he could trace the split birds in this collection using trial pairings. He then started with those. These pairings then resulted in these birds with the orange forehead band. He suspects that this bloodline probably has nothing to do with the turquoise Agapornis fischeri coming out of America.

I again gave him some suggestions to set up trial pairings. You must understand, that, with the idea of the possible existence of a blue type2 already in mind back then, I was not able to find a sound answer straight away. It was requested to set up trial pairings and to keep us informed about the further developments. Which is the standard practice for these matters. I know that this does not make me very popular, people do expect answers, preferably yesterday already. But again, everything must be investigated thoroughly, quick answers and guesses can and should not be expected from us.

In May 2018 he also mentioned in his article about blue type2, these blue birds with the orange forehead. This article contained new information about the completed trial pairings. Wicus called these blue phenotypes South African turquoise (orange fronted) or SA OF turquoise. He wonders whether this phenotype might be the real turquoise? A valid question, but the answer is not so easy.

To make is a bit simpler I would first like to talk about the name SA OF turquoise. If we follow the international agreements, we know that mentioning locations/countries in a name is not desired. In addition, orange fronted refers to an effect of the phenotype in Agapornis fischeri. Should this mutation for instance appear or be transferred through genetic introgression (transmutation) to for instance Agapornis personatus or Agapornis nigrigenis, then the characteristic would not be applicable in the same degree or even at all. And lastly the mention of turquoise is not 100% correct either according to us. In turquoise, as we know this in other species, both lovebirds and other parakeet species, we see that these have a lot more psittacine in the feathers than this SA OF turquoise. Here we are dealing with a 95% blue bird, some psittacine hue on the forehead, cloaca and wings. So, if it is not a turquoise, then what is it?

Well, we must first and foremost realize that there are multiple genes which can influence the creation of the psittacine. The best proof is the orange face and pale headed mutation found in lovebirds. In this mutant the creation of the psittacine is also affected, even though this has nothing to do with blue.

The chance that a new allele is created on the bl-locus or that the blue type is modified because of an extra epistatic gene, is very real. So, in consultation with a number of scientists I looked for a more fitting working name for this phenotype and ended up with sapphire. Sapphire is a gem which is mainly blue, but in this blue colour sometimes ‘polluted’ iron particles can be found. With some imagination we can compare the limited amount of psittacine present in the blue feathers to this. Additionally, the name starts with SA which might be a small, subtle, indication of South-Africa.

Of course, this name is placed between asterisks for the time being, hence *sapphire*. This way a separate series can be foreseen, if required, in the BVA-International nomenclature. So, these birds can participate in shows. Their owners can therefore continue setting up trial pairings with these birds.

We then continued, based on the data in the article and some reports from other breeders who (claimed) said to own these birds. What we mention here is only preliminary. None of the results mentioned in the article have been verified by us.

I will list them:

  • Combination of *sapphire* x *sapphire*. The precise number of young is not known, only that some were *sapphire*.
  • According to the author the combination of green/ *sapphire* x *sapphire* resulted in green and *sapphire* young.
  • *sapphire* x blue resulted in all *SapphireBlue* young. The photo I saw of this bird showed a beautiful nearly entirely light blue bird with a light yellow/orange mask. For the sake of clarity, this was a photo, and only one specimen.
  • Tests with *SapphireBlue* x *SapphireBlue* resulted in *sapphire*, blue and *SapphireBlue* young.

If these tests are all accurate and nothing has been missed, then it does indeed appear to be a separate allele or an epistatic gene, but again this must be investigated further. The results mentioned fit perfectly in the known survival patterns. Unfortunately, these are not yet sufficient to reach a definite conclusion. Surprises are still possible, we should not leave anything up to chance. The trial pairing with green and the resulting young is also an absolute necessity and these are currently still lacking in the list.

For now, it is assumed that this *sapphire* has a different origin than the turquoise and/or blue1 type coming out of America. However, there is someone who claims to have bred a young from the combination green/blue2 x green/blue2 – i.e. splits of birds with an American origin – which developed an orange forehead band after a few weeks. This might indicate that: either the two types were mixed up by this breeder and he had nevertheless paired birds from the African bloodline, or there is indeed an epistatic gene which is present somewhere in the genome of Agapornis fischeri.

Once this epistatic gene is present in a bird together with blue type2, the combination could be anchored in the genome through selective breeding programmes. So, some extra work to research this properly and this can only be done by pairing green/*sapphire* with green/*sapphire*. If these are linked or epistatic factors this will surface sooner or later. Once the trial pairings with pure green birds are confirmed by multiple breeders, and we still get *sapphire* young we could consider removing the asterisk in the name. As I understand it members of the SA lovebird club have also set up a joint project to develop these trial pairings. This can only make us happy.

We have also recently taken blood samples from these *sapphire* birds. We hope to scan these samples during the next DNA sequencing. This might confirm or deny the fact that this is also an allele of the bl-locus. This is not a certainty, no matter how easy it might look. Besides, during a recent conversation with Henriette van der Zwan about this research, it soon came up that the correct weeding out of only the blue and the PPR types is sufficient for at least two doctoral dissertations. So, if anyone is up to it ….
Hopefully multiple breeders can start with this information and friends: do not forget combinations with the pure wild type. We definitely need these.

*Teal* Agapornis taranta

A few years ago, the Belgian breeder and BVA judge Eric Pauwels found a possible ‘PPR’ A. taranta in a nest of two wild type birds. The bird had a reduction of about 15-25 percent of the psittacine on the chest, the wings had a lesser reduction. At first, he thought he was dealing with a modification, but even after the adolescent moult the bird retained its ‘parblue’ colour. The bird, a female, was paired with a wild type male two years later. The results were as expected green. Two years later when the F1 generation was paired, there were again ‘PPR’ young in the nest. Later on, birds with similar characteristics also surfaced in Italy and Spain. The birds also came from a Belgian bloodline.
This is proof that we are dealing with an autosomal recessive mutation. The phenotype of these A. taranta is not identical to turquoise A. roseicollis, yet I did think that we could possibly be dealing with a turquoise. But eventually we did realize that this was probably a new PPR mutation and we proposed to call this phenotype *teal* for the time being, which basically means green-blue.

Normally mutations of the bl-locus require a crossing-over to combine these with a dark factor. The breeding results of this *teal* with dark factor appear to be occurring without a crossing-over. This might indicate that this phenotype is caused by a different gene than the normal blue.
Time will tell.

Blue x blue is green?

And lastly in this ‘article’ about blue and related matters, I would also like to mention this. If for no other reason than to show that blue and everything related, is not so easy.

I was a guest in Spain in May 2017 for a study day. In the afternoon Antonio Pradas came by with a number of birds. The man was faced with a riddle. He had a pair of Agapornis personatus D blue x SF violet blue. To his surprise he also had a D green and a normal green young in his first nest in addition to two D blue, one SF violet D blue. All young were of course male. Later on, these birds were paired with another blue partner. The female then had 100% blue offspring, the male however, again had blue and green young.

We all know that blue x blue always results in blue young, so everyone ridiculed Antonio when he announced this through social media. He was of course immensely relieved when I could confirm in front of quite a lot of aviculturists, that this was indeed possible.

How is it possible? Well the answer must be sought in what we currently call epigenetics. In epigenetics the scientist researches the question why a gene gets a different expression, without a change in the sequence of the bases. In other words, the same DNA but a different type [21]–[23].
I definitely do not want to complicate it too much, but one of the most likely trains of thought is that genetically the male is probably green/blue. Remains the question how it is possible that he is phenotypically blue?

The answer is probably: epigenetic gene regulation. This can occur through various mechanisms, but the most likely are methyl labels [24]. Simply put: during the methylation process (methylation takes care of the maintenance and repair of the DNA in all cells and tissue in the body). During this process a methyl group can bind to the DNA. This can affect, or even stop the operation of the gene concerned. Methyl labels are very stable and usually remain during the cell divisions. Only during the meiosis (in the male the creation of sperm cells and in the female the creation of egg cells) will these methyl labels be removed and will the unmutated blue gene again be transmitted in the reproductive cells.

Complicated? Not really, just keep in mind that this is a green split blue male. An extra factor (methyl label) eliminates the functioning of the unmutated bl allele in the DNA of the bird (heterozygotic for blue) and the bird actually becomes homozygotic blue. Because of this the blue bird cannot produce psittacine and he is therefore blue. But his DNA (genome) is just a normal green split blue. During the creation of the sperm cells the label is removed and as a result an unmutated bl gene is passed to the young.

How and why this happens and what is the cause remains a mystery for now. Daily new research is conducted, new answers are found and definitely new questions. It never stops.

I hope you now realize that in genetics nothing is as simple as it seems or that the aviculturists would want and foremost there are still a lot of questions. Genetics is much more than just genes, alleles and loci. Those who have a different opinion have made all this effort for nothing.

Voila, my friends, I hope I have dealt with all your questions and assumptions/suspicions. Hopefully things are clearer now.



[1] D. Van den Abeele, Lovebirds Compendium, 1ste dr. Warffum- The Netherlands: About Pets, 2016.
[2] D. Van den Abeele, “Agapornis fischeri turquoise”, Agapornis, nr. 2, 2012.
[3] Read “Responsible Science, Volume II: Background Papers and Resource Documents” at .
[4] G. E. Hill en K. J. McGraw, Bird Coloration. Volume 1. Mechanisms and measurements. Harvard University Press, 2006.
[5] R. Hooke, “1665 Micrographia”, Lond. Martyn Allestry, 2003.
[6] J. Dyck, Structure and spectral reflectance of green and blue feathers of the Rose-faced Lovebird (Agapornis roseicollis). Munksgaard, 1971.
[7] R. Morelli, R. Loscalzo, R. Stradi, A. Bertelli, en M. Falchi, “Evaluation of the antioxidant activity of new carotenoid-like compounds by electron paramagnetic resonance.”, Drugs Exp. Clin. Res., vol. 29, nr. 3, pp. 95–100, 2003.
[8] H. van der Linden, Grasparkieten. Van Spijk, 2002.
[9] R. O. Prum, R. H. Torres, S. Williamson, en J. Dyck, “Coherent light scattering by blue feather barbs”, Nature, vol. 396, nr. 6706, pp. 28–29, nov. 1998.
[10] T. F. Cooke e.a., “Genetic Mapping and Biochemical Basis of Yellow Feather Pigmentation in Budgerigars”, Cell, vol. 171, nr. 2, pp. 427-439.e21, okt. 2017.
[11] H. van der Zwan, F. van der Westhuizen, C. Visser, en R. van der Sluis, “Draft De Novo Genome Sequence of Agapornis roseicollis for Application in Avian Breeding”, Anim. Biotechnol., okt. 2017.
[12] H. van der Zwan, “Agapornis Genome Study – Information for breeders on the Agapornis genome PhD study”, Agapornis Genome Study – Information for breeders on the Agapornis genome PhD study. [Online]. Beschikbaar op: [Geraadpleegd: 11-sep-2017].
[13] Peter Bergman, “Gene function in Yellowface Budgerigars”, Australian birdkeeper, 1998.
[14] P. Bergman, “Gene function in Yellowface Budgerigars”. [Online]. Beschikbaar op:
[15] E. T. Domyan e.a., “Epistatic and Combinatorial Effects of Pigmentary Gene Mutations in the Domestic Pigeon”, Curr. Biol., vol. 24, nr. 4, pp. 459–464, feb. 2014.
[16] H. J. Cordell, “Epistasis: what it means, what it doesn’t mean, and statistical methods to detect it in humans”, Hum. Mol. Genet., vol. 11, nr. 20, pp. 2463–2468, jan. 2002.
[17] R. Stradi, E. Pini, en G. Celentano, “The chemical structure of the pigments in Ara macao plumage”, Comp. Biochem. Physiol. Part B, vol. 130, nr. 1, pp. 57–63, 2001.
[18] K. J. McGraw en M. C. Nogare, “Distribution of unique red feather pigments in parrots”, Biol. Lett., vol. 1, nr. 1, pp. 38–43, mrt. 2005.
[19] F. Hausmann, K. E. Arnold, N. J. Marshall, en I. P. F. Owens, “Ultraviolet signals in birds are special”, Proc. R. Soc. Lond. B Biol. Sci., vol. 270, nr. 1510, p. 61, 2003.
[20] M. L. Berg en A. T. D. Bennett, “The evolution of plumage colouration in parrots: a review”, Emu, vol. 110, nr. 1, pp. 10–20, 2010.
[21] E. Jablonka en M. J. Lamb, “The Changing Concept of Epigenetics”, Ann. N. Y. Acad. Sci., vol. 981, nr. 1, pp. 82–96, jan. 2006.
[22] A. Petronis, J. L. Kennedy, en A. D. Paterson, “Genetic anticipation: fact or artifact, genetics or epigenetics?”, The Lancet, vol. 350, nr. 9088, pp. 1403–1404, 1997.
[23] R. Jaenisch en A. Bird, “Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals”, Nat. Genet., vol. 33, pp. 245–254, mrt. 2003.
[24] Richard C. Francis, Epigenetics: How environment shapes our genes. WW Norton & Company, 2011.