Russell thank you very much for your post! I found it most informative and leading to tons of questions, mainly surronding maternal and paternal traits. How can I learn more about what traits are maternal or paternal and which traits are recessive... or is this information only garnered from years of observation and reading... Is there one good publication that would detail traits?
I'd been hoping for a list like this to come along, and for Bob Russel to open up a bit on it. I was wondering what percent of your colonies were selected for breeding, or if it was done strictly on traits shown? The best half of your group will have more genetic diversity, but less progress toward your selected trait than will the best quarter of your group, and so on as the selectivity increases. What level of selectivity do you prefer? Or is the percentage of selected breeders in the group not as important as the degree to which the desired trait shows up?
Not to answer for RRussell, but once you stabilize a given set of traits, it gets down not to selecting based on a specific trait, but rather to excluding a specific line because one of the traits is not adequately expressed.
Lets say the set of selected traits are:
3. low production of drones
4. high egg laying rate
5. low level of brace comb
Now you score a colony on a scale of 1 to 10 where 1 is worst and 10 is best for the given trait:
1 - 8
2 - 10
3 - 7
4 - 9
5 - 5
Presuming most of the colonies being rated gave comparable results, this colony might be excluded from breeding by the low score on brace comb. This colony exhibits excellent gentleness and high egg laying rate but has a slight tendency to swarm. If this colony were being used for breeding, it could be used to supply drones which would be mated to queens reared from a line that scored higher in non-swarming and brace comb.
A good selection program would obviously use many more traits than the subset above.
DarJones - The universe is not only stranger than we imagine, its stranger than we can imagine - JBS Haldane
Out of the list of traits which are on the queen side and which are on the drone side?
Well for color, either or.
Couple things to think about though, if you breed italian queens then open mate with mostly dark drones, you'll get bees af many colors in the hive as there will be many fathers. But the drones will be italian. So If you are looking for long term change to a dark bee breed dark queens, which will produce dark drones, to continue the dark color into the future.
"We don't need no education" (Pink Floyd) - Yes you do, you just used a double negative.
A boiled down synopsis of color inheritance in bees is that there are 2 major genes affecting color and 6 or 7 modifying genes. Yellow from the Y gene is dominant over black. The Scutellata gene AC is dominant. From that point, segregation gives a range of colors from pure yellow to solid black.
Brother Adam made the point that color should be the least of the worries of a bee breeder. I would add a caveat that there is a survival advantage for black bees in colder climates.
DarJones - The universe is not only stranger than we imagine, its stranger than we can imagine - JBS Haldane
As a side note, Im not looking to make my queens the next super bee for mass market. Im simply wanting to raise a few for myself and have fun doing it while I learn.
Volunteer - I would consider starting to keep meticulous records and devising a grading system for individual traits, as Robert Russel hinted at. Since he has not divulged all the male-passed versus female-passed traits on his 200+ trait tracking queen sheets, one would have to go about AI breeding both ways (drone mother trait and queen mother trait) and grading the resulting offspring queens to determine which was passed by whom. Color is a side note, and as Old Timer noted, if you are open mating, you'll get as many colors as fathers (and little valuable info).
You might not be trying to make the next super bee, but you'll learn more than by guessing from colors.
I find this thread very interesting but I do have a couple of questions.
I am interested in the education that is behind all of this. I realize the list of people can get very long so that question may not really be answerable. I am mainly thinking of the list presented here how ti was compiled and by whom. More importantly those that are behind the methods of selection.
At one point the number 200 has come up in regard to traits that can be selected for. Are colonies evaluated on all 200? IF so I find it hard to believe that any effective selection can happen. As an example if you have a colony that is the best you have ever seen on 1 trait. but losses ground on 199. how will you ever make forward progress? It is my understanding that if you have 200 traits you probably need 200 breeding programs. each one focused on a single trait or at best a set of traits that are closely linked to each other. That trait is developed to an extreme degree and then those specimens are crossed with specimens from another breeding program in an attempt to combine them and improve both characteristic in the new line. Even this usually results in a loss of quality in both traits although an overall improvement from past generations.
As an example you have colony 1 that produces 50 lbs of honey a year and swarms terribly. You then develop from this hive line A and line B. Line A is bred and selected for honey production. Line B is selected for lower tenancy to swarm. After 5 generations you then breed line A back to line B and the result is colony 2 which does not produce as much honey as line A or is as good in regards to swarming as line B but it does out preform Colony 1 in both regards. And the bees are mean as all get out. So once again you split colony 2 into lines A and B still focused on the same trait while you also add a line C that addresses temperament.
Now think this out across 200 traits. You can't select a colony based upon temperament one generation and then select from that line for honey production the next generation. What progress have you made and you just threw it away when you changed what you are selecting for.
Even worse. Unless the selection is continued all traits are lost anyway. They will be lost very quickly. This means that even if you managed to selectively breed the mite immune bee. you could sell those queens and within just a few generations that immunity would be lost. The Immunity trait would have to be continuously supplied from the original breeding program and that only by continuing the focused breeding and selection.
As a real life example found in the link I posted earlier. Roosters through very intense breeding and selection have been developed that grow 12 inch long or longer saddle feathers. This alone was considered to be impossible. Not only that but these impossible feathers are being produced with qualities that make them suitable for some of the most demanding qualities a feather can be asked to have. Consistent length of fibers the length of the quill. fibers that are not to course and not to fine, color. A quill that is not to think as to be fragile but not so thick it cannot be wrapped around the shank of a hook. not only will the quill wrap around the shank of a hook but it will lay flat and not twist in doing so. As you can see even that list gets pretty long. But if the selection pressure was removed for even a few generations. every one of those roosters would revert back to producing short spear point shaped saddle feathers that are useless for making dry flys.
Okay I wrote all that to give you a feel for how I understand thing. but based upon that understanding I question if breeding is the answer. or just leaving the bees alone and letting them revert back to being just bees? Not making a claim here. just making a case for a train of thought that basically says it is the meddling of man that is the root of the problem. maybe not meddling is the answer. Obviously far to simple but it is the basic idea.
That is a pretty badly flawed summation. Here are some ways to change your mind and start to see how breeding works.
When working with an organism, you are working with a multitude of traits. These traits rarely can be broken down to single genes that control the trait. Consider that each gene is on a chromosome which can be thought of like a chain. While crossing over does occur, for most purposes chromosomes tend to stay together. This means that if two genes are relatively close on a chromomsome chain, then they tend to stay together and are inherited in a linked fashion.
You mention 200 traits that can be selected which was supposedly given by RRussell as the checklist he works with. I don't have 200 traits, but if you go down to the fine details, you could easily expand the list I provided into subtraits that could number in the thousands. Consider non-swarming as a trait.
1. Does not build queen cells when crowded
2. Tends to store honey away from the brood nest
3. Tends to build comb very early in the nectar flow.
4. Does not start drone cells early in the season
Now you have a single high end trait broken down into 4 sub-traits that contribute to non-swarming. Each of those three traits can now be broken down further.
Your example of splitting a breeding line into lines A and B is off just a bit. I would normally select two existing lines that exhibit the traits required and initiate the breeding program with existing selected stock. A highly productive line that tends to swarm would then be crossed with another line that may not be as productive but rarely swarms. Selection would focus on identifying colonies that don't swarm yet still produce a large crop of honey. Please note that what you are selecting for is hidden in this example. You are not selecting for high production or non swarming, you are really selecting for bees that produce the largest crop of honey with the least cost of management. In other words, bees that are efficient when managed.
Your real life example of the roosters is not valid. That is a highly selected population that has already passed the tipping point where selection is no longer a case of trying to include new traits but rather is trying to exclude negative traits. Even if you discontinued selection with the population, the traits for producing high quality hackle feathers would not disappear, they would just tend to median range of expression for the species, population, and traits in question. So long as the population of birds producing high quality hackle feathers is maintained separate from all other chickens, they will tend to retain the selected traits.
You state that the traits can be quickly lost. Point of fact is that once traits are selected and increased in a population, they never entirely go away unless the population is wiped out. Consider the dog. Does a dog behave like a wolf? Yet the dog was selected out of a population of wolves.
Now I will give a real life example for honeybees. Beekeepers figured out nearly 100 years ago (see Rothenbuhler) that some bees are more hygienic than others and that hygienic bees are more resistant to disease. When varroa mites hit, we found that a form of hygienic behavior controls uncapping and removal of mite infested brood. Over the last 20+ years, enough breeders finally got on the band wagon of selecting for hygienic behavior that the entire United States now has an increased population of hygienic bees. At the same time, natural populations of feral bees have been under pressure from varroa mites so that they too have developed increased hygienic behavior because the only bees that live are the bees that control varroa. The result is that we are now at the tipping point where the entire U.S. population of honeybees is becoming hygienic. This is an example of a group of traits increasing dramatically in a population where the selection pressure is extreme and unremitting. All that is needed now is for the remaining beekeepers who treat for mites to stop treating.
Just in case you don't believe it, watch for articles in Bee Culture in the next few months. There will be loud screaming and caterwauling as various "scientists" debate the issue. Pull up a chair and watch the show!
Take a look at this link and see how many breeders are involved in propagating VSH in their bees.
DarJones - The universe is not only stranger than we imagine, its stranger than we can imagine - JBS Haldane
It appears to me that the selection effort with the bees is not so much trying to develop something new in the bees. Such as is the case with the chickens. But to make a conscious effort to examine colonies, recognize certain behaviors or traits that are considered beneficial. Then get those particular colonies widely spread as possible. No process of eliminating negative traits?
In very simple terms. Find the better bees and help them take over.
In the artical I linked to on the chickens it is even mentioned that those methods quickly become a matter or selecting against (Finding negative traits) rather than selecting for (Finding positive traits). In the case of bees the goal is the proliferation of colonies that have been selected for and that the best colony would be graded according to all positive qualities.
It makes since in my head so I am not sure I am getting it in type the way I am understanding it. I do see a clear distinction between what I think you have described and what I did.
Actually that sounds a heck of a lot easier than the chicken stuff.
Last edited by Daniel Y; 12-31-2011 at 07:55 AM.
A very good thread, and a very good read. It took me a few days to get through it, and to be honest I'll probably have to read it a few more to fully grasp everything that was in here.
There is one thing that I just can't seem to get a grasp on. Everyone appears to be stressing the importance of taking uniform standard, periodic and accurate records of all the traits that the queens possess. Only by having all the data on all the traits that the queen's offspring show can you fully evaluate the genetic material's potential for future breeding. That makes sense.
What I'm a little confused about is what I should be looking for, how I should be looking for it, and how often I should be looking for it. I can come up with a list of 20 things I should be checking for, that isn't the hard part. But if some of those traits are connected, I'm actually only checking up on 10 things, even though I am collecting data on the others, if any of that makes sense. But for the time being, too much data is less of a problem than not enough. So how do you create a uniform standard in which to compare multiple hives? How do you objectively classify a hives "gentleness"? Or it's comb building abilities? Do you count the number of stings you get each visit, and assume the open hive popped you? How much of that comes down to operator error? How much propolis is too much? If you made a scale of 1-10, am I just arbitrarily assigning a number? What's to ensure that my "7" in one hive was actually less propolis than my "8" in another hive two weeks later?
I can see the ability to compare one trait over multiple hives, if you are only observing one trait, objectively through a means of a test. For example, placing a propolis trap on a hive and timing how long it takes them to fill it up would measure the hive's ability to collect propolis. But that would be sacrificing several other selected traits to spend a week testing one trait. By the time the summer is over I may be able to test each characteristic, but not each characteristic for each period throughout the summer. Plus, who is to say that my constant "testing" isn't changing the results? Maybe a hive didn't make it through the winter because I kept toying with them. Or maybe one hive doesn't collect enough pollen on week 2 because I caused all their foragers to go out and collect propolis in my experiment on week 1.
But, even assuming that I can wrap my head around that, how often should I really be checking for this stuff? Once a week?
Here's a good example. Lets say we are only looking at a queens ability to lay eggs. The more eggs the better (within some reason). This will vary based on the time of the season. Spring is likely to produce more eggs than fall. So, I obviously need to check for eggs laid in spring and fall, as well as periodically throughout to tell when they start to dial back. The only objective way to tell this is by taking the total square inch surface area of the open brood, dividing it by the time the open brood took to develop, to get an average square inch surface area laid per day. Doing it that way, I would have to re-check every 28 days, and re-measure the surface area of each hive every 28. If I have five hives, not bad. If I had 100, alot worse. Plus, doing it once every 28 days will only give me an average number for the preceding month. Is that acceptable? If I did it every 14 days and divided by two, it would be more accurate.
Plus, the more I open the hive the more potential I have to change the outcome. The more dis-oriented the bees get after each inspection, the less productive they are, and the more "reorganization" and "recleaning" needs to take place.
But perhaps I'm WAY over thinking this (as, from time to time, I'm known to do). Does any of this sound like actual, logical concerns?
In the end:
1. What traits do you look for?
2. How do you measure those traits?
3. How often do you measure those traits?
Michael Palmer says something to the effect that a third of hives are great, a third are okay, and a third are junk. You're gonna know pretty quickly if a hive is very mean, and it's probably easy to tell if it's pretty mean, or even kinda mean.
I guess what I'm saying is unless you're searching for some obscure trait, the standouts should be fairly easy to find both in positives and negatives. But this is just conjecture, I've never had that many hives. For me, it's fairly easy to see which hives I want to breed from. This last year, there was only one of five. Currently, it's looking more like three of ten.
Somebody tell me if I'm off base, I'm trying to learn this as much as anyone.
Good post SpecK. You pose some good questions. As far as gentleness goes that is a tough one. I have marked hives that seem exceptionally mean one day only to find them a few weeks later not even remotely aggressive. Determining aggressiveness is so dependent on conditions such as weather, honey flow or even nightly skunk disruption that you have to judge them more on an average. Honey production also can be influenced by drift, tendency to rob and hive maturity just to name a few factors. My advice is to garner as much info as possible and take one long look at your best candidates in early spring and make your best guess. In the final analysis what determines the success of your new queens may be determined by how well mated they are more than any other factor.
"Ve are too soon olt und too late schmart."- A nameless German philosopher
Selecting for some 200 traits as has been suggested isn't going to happen. Not in most operations. Even at 20 traits, as has also been suggested, the amount of work involved limits what can be done. And as SK says, how many traits are connected to others...like spring population, brood quantity, and honey production?
I don't keep track of much, only the things that matter to me in my operation, and only traits that I can measure in some way.
I keep bees in Vermont. What do you think is the mosdt important trait I want in my bees. Wintering, yes? If the bees don't winter, there will be nothing to select from anyway. I use two lines on my yard sheet for the ability of a colony to winter properly.
1. Population in early spring. You know that day you open the inner cover for the first time? Brood rearing is just gettint underway. Not much flight yet...just some cleansing flights through the winter. Not enough flight to effect the population...colonies lose population with the beginning of spring foraging as the loss of old bees is greater than the replacement by emerging brood. How large is the cluster compared with what went into winter? A colony whose bees winter well will have the same cluster of bees at the beginning of spring flight as at the end of flight in the Fall. There won't be large amounts of dead bees on the bottom board or on the bround in front of the hive.
2. Population and Brood quantity at Dandelion bloom. If they dwindle for some reason in the spring, you'll know it at this time. I want a colony that doesn't lose its population when foraging gets underway in the spring and has at least 9 frames of brood at Dandelion bloom. I also look at brood pattern at this point.
Honey production is measured by apiary and colonies that produce the most above the colony average are tagged as possible breeders.
Total production is adjusted by amount of sugar required to support colony during the year.
Propensity to swarm ...possible breeders are eliminated if they start swarm preparations when other colonies in the yard, with similar populations and brood counts don't.
Disease resistance..colonies are eliminated if they show any brood disease. Since I see very little disease in my operation, Chalkbrood is a valuable tool in selecting for hygienic bees. At one time I had an operation that was full of Chalk. Now it's difficult to find.
Varroa...does the population explode during the summer and overwhelm the colony before the honey can be harvested and/or are there DWV bees emerging and leaving the hive? Or can it go through the season and into Fall before any treatment might be necessary.
Temper...Under all conditions...temper can be broken down into three categories...read what Guzman says about temper. You're looking at not just how many stings you take when handling a colony. You're looking at the calmness/nervousness of the bees.
1. Propensity to sting
2. Runniness on combs.
3. Flightiness...think head butt-ers
Longevity of queen, or stock. I never select a queen for breeding that hasn't gone through at least two seasons. If a colony sucessfully supercedes their queen...a queen that has proven herself over time...and that colony continues to be a top performer in every category..that colony is still a candidate. I don't try to maintain some pure representation of my stock. I accept supercedure by my bees and in fact think it a natural process the bees employ.
Anyway, that's really just about it. I'm considering keeping track now of colonies that pack the broodnest so the colony becomes pollen bound. I'm trying to determing whether or not the trait is genetic or induced by some outside influence. Tarpy believes it's genetic. Hackenburg/Mendes believe it's caused by neonics. I haven't a clue.
Food for thought...
"Sometimes, breeder selection comes down to a gut feeling. You just have some feeling that the queen is special". Marla Spivak
MP, I can answer that question. The tendency to hoard excessive amounts of pollen is genetic. It is also a trait that is highly linked to bees that winter well and build up explosively in the spring. It has two significant negatives. Bees that collect high amounts of pollen also tend to pack pollen into honey storage areas and they tend to swarm very early in spring. This doesn't matter much if you are doing extracted honey, but if you want comb or chunk honey it can be a bit disconcerting. When I've had bees that showed this trait, there were three linkages, 1 - Overwintering with relatively small colonies on a huge reserve of honey and pollen, 2 - Tendency to build up explosively in spring, 3 - Tendency to pack cells of pollen randomly in honey storage areas with more pollen cells close to the brood nest. You can use the last trait alone to select for these bees. Just mark your supers with source colony and location of the super relative to brood nest, then, as you extract, keep an eye out for random pollen cells. I might add that this trait does NOT exist in any Italian strain I've ever tried.
Dar, I'm certain that Michael is aware of genetic source of excessive pollen hoarding...the question is if it is the cause of the excessive hoarding he is seeing.
One possibility (and I've shared this with Michael in the past) is that fungicides will reduce the water soluble protein in the pollen (by preventing fungal fermentation). Sharing of protein around the hive is part of the feedback mechanism used to determine how much pollen to collect, and if the protein is lacking, they may well be cued to collect more, and more, and more....
Dar, I'm not really seeing it in the spring, and it's not pollen being packed in the honey combs I'm concerned with. This pollen "hoarding" if you will, is happening later in the season. When for whatever reason a queen slows down, the colony packs the entire broodnest with pollen. So much pollen in the broodnest that there are no empty cells for the queen to lay. You have a longer build-up season so maybe you see it differently than I.