I'm an early riser, routinely out in the garden before the bees and the birds.

Be aware that some daylilies start to open their flowers in the early afternoon, others in the evening and yet others during the night. There are night-flying moths and some of the daylily species were/are adapted to being naturally pollinated by night flying moths (and possibly other night flying insects) and other daylily species were/are adapted to being naturally pollinated by day flying insects. It has been noted that hand pollinations of some daylilies in the morning may never take - because the night flying insects beat the hybridizer and deposited pollen of the incorrect ploidy or self-pollinated a self-incompatible daylily. Imagine the problem if the night flying insects pollinated such daylilies and were partially successful and the hybridizer's hand pollinations were also partially successful.

megdavis said:

Thanks Rob!! Incredibly useful advice, I'm taking it all to heart! I may be skilled at the design of scientific experiments and data collection, management, and analysis--but I'm a total novice to daylily hybridization. Love your perspective!

Hi Meghan! We all start out with stars in our eyes, and end up with dirt on our hands.
No matter how long the road to achieving our hybridizing goals, the benefits definitely outweigh the downside. (Emotionally and spiritually if not financially!)

admmad said:
Be aware that some daylilies start to open their flowers in the early afternoon, others in the evening and yet others during the night. There are night-flying moths and some of the daylily species were/are adapted to being naturally pollinated by night flying moths (and possibly other night flying insects) and other daylily species were/are adapted to being naturally pollinated by day flying insects. It has been noted that hand pollinations of some daylilies in the morning may never take - because the night flying insects beat the hybridizer and deposited pollen of the incorrect ploidy or self-pollinated a self-incompatible daylily. Imagine the problem if the night flying insects pollinated such daylilies and were partially successful and the hybridizer's hand pollinations were also partially successful.

Great point! We're a pollinator-friendly garden and I have a fondness for nocturnals and EMOs due to my day job, so the struggle is real! I've sometimes used the same pollen source to pollinate the same pistil a couple of times: once in the early morning, and again later. Any data or experience on how useful that is?

I am inbreeding and line breeding. Hadn't considered five generations tho, so thanks for the tip! Is there a standard breeding approach with sibling pairs? I have specifically been acquiring some interesting parent-child registered pairs to get a jump on the line breeding.

Five (or more) sometimes six generations is considered the minimum - of self-pollinations with diploids. Sibling crosses requires more generations. That is to produce what is generally considered an inbred line. "Line" breeding is typically considered to be less about inbreeding and more about having a closed population of closely related individuals with the same specific selection [breeding] goals.

Tetraploids require many more generations to produce a formal inbred line.

The catch with inbreeding is (especially with diploid daylilies) that inbreeding depression is severe. According to both Stout and Arisumi (a geneticist) it can be extremely severe. Arisumi tried to do a formal genetic analysis of the inheritance of red flower colour in diploid daylilies (by making the appropriate crosses for the F1, F2, etc.,) and basically gave up because of the problems of self-incompatibility and inbreeding depression in diploid daylilies. He salvaged what he could with some crosses but only ever published the results in the AHS (not peer-reviewed) magazine.

If you are trying to produce inbred lines in diploids I would start with known self-compatible cultivars, even if you are not going to self-pollinate. Then, assuming sib crosses, start with as many as possible. I have not tried producing an inbred line of daylilies, so I do not have any idea how bad it can be. I have noticed that self-pollinations of 'Stella de Oro' produce only about half as many seeds as cross-pollinations so that is already noticeable inbreeding depression. As an example, say 50 paired sib crosses - that requires 100 siblings from the same initial cross. Number the crosses. Pollinate one flower for each cross successfully. When collecting the seeds, use a random number table to choose the crosses from which to collect seeds (sequentially until there are more than 100 seeds). Plant those seeds (typically disregard their identities).

With self-pollinations selection will be strongest for self-compatibility. With sib crosses although self-compatibility will play a role there will also be selection for different compatibility factors within sib crosses. That could counter the desired result from inbreeding. Selection for fertility has to be minimized as it will slow the inbreeding process and will occur no matter what one does, to some extent.

I've sometimes used the same pollen source to pollinate the same pistil a couple of times: once in the early morning, and again later. Any data or experience on how useful that is?

Sorry, no I don't and I do not know of any published research for it in daylilies.

admmad said:Five (or more) sometimes six generations is considered the minimum - of self-pollinations with diploids. Sibling crosses requires more generations....Tetraploids require many more generations to produce a formal inbred line.

Impressive! Most of my (limited) training is for line breeding in animal herds (one or two generations, typically). But I think for different goals (usually some balance between trait selection and health). This is terrific insight, thanks! I see that if I want to understand the inheritance of a desired trait, that is a different goal than understanding what (good or bad) recessive genes or rare traits may be lurking in my lines, and the pursuit of this knowledge may be different than the goal of production (and perhaps introduction) of a healthy plant with the traits I seek. I appreciate the advice!!

I have one residual question that goes back to the original line of this thread: registration and parentage. My question may really be one of etiquette.

Let's say you have (finally, after generations) met your goal and have a great plant with your desired traits, plus wonderful increase, and you are ready to register and introduce. You've got enough fans to start sharing with the world. But let's say the parentage back to registered cultivars is multi-generational and these parent plants are solid but nothing exceptional, and you would not have chosen to register and introduce them otherwise.

Is it better to sdlg x sdlg and then give (through breeder's notes or your own website) the full parentage back to known cultivars, or
Register a key generation of plants (parents or grandparents, depending on the number of generations needed to get back to registered cultivars) to allow for the full heritage to be shown in the database, regardless of whether you will actually send those plants out to the world.

Thanks for any and all opinions on this! Clearly, it will take me a decade or two to get to this point, but I've been wondering if there is a standard.

Jillz said:Love the graph but I would think it would be more interesting to see Tets registered relative to the total registrations each year rather than the absolute numbers. (Just giving you more to do Meghan )

My pleasure, and thanks to Dennis for abstracting more data so I could do this!

Here is the graph by count of registrations, showing the full time bar.


If we do this by percent, and limit it to the period after 1945 (around when tet registrations began), you can see that the tets became dominant in the late 90s and continue to dominate registrations:


Here are the tets by percent of total registrations by themselves (note the scale on this is only to 80% and on the prior graph the scale went to 100%:


Dennis sent some bonus data, more on that soon. Enjoy!
Meghan

More fun data from Dennis -- this is a graph of the number of hybridizers over time by ploidy:


If we compare it to the graph I uploaded earlier for the number of registrations, the two graphs show well correlated data, as you might expect!

I can understand linebreeding in animals.

"Linebreeding is mating animals so that their descendants will be kept closely related to some animal regarded as unusually desirable. It is accomplished by using for parents animals which are both closely related to the admired ancestor but are little if at all related to each other through any other ancestors. "

"WHY LINEBREEDING IS PRACTICED
Animals do not live long enough for the breeder to get all the sons and daughters he wants from the best ones. Often an animal is old or even dead before its real superiority is recognized. If its sons and daughters are mated to unrelated individuals, the offspring will get only about one-fourth of their inheritance from this outstanding grandparent. If these in turn are mated to unrelated individuals, the influence of the outstanding ancestor is again halved. Unless some form of
linebreeding is practiced, it is only a matter of three or four generations until even the most outstanding animal's influence is so scattered and diluted that no one descendant is very much like it. Linebreeding takes advantage of the laws of probability as they affect Mendelian inheritance to hold the expected amount of inheritance from an admired ancestor at a nearly constant level instead of letting it be halved with each generation, as would happen if all the matings were outbreeding."

However, the "best ones" in a perennial plant do not ever die. In a plant such as daylilies where thousands of potential "best ones" are registered every year and cultivars can become outdated rapidly, why linebreed?

megdavis said:More fun data from Dennis -- this is a graph of the number of hybridizers over time by ploidy

Are the graphs of the number of hybridizers over time or are they of the number of hybridizers who registered at least one daylily in each year or ? How are years in which a hybridizer did not register any daylilies handled (gaps in their registrations)?

How were the hybridizers who registeed both tetraploids and diploids handled?

How were, the admittedly, probably quite few, hybridizers who hybridized one ploidy for a few years and then switched to the other ploidy handled?

admmad said:I can understand linebreeding in animals.

"Linebreeding is mating animals so that their descendants will be kept closely related to some animal regarded as unusually desirable. It is accomplished by using for parents animals which are both closely related to the admired ancestor but are little if at all related to each other through any other ancestors. "

"WHY LINEBREEDING IS PRACTICED
Animals do not live long enough for the breeder to get all the sons and daughters he wants from the best ones. Often an animal is old or even dead before its real superiority is recognized. If its sons and daughters are mated to unrelated individuals, the offspring will get only about one-fourth of their inheritance from this outstanding grandparent. If these in turn are mated to unrelated individuals, the influence of the outstanding ancestor is again halved. Unless some form of
linebreeding is practiced, it is only a matter of three or four generations until even the most outstanding animal's influence is so scattered and diluted that no one descendant is very much like it. Linebreeding takes advantage of the laws of probability as they affect Mendelian inheritance to hold the expected amount of inheritance from an admired ancestor at a nearly constant level instead of letting it be halved with each generation, as would happen if all the matings were outbreeding."

However, the "best ones" in a perennial plant do not ever die. In a plant such as daylilies where thousands of potential "best ones" are registered every year and cultivars can become outdated rapidly, why linebreed?

Thanks Maurice! I was planning to self-cross and line breed (backcross) to look for recessive traits, but my instincts could be waaay off when it comes to plants. Thanks for your insights!

admmad said:Are the graphs of the number of hybridizers over time or are they of the number of hybridizers who registered at least one daylily in each year or ? How are years in which a hybridizer did not register any daylilies handled (gaps in their registrations)?

How were the hybridizers who registeed both tetraploids and diploids handled?

How were, the admittedly, probably quite few, hybridizers who hybridized one ploidy for a few years and then switched to the other ploidy handled?

Great questions -- and Dennis has most of the answers since he abstracted the data. I only had the aggregated data by year.

Wow, the information that all of you discuss and share is so interesting! Maurice, is there an instance where inline breeding in daylilies a good strategy? Would you say that line breeding for broken colors help to increase the odds of getting a broken color seedling or the odds are better if the seedlings are crossed back to the striped parent? Also, would inline breeding increase the probability of recessive genes expressions?

Rob, thank you for sharing your hybridizing insights and experience. I put the post in my notes for future reference.

@megdavis Are you going to be working with diploids primarily or tetraploids primarily or both?

I was planning to self-cross and line breed (backcross) to look for recessive traits

Dominant and recessive are relatively simple when working with inbred diploid lines (Mendel worked with peas which are natural inbred lines). When working with outbred daylilies they become more complex. In particular the concepts actually apply to how we measure traits more than to the traits themselves. For example, using inbred lines if I cross a purple flowered plant with a white flowered plant I may find that their F1 seedlings are all purple flowered. Using just my eyesight I may not be able to distinguish the F1 purple flowers from the purple parent. However, if I extract the purple pigment I may find that the F1 seedlings have substantially less purple pigment in their flowers than the purple parent has. When something like this was actually done in petunias the F1 only had about 70+% of the pigment. Using just eyesight the characteristic was a "complete" dominant. Using the quantity of pigment it was a partial dominant or quantitative. If I take the same purple allele and "white" allele and cross them into a number of different inbred lines I will find that the amount of pigment varies in the different lines, possibly both in the homozygous parent and in the F1s. When something similar to this has actually been done it has been possible to find genotypes in which an apparently "dominant" characteristic in some/many inbred lines is "recessive" in others. Each inbred line is the equivalent of one genotype or one genetically unusual genotype. This can be described as background genotype effects or modifier gene effects, etc.
In tetraploids the situation becomes even more diverse. Simple, classical dominant and recessive characteristics are more difficult to identify than in diploids. That may be because there are few situations in diploids where the heterozygote is 100% the same as the homozygote (where the "dominance" is complete). In a tetraploid where there are five possible genotypes at one gene (AAAA, AAAa, AAaa, Aaaa, aaaa) if one measured the amount of pigment present in a flower one might find AAAA - 100%, AAAa - 80% AAaa - 60% Aaaa - 40% aaaa - 0% for example and quantitative inheritance. Finding AAAA - 100%, AAAa - 100%, AAaa - 100%, Aaaa - 100% and aaaa - 0% would be unusual.

I typically suggest considering most characteristics in daylilies to be quantitative/additive (affected by many genes) and for what to expect in seedlings to be based on whether the characteristic involves "making" something. The ability to make something, for example, a pigment, is usually inherited in the seedling but is not necessarily present to the same extent as the parent (depending on the other parent).

kousa said: Maurice, is there an instance where inline breeding in daylilies a good strategy?

Depends on what you mean by "inline". If you mean backcross an F1 seedling to one of its parents. Then yes it can be a good strategy. When it is a good strategy will depend on what the goals were when the original cross was made and which of those goals were achieved and to what extent.

Would you say that line breeding for broken colors help to increase the odds of getting a broken color seedling or the odds are better if the seedlings are crossed back to the striped parent?

Broken colour (transposons) and striping are likely to be independent characteristics. When gametes are formed by the broken colour parent the broken colour may often be repaired. Depending on the characteristics present in the striped parent, even if it is not repaired, broken colour may not be visible in the seedlings - it all depends on the parental characteristics. If the seedling did not inherit an active transposon then crossing it to the striped parent cannot change that but only affect striping.

Also, would inline breeding increase the probability of recessive genes expressions?

Please define what you specifically mean by inline breeding. Which recessive genes or what characteristics?

Since we're doing a deep dive into statistics, it might be interesting to know the percentage of AHS members who have registered a daylily. I feel as if it's pretty high. I was initially wondering what proportion were hybridizers, but since a good number of hybridizers probably haven't registered a daylily, I don't think there's a way to get that data, short of a survey. There may be quite a few hybridizers who aren't AHS members.

JamesT said:There may be quite a few hybridizers who aren't AHS members.

It seems to me that all hybridizers should be required to be members of the ADS, otherwise why even have the ADS?
I noticed that at least one hybridizer offers the ability for customers to name and register one of their seedlings.

I think there is a continuum of value among the 80000 cultivars in the database, from these otherwise-culled plants through the best work of serious hybridizers.

admmad said:@megdavis Are you going to be working with diploids primarily or tetraploids primarily or both?

Maurice, you are an absolute treasure! I've been trying to wrap my head around pigment genetics, especially in tets, and your explanation here really helped me focus my thinking more explicitly. Thank you!

I'm working with both dips and tets, with the caveat that tet genetics has so confused me, I was going to focus a lot of the inbreeding program on the dips with only a few goals for specific traits (actually, mostly traits *other* than bloom pigment). I'll still do some inbreeding in the tets, but I've assumed this will be exploratory and to grow my own experience. Sure, I eventually want to figure out if it is even possible to get a daylily the exact Pantone color for the Raven's team purple, but my gut feeling is I'll need tets for that. Maybe I'll get lucky with my "fun" crosses, or maybe I'll knuckle down in five or ten years and work on this as a main focus, but for now, it's in the garden goal/side project category. Or maybe someone here has already done it and wants to sell me a daylily.

For diploids, I am developing a list of (Mendelian) expectations for each of my few chosen traits given classic inheritance patterns (dominant, recessive), and my plan was to compare observed to expected, and develop probabilities for the trait (I may or may not use a Bayesian approach...that's a teeny bit advanced for my statistical skills). I've been working under the hypothesis that I may, for some traits, be able to see partial penetrance using this approach. (Your note that objective measurement is needed to do this with pigment is well taken!! I'll have to consider how this might occur with the traits I want.) Once I have enough data, I hoped to be able to "predict" what range or distribution to expect from certain crosses, at least within my own lines, and then try combinations of goal traits. My ultimate goal was to use this knowledge to help me optimize later goals where I wanted a new trait PLUS the trait I'd evaluated. Hope springs eternal, right?

As a side note to those who are (sort of) following this, but who have never had genetics (and may want to explore why genetics, parentage and data analysis are all bouncing around the same thread together), I found this relatively straightforward site you might find useful: https://www.nature.com/scitabl...

Once again, thanks Maurice!!

I could see having to be a member of the AHS in order to register a daylily, but not to be a hybridizer. I may never register a daylily, but it is fun to try to reach a goal. I am a member of the AHS, but if I were not, I would still be dabbing pollen on the daylilies.