Dennis616 said:
Meghan, I do not guarantee accuracy but I believe this is a mostly accurate count of tetraploids officially registered. The late 1960's – early 1970's were when they really started being produced in quantity:
RegYr Ploidy Ct
1949 Tetraploid 1
1951 Tetraploid 4
1954 Tetraploid 1
1955 Tetraploid 1
1959 Tetraploid 27
1960 Tetraploid 2
1961 Tetraploid 14
1963 Tetraploid 6
1964 Tetraploid 2
1965 Tetraploid 5
1966 Tetraploid 14
1967 Tetraploid 58
1968 Tetraploid 69
1969 Tetraploid 80
1970 Tetraploid 89
1971 Tetraploid 145
1972 Tetraploid 122
1973 Tetraploid 180
1974 Tetraploid 243
1975 Tetraploid 156
1976 Tetraploid 201
1977 Tetraploid 166
1978 Tetraploid 238
1979 Tetraploid 340
1980 Tetraploid 218
1981 Tetraploid 303
1982 Tetraploid 260
1983 Tetraploid 255
1984 Tetraploid 263
1985 Tetraploid 287
1986 Tetraploid 373
1987 Tetraploid 345
1988 Tetraploid 356
1989 Tetraploid 330
1990 Tetraploid 407
1991 Tetraploid 349
1992 Tetraploid 450
1993 Tetraploid 394
1994 Tetraploid 463
1995 Tetraploid 589
1996 Tetraploid 744
1997 Tetraploid 468
1998 Tetraploid 762
1999 Tetraploid 995
2000 Tetraploid 1048
2001 Tetraploid 967
2002 Tetraploid 1063
2003 Tetraploid 1569
2004 Tetraploid 1427
2005 Tetraploid 1508
2006 Tetraploid 1754
2007 Tetraploid 1543
2008 Tetraploid 1567
2009 Tetraploid 1785
2010 Tetraploid 1647
2011 Tetraploid 1875
2012 Tetraploid 1708
2013 Tetraploid 1986

As promised, here is a graph of the data Dennis helpfully extracted from the registration database on tetraploid registrations over time! You can see there was remarkable growth between 1995 and 2005. Still increasing, but the rate of increase has slowed somewhat of late. (Click on the image to enlarge it.)



Enjoy! This is the fun part of having a registration database.
Meghan

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 )

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 )

Sure! Just get me the data in .csv or excel format, and I can do almost any graph you want!

Fascinating, thank you!
Interesting to introduce epigenetics to this discussion. I have only seen it mentioned in regard to studies on ageing. Cells can "forget" their original purpose when exposed to stressors like time, pollution etc....

so to use this concept in hybridization... well, perhaps the use of colchicine (sp?) for conversion is a kind of epigenetic modification... what do I know?
But I find it interesting to contemplate.

And then there is Clown Pants. Steve says that this very seemingly-rare trait of striped sepals is now actually dominant (he thinks). Looks that way from seedlings that are coming out!

megdavis said:

Yeah, EXCELLENT point. Missing data is a big problem, and these would probably be missing for reasons, rather than missing at random, which really challenges our interpretation of the statistics.

So now we need to do focus groups too before we make too many changes....LOL!

Ahem, missing records is a lot bigger problem for statistical analysis than random missing data within the database. It's a lot better to have a vibrant, growing database that everyone is happy conrtibuting to, than have a dying database that fewer and fewer people are interested in. Having 100,000 partial records is better than having 10,000 "perfect" records.

Why do I get the feeling we're just being trolled by people who don't know a trowel from a trug? Good luck with all this.

Lyshack said:
Ahem, missing records is a lot bigger problem for statistical analysis than random missing data within the database. It's a lot better to have a vibrant, growing database that everyone is happy conrtibuting to, than have a dying database that fewer and fewer people are interested in. Having 100,000 partial records is better than having 10,000 "perfect" records.

Yes, I agree! Missing records are bad, especially when they are not random, and keeping the registry thriving is a critical goal!!

(Quick statistical point: MNAR - missing not at random - within a dataset is also terrible, both your point about missing records (selection bias) and missing data points (MNAR) can lead to flaws in the analysis such that the findings can actually be *wrong*. We're better off not collecting data points than collecting them badly, so I would argue for 100,000 smaller records that are mostly complete than 100,000 larger records with lots of partial data, and I agree that we want the 100,000 not the 10,000.)

If we want to ask a specific question of the data, then maybe the solution is not to collect such data through the registry, but instead do it as part of a survey outside of the registry process. That way, people can participate or not as they choose. But that's starting to get researchy....and not very trowel-friendly....would have to be a winter exercise for sure!

I do think there may be consensus (on this thread at least) that an open-ended optional breeder's notes section might be a nice addition to the registry. I don't think that would discourage too many people, and has potential to be a helpful feature.

I would also argue to create hybridizer profiles within the data system (a one-time, free activity, and then have hyperlinks from hybridizer names on the registration display to this profile) where some useful data like zone of hybridization could be collected, since I think the question of what hybrid survives where seems to be a perennial (pun intended) theme on this forum. Hybridizers could also be incentivized to fill out this profile if there was a place for website address (another recent forum topic: The thread "Forum Member websites featuring their daylilies" in Daylilies forum). Since I hear all serious hybridizers eventually move to Florida, we'd have to be able to log updates to the zone... Guess I'll never be a serious hybridizer: it's too warm there for me!

SueVT said:Fascinating, thank you!
Interesting to introduce epigenetics to this discussion. I have only seen it mentioned in regard to studies on ageing. Cells can "forget" their original purpose when exposed to stressors like time, pollution etc....

so to use this concept in hybridization... well, perhaps the use of colchicine (sp?) for conversion is a kind of epigenetic modification... what do I know?
But I find it interesting to contemplate.

Epigenetics is fascinating and there's so much we don't know!

I've actually been wondering for a while about colchicine - we know it does modify the genes themselves (they double) and based on a quick scan of some papers, it looks like it could influence the epigenome too.

Getting back to Blame it on the Rain, and the importance of data on lineage - does anyone know if ALL the stripey daylilies are from tet lines, perhaps certain tet lines? (As in, somewhere in the past, a parent was treated with colchicine? I heard there was another new drug being used that was less human-toxic, what about that one?) Or could these just be like calico cats, and be based on different chromosomal expression? [Edited to add: if all the blooms look different, I think that could be an argument for the calico cat option....]

OK, I really do need to stop procrastinating and go edit student papers....

I have been following this thread with fascination. Many technical stuffs are way above my head but I appreciate all who try to explain these concepts in a simple way to understand. My question from all this is what information do seasoned hybridizers use to make their crosses and select the seeds to plant? Do they actually rely on these statistical data analysis to make their decisions? Take Jamie Gossard and Karol Emmerich for example who have been hybridizing for years. Do they collect genetic data on their seedlings, analyze the data and use them to make their crosses? What results do hybridizers actually get from the crosses they made from these statistic analysis? Do they provide impressive results? It seems to me a lot of the unique and different daylily flowers are made by chance rather than by statistical analysis. The broken colors and sculpted forms daylilies are such examples. I don't believe that the hybridizers planned for stripes when they did the crosses that produced Undefinable, Wacky Wednesday, or Explosion in the Paint Factory.

Everyone who hybridizes daylilies and is interested in the parentage of registered daylilies should be aware that there are errors in the recorded parentage and errors in the ploidy (diploid/triploid/tetraploid) of registered daylilies.
The errors may have several different causes but one of them is that daylily hybridizers do not use what have been called "safe" hybridizing techniques. That is, most simply assume that there are no natural pollinations and only their hand pollinations work.
A rough estimate of the number of errors in parentage, based on the errors in ploidy was 6%. A more recent study found 16% errors in parentage also based on errors in ploidy.
Errors in ploidy will only capture some errors in parentage. I would expect that errors in ploidy would significantly under estimate true errors in parentage.

kousa said: My question from all this is what information do seasoned hybridizers use to make their crosses and select the seeds to plant? Do they actually rely on these statistical data analysis to make their decisions? Take Jamie Gossard and Karol Emmerich for example who have been hybridizing for years. Do they collect genetic data on their seedlings, analyze the data and use them to make their crosses? What results do hybridizers actually get from the crosses they made from these statistic analysis? Do they provide impressive results?

I have never read any evidence that "backyard" daylily hybridizers analyze their crosses in the same way that professional plant breeders and geneticists would analyze crosses. I doubt that they collect the necessary sorts of statistics.
Most new daylily characteristics have been found by chance when crosses between related daylilies have been made and the hidden effects of new mutations have appeared. In a few cases new mutations have obvious effects immediately (typically about 5% of new mutations are "dominant").

admmad said:I have never read any evidence that "backyard" daylily hybridizers analyze their crosses in the same way that professional plant breeders and geneticists would analyze crosses. I doubt that they collect the necessary sorts of statistics.

I realize a single case only shows potential and not frequency, but I'm definitely "backyard," I'm doing extensive research before I make my crosses and I'm tracking all kinds of data, although I'm not (yet) using any professional plant hybridization sofware programs (any suggestions???):
https://garden.org/thread/view...

But as I've said before, I'm used to being an outlier....

Getting back to Blame it on the Rain, and the importance of data on lineage - does anyone know if ALL the stripey daylilies are from tet lines, perhaps certain tet lines? (As in, somewhere in the past, a parent was treated with colchicine? I heard there was another new drug being used that was less human-toxic, what about that one?) Or could these just be like calico cats, and be based on different chromosomal expression? [Edited to add: if all the blooms look different, I think that could be an argument for the calico cat option....]

Stripes of various sorts in plants can have many different genetic bases (and in some cases completely non-genetic bases). In daylilies they have been found in both diploids and tetraploids.
No one has done the necessary genetic analyses to determine the possible causes of any of the striping characteristics in daylilies so the following is based on what has been studied in other plant species.
Several of the striping characteristics in daylilies are likely to be based on what are commonly called "jumping genes" or transposons. Those are basically sections of DNA that can be inserted into different locations of a plant's chromosomes (diploid daylilies have 11 pairs of chromosomes - approximately 30,000 to 40,000 different genes ). When inserted they typically disrupt the gene into which they have been inserted and may show a visible characteristic. The transposon has the ability to be removed from its location. That may happen whenever the cell divides. Once removed it is unlikely to be re-inserted into the same gene and in the same location. Therefore it may not have a visible effect. When inserted into one of many genes that affect flower colour its typical action is to stop that gene from producing its normal product and the end result is often a change in colour. Since the effect is inherited all the descendants of the cell will show the change in colour - the result is a stripe of a different colour. A typical effect when the transposon is inserted is a loss of colour. When it is removed ("jumps out") is a return in colour.
Other striping effects are probably caused by mutations that alter the way that the genes respond to environmental factors. Some of those environmental factors are external for example, light and temperature and others are internal.
Every cell in a plant has the same genes (more or less barring somatic mutations of various sorts, etc.) Yet all the cells of a plant are not identical in appearance, etc. That means that although all the cells are genetically identical which genes function in a particular cell and for how long they function in each cell and when then function in each cell can be different. That is, the regulation of the genes varies. Striping of various sorts can be caused by changes in the regulatory DNA, which may be regulatory genes or other regulatory DNA.
There are many genes involved in flower colour. Some are described as "structural". They act to produce and modify the flower pigments. Others are described as regulatory. They regulate the structural genes. Striping can be caused by changes in any number of those genes.

From the description of how 'Clown Pants' acts in crosses I expect that its cause is different from any of the other examples of striping in daylilies.

megdavis said:Getting back to Blame it on the Rain, and the importance of data on lineage - does anyone know if ALL the stripey daylilies are from tet lines, perhaps certain tet lines?

Never mind, the answer is that stripes can be either dip and tet, so if it is epigenetic, it's not necessarily from colchicine:
Registered Dip: Daylily (Hemerocallis 'Cherry Stripes')
Registered Tet: Daylily (Hemerocallis 'Clown Pants')

[Edited to add: THANKS MAURICE! We crossed posts, and yours is such a lovely description of the potential biological mechanisms of striping, much appreciated! Is there any sense if the transposition could be mediated by plant viruses? Next thing you know, we'll be using CRISPR-CAS to build custom mutant hybrids...]

megdavis said:I realize a single case only shows potential and not frequency, but I'm definitely "backyard," I'm doing extensive research before I make my crosses and I'm tracking all kinds of data, although I'm not (yet) using any professional plant hybridization sofware programs (any suggestions???):

Tracking data for valid genetic analyses is not simple. It is the preliminary steps that are required to insure that the plants from which the data is collected are an unbiased sample of the necessary plant generations that is vital. Then depending on what sort of data (what characteristics are of interest) the environmental factors that affect plant growth and development have to be effectively randomized.

When you make your crosses, are you insuring that your pollen is the only possible one that produced the seeds? Professional plant breeders typically force open flower buds before they open naturally, remove the stamens, pollinate the flower and then close the flower and enclose it in a "bag"- to insure that only the pollen they used was effective.

Are you making the same crosses repeatedly to produce a large number of seeds? Do you repeat the same crosses in different years? Do you grow the seedlings in different locations in your garden - in other words do you replicate the cross both in time (years) and locations?

To study inheritance the hybridizer cannot select. When seeds are collected they cannot be chosen for any characteristic (unless one specifically wants to study the effect of choosing for that characteristic) . When they are planted they cannot be chosen for any characteristic. When(if) seedlings are transplanted they cannot be chosen for any characteristic. Etc.

Those are the fundamental requirements that are basic to studying inheritance in plants.

Then there is the data collection. It needs to be as objective as possible and on substantial numbers of seedlings.

An example:
Let's say one was interested in the inheritance of the number of flowers produced.
One daylily (A) produces 10 buds per scape. Another (B) produces 20 buds per scape.
Over the course of the next 5 to 10 years depending on how many years it takes for the plants to flower we produce F1, F2, F3 and two backcross generations so that they flower in the same year and are the same age so that various characteristics can be determined/measured, etc.
It is likely that we would need a minimum of 50 plants for each generation. All plants would have to be grown in the same way. Typically that would mean measuring the amount of fertilizer each plant receives, the amount of extra water, etc. Then means and variances of bud counts would be collected from the plants in the different locations and different years.
Since bud count is a quantitative characteristic (as most are) a biometrical analysis (quantitative genetic) would be done (in the pre-DNA analysis past). Nowadays DNA would be collected and the plants would have been inbred for 5 generations before any crosses, etc.

For a qualitative characteristic, say flower colour, the same generations would be required, the number of seedlings might need to be even greater since the statistical analyses would depend on the possible number of genes involved in the flower colour and different statistical analyses would be appropriate. The classification of the flower colours would be the difficult part.

Is there any sense if the transposition could be mediated by plant viruses?

It has been generally considered that "genomic shock" may re-activate transposition. There are enormous numbers of transposons and portions of transposons in most living organisms. They nearly all have been inactivated. I am not sure that any specific type of genomic shock has been experimentally shown to re-activate transposons but virus infection is one possibility. Another suggestion is interspecific hybridization. Obvious possible shocks are environmental, etc. The use of colchicine could be another - although one would expect that the appearance of an activated transposon would be in the same generation as the colchicine was used (or any other "shock" occurred).

I hesitate to post in this technical thread since my math and statistics skills could fit in a thimble, but from what I have learned so far in my near decade of hybridizing (which still makes me a total novice, in my view) and seeking as much advice as I can from experienced hybridizers (who are usually generous with their knowledge and willing to help if you ask), I just wanted to make a few quick points:

1) All serious hybridizers do NOT eventually move to Florida. The warm climate in FL allows one to make crosses and have enough fans of the resulting seedlings in 2 years to bring it to market. This allows a rush to cash in on the latest look, but many serious hybridizers want to advance the overall quality and performance of the daylily over the long run. It's impossible to test and develop daylilies that will perform well in the northern part of the country if you are hybridizing and growing in FL. There are daylilies that do well in the north, and daylilies that do well in the south. There is some overlap, but it is limited, as many people have learned at the expense of their pocketbook.

2) To fully evaluate a daylily intended to be grown, say, above the Mason-Dixon line, takes 5-8 years.

3) A daylily that has a pretty bloom is not necessarily intro-quality - it needs to have a good plant under it. Some newer hybrizers have discovered this when people buy their early, pretty intros and then complain (online) that they don't perform particularly well. A pretty bloom is only half of it. So it's back to the drawing board for scapes, foliage and overall plant quality.

4) Even a spectacular daylily with a good plant under it is not guaranteed to be a good parent. Some of the best performing intros can be duds as parents. This is why a long evaluation period is required - to see what kinds of kids they throw and how those kids perform. Simply being pod and/or pollen fertile is no guarantee you're going to get great kids, and it's no guarantee the kids will have the parents' trait you want unless that trait has been shown to be passed along to the intro's kids. Which takes time.

5) Daylily hybridizing is largely a numbers game. While all of us can get lucky from time to time, the old rule of thumb that 1 out of 200 seedlings will be intro-worthy still holds. Even hybridizers who have been at it their whole lives and have developed awesome, high quality programs, still plant thousands of seeds every year.

6) Hybridizing seems to be largely a combination of long, hard experience and the wisdom that is gained from that, together with good intuition, which improves over the years with the hard-earned experience.

Hybridizing by its nature is a very humbling experience that requires lots of patience over the long haul. I thought (or at least hoped) that I could do a short-cut that would render all the years of sweat and tears unnecessary but I soon learned what everyone else does - Mother Nature will give you failures where you were expecting successes, will make you wait when you are in a hurry, and will refuse to give you the looks or traits that you were expecting. But she will also - and this is what keeps us all going - from time to time grant you a miracle seedling that makes your heart go pitty-pat, puts a huge grin on your face, and suddenly all the years of hard work seem worthwhile. Math and statistics may help and I am always in favor of more analysis and information, but they're not going to let you do an end run around the trials and tribulations of hybridizing!

Even a spectacular daylily with a good plant under it is not guaranteed to be a good parent. Some of the best performing intros can be duds as parents. This is why a long evaluation period is required - to see what kinds of kids they throw and how those kids perform. Simply being pod and/or pollen fertile is no guarantee you're going to get great kids, and it's no guarantee the kids will have the parents' trait you want unless that trait has been shown to be dominant and passed along to the intro's kids. Which takes time.

"Good" parents are not necessarily determined by the characteristics they show. A "good" parent is determined by the characteristics that its offspring show. The parent does not need to show those same characteristics to pass them on successfully to its offspring. Most characteristics are neither simply dominant nor simply recessive. Most characteristics are affected by many genes with the typical end result being quantitative even though some of the genes may be treatable as "dominant" and others may be treatable as "recessive".

Maurice, I love that you verify, and essentially yes, with the note that I'm really new to hybridization and some of these are planned protocols rather than years of proof. I should caveat that this is yes for my few primary hybridization goals, I've given myself permission to slack off for side projects that are mostly meant to grow the garden and entertain my family.

admmad said:When you make your crosses, are you insuring that your pollen is the only possible one that produced the seeds? Professional plant breeders typically force open flower buds before they open naturally, remove the stamens, pollinate the flower and then close the flower and enclose it in a "bag"- to insure that only the pollen they used was effective.

This is the one area where I have considered kicking it up a notch. I'm an early riser, routinely out in the garden before the bees and the birds. What I haven't done is force open a bloom and bag it, but I've been pondering to do this for my high priority crosses. I'll upweight this given your recommendation.

admmad said:Are you making the same crosses repeatedly to produce a large number of seeds? Do you repeat the same crosses in different years? Do you grow the seedlings in different locations in your garden - in other words do you replicate the cross both in time (years) and locations?

Yes to all seeds, yes to different locations, yes to replicates. (I am a scientist.)
I don't have enough space to start ALL the seeds in my entire collection at the same time, so I try to divide them across batches and track this. Some, I do the whole pod at once. Some, I divide across two batches. I stratify to aid in more consistent emergence. I log the data on seed starting and emergence.

admmad said:To study inheritance the hybridizer cannot select. When seeds are collected they cannot be chosen for any characteristic (unless one specifically wants to study the effect of choosing for that characteristic) . When they are planted they cannot be chosen for any characteristic. When(if) seedlings are transplanted they cannot be chosen for any characteristic. Etc.

Yeah, it's a free for all in my garden - once they get systematically sorted (by seed number, which are assigned to seeds using a convenience random method -- but admittedly not a random number generator) into their different spots, they are on their own for a while. I'm also systematically dividing parents into different parts of the garden to see how they perform under the (surprisingly varied) conditions within my acre of usable turf. I've got dry, I've got wet, I've got sun, I've got shade, I've got well-amended soil, I've got total clay, I've got morning sun, I've got afternoon sun. What I don't have is FLAT. Oh well.

admmad said:It is likely that we would need a minimum of 50 plants for each generation. All plants would have to be grown in the same way. Typically that would mean measuring the amount of fertilizer each plant receives, the amount of extra water, etc. Then means and variances of bud counts would be collected from the plants in the different locations and different years.

Have you done a power calculation? Might need more than 50!
Just kidding, I standardize my watering, my compost amendments, and my neglect. I don't use synthetic fertilizer or chemical amendments. Your point about measurement is well taken - if I had a flat surface to have a greenhouse, where I could actually measure RH, °C, and standardize water inputs through an automatic system, I seriously would. I'm planning to put in a fully-automated watering system once I save up for it. I've got my eye on some fancy soil moisture meters.

Call my approach epidemiology and not experiment; it's close, but you are correct: not quite the same. Nonetheless, I am data driven. I have been tracking rainfall for a couple of years and have considered to actually start a log file since the NOAA data don't match my personal microclimate. (The NOAA data are pretty good for temp, and I do have local stream flow since there's a gauge on the stream near my house: I've considered using this as an alternative to the hand-captured rainfall log since, well, I'm away a bunch for work and also some of my colleagues have shown that river flow correlates well with rainfall in the Chesapeake area. My expectation is that stream flow should also correlate highly.)

admmad said:Since bud count is a quantitative characteristic (as most are) a biometrical analysis (quantitative genetic) would be done (in the pre-DNA analysis past). Nowadays DNA would be collected and the plants would have been inbred for 5 generations before any crosses, etc.

While I have the technical ability and pipeline to do the genetic analysis, I *don't* have the money or time for a personal project. Bioinformatics is a time hog!

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.

admmad said:For a qualitative characteristic, say flower colour, the same generations would be required, the number of seedlings might need to be even greater since the statistical analyses would depend on the possible number of genes involved in the flower colour and different statistical analyses would be appropriate. The classification of the flower colours would be the difficult part.

I'm the one who posted about her Pantone swatches and incited the counter-argument against standardization of color as a requirement of registry (and I totally defer to the counter argument; the comments I make about using a standard color wheel are just for my own program and not as a requirement for registration.)

What I really love about your post, Maurice, is that you really do break it down in a concrete way for newbies like me and for more established hybridizers looking to up their game. Huge kudos. I would *love* more input and mentorship! I'm a bit of a bad mentee, since my job can hijack me for months on end, but I care a lot and am open to critique. Importantly, I'm not looking to make daylily hybridization a profession. I've already got a professional career that's so needy, I can barely manage it.

I feel there's a tag line here: Nature-inspired, data-driven, and 100% backyard....
I've got a problem. I admit it. Twelve step program?

Thanks, Maurice. I figured you'd clarify my use of "dominant". Not being a scientist, for me decisions are based on observation on whether the kids show traits I'm looking for, without knowing the proper nomenclature. Thanks as always for the clarification.

RobLaffin said:Hybridizing by its nature is a very humbling experience that requires lots of patience over the long haul. I thought (or at least hoped) that I could do a short-cut that would render all the years of sweat and tears unnecessary but I soon learned what everyone else does - Mother Nature will give you failures where you were expecting successes, will make you wait when you are in a hurry, and will refuse to give you the looks or traits that you were expecting. But she will also - and this is what keeps us all going - from time to time grant you a miracle seedling that makes your heart go pitty-pat, puts a huge grin on your face, and suddenly all the years of hard work seem worthwhile. Math and statistics may help and I am always in favor of more analysis and information, but they're not going to let you do an end run around the trials and tribulations of hybridizing!

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!