The gist of this post: When we take two plants of the same cultivar and plant them side by side to compare them we should always assume that they will be different even if we divided a double fan crown to produce the two single fans (or a four fan crown to produce the two double fans) to plant and nothing had ever been done differently to those fans. We should never assume that the two plants will be the same or should be the same just because they are planted beside each other.

If we look at a clump of a daylily cultivar in flower we will notice many differences. Some of the fans are small and some are large. Some develop a scape and some do not flower. Some scapes are taller than others. Some scapes have more branches than others. Some flowers are smaller than others. In other words there is a large amount of variability.

There is variability in everything that is related to growing plants. One of the most important causes of some of that variability is the soil. We often make the assumption that the soil is the same for plants because we water and we fertilize. But it is not. In any area there are patterns in the soil that affect plant growth. Those patterns mean that even if we place two plants side by side they will be growing in different conditions and that means that they will be different because of those different conditions no matter what effect we thought we were testing by planting them side by side.

The figure below is a simple test done more than 80 years ago. An entire field was planted with one variety of beets. Rows were 22" apart and the beets were thinned to 12" apart in the rows. When the beet roots were harvested their weights were measured along with their location in the field. The figure shows how much the weight of the roots differed from the overall field average for different locations in the field.




There are gradients throughout the field. No matter how close one puts two plants one plant will be in a different location in the gradient versus the other. The different locations will have different minerals (N, P, K, etc.) and differences in other factors that affect plant growth and all those unknown differences will cause the plants to grow differently.

More recent research into the variability of soil has looked at the actual amounts of various minerals in the soil. Nitrogen is one of the most important for plant growth. When 1 m x 1 m sections (areas 3.3 ft x 3.3 ft) of a field were examined there could be as much as two to three times the amount of nitrogen in different areas. When 12.5 cm x 12.5 cm sections (5" x 5") were examined the amount of nitrogen could vary by as much as 10 times. Those differences will cause plants (even divisions or clones) to be different even when planted side by side or near each other.

How do we design a test so that we can more or less safely assume that the natural variability in the soil (and in all other factors that affect plant growth, such as amount of shade, temperature differences, etc) are not the cause of any differences we see? We have to use many plants and we have to plant them in several different locations at random. By doing so we even out the differences between locations (when we use averages of the measured characteristics we are interested in comparing). And we have to repeat the entire test several times.

Lets say we do the test only once. For this example consider two opaque containers. In both containers someone has placed some unknown number of white beans (from 0 to 50) and some unknown number of red beans (from 0 to 50) for a total bean count of 100 beans in each container. We are going to remove two beans from each container. We cannot see into the containers. From container one we get two beans and both are red. From container two we get two beans and both are white. We decide that container one is full of only red beans and container two is full of only white beans. We are probably wrong. Both containers may have exactly the same number of white and red beans. Even if both containers have 50 white beans and 50 red beans we could get just two white beans from one container and just two red beans from the other container. If that was the case we would actually get two white beans one quarter of the time from one container and two red beans one quarter of the time from the other container. If that was what happened when we did our sampling then statistics tells us that we should not conclude anything. We do not have results that are statistically meaningful or informative. We need to repeat the test several times. As well we should increase our sample size - instead of just two beans perhaps we should take out four beans from each container. Simply because we see a difference between two groups or samples does not necessarily mean that the difference is significant. It does not necessarily mean that the same difference will be present if we repeat the entire test. It does not necessarily mean that the difference is real in general. We use statistical tests to help us decide if the difference is significant and reliable.

For a test to produce results that actually relate to the effect the grower is interested in testing and not caused by all the other unknown factors that affect plants one should probably have as a very minimum 30 plants, three sets of five treated one way and another three sets of five treated in the other way. The plants would need to be chosen at random for each set of five. One set of five from one group would need to be paired at random with a set of five from the other group. Each of the three sets would need to be planted in a location chosen at random in the growing area. In each of the three locations the first group of five to be planted would need to be chosen at random.

One possible scenario (choosing at random means using dice or random number tables, etc. to make the decision of which group to plant where in each location, for example, roll one die - if it is 1, 2 or 3 then row 1 is group A otherwise it is group B, in location 1):

Location 1 row 1 (chosen at random) five plants of A, row 2 five plants of B.
Location 2 row 1 (chosen at random might be) five plants of A again, row 2 five plants of B
Location 3 row 1 (chosen at random) five plants of B, row 2 five plants of A.

Whatever we measure should be averaged and compared using simple statistics as a minimum, to take into account the variability that is always present no matter what we do. Most of the simple, but informative statistical tests are available on the internet (or often in spreadsheet programs); one just needs the measured characteristics.

This would explain something I have noticed in my yards. I quickly noticed at our old house there spots I called "hot spots". For the most part no matter what plant I put there they grew like crazy! I have couple spot here at this house that seem to be the same way.

I had always thought my soil was the same every where in my yard but obviously not.

Interesting topic as I sit here recovering from this morning's taks of digging out a 20 foot x 30inch deep flower bed along the side of my house. That soil was so different from my other flower beds. Once I took the top layer of grass off, it was evident that it was all backfill. Lots of little loose rocks, clayish soil and no bugs in the soil, no worms either.

There will be a lot of amending done before I put in all my new daylilies.

First thing I plan to do is to take all of the top stuff and put it through my old wire sifter to save any soil that I can and then I will be adding new top soil and possible some composted manure and peat to lighten it up and make it a joy to work with.

This reminds me of taking soil samples for testing. It is recommended to take 10 1/2 cup samples from various locations in the test area, even a small planting bed. By mixing the samples up thoroughly the differences through-out the bed will be “averaged out” and the final 1 cup sample from that mix will be generally representative of the “average” conditions of the whole bed.

It might be possible, with a fair amount of effort and cost, to generate a soil map of a test bed by performing a large number of soil tests. The more the soil differences are known the more they can be taken into account when testing, allowing for increased accuracy.

Additionally, the soil uniformity of a test bed could be increased. One could thoroughly rototill the soil, test the soil, and add amendments as needed to meet particular deficiencies in particular areas. If this process is repeated for several years one could that imagine soil uniformity could be greatly increased.

Of course factors other than soil variations can also present different growing conditions to daylilies planted right next to each other. For example, I have two daylilies planted in the same bed, 3 feet apart, in afternoon shade. However, I noticed that for most of the summer one of them by pure chance happened to be in an exact spot that received much more dappled afternoon sunlight than the other. One might be able to mitigate this by placing the test bed in full sun or ensuring somehow that shade, when it occurs in the test bed, is homogeneous. One can imagine still other factors, such as actions of wild insects and moles, randomly affecting one plant more than another.

It might be possible, with a fair amount of effort and cost, to generate a soil map of a test bed by performing a large number of soil tests. The more the soil differences are known the more they can be taken into account when testing, allowing for increased accuracy.

Additionally, the soil uniformity of a test bed could be increased. One could thoroughly rototill the soil, test the soil, and add amendments as needed to meet particular deficiencies in particular areas. If this process is repeated for several years one could that imagine soil uniformity could be greatly increased.

Of course factors other than soil variations can also present different growing conditions to daylilies planted right next to each other. For example, I have two daylilies planted in the same bed, 3 feet apart, in afternoon shade. However, I noticed that for most of the summer one of them by pure chance happened to be in an exact spot that received much more dappled afternoon sunlight than the other. One might be able to mitigate this by placing the test bed in full sun or ensuring somehow that shade, when it occurs in the test bed, is homogeneous. One can imagine still other factors, such as actions of wild insects and moles, randomly affecting one plant more than another.

One could never be certain that enough of the variability/differences had been removed to make a single test/comparison in one location valid (since we can never be certain that we even know what all the factors are that could affect the characteristics we are interested in comparing). The result would be that if a difference was statistically significant one could only conclude that the two cultivars were different in that location - and at that time and not that the effect we were interested in testing was the cause of the difference. One would not have sufficient information to have some confidence that any differences found would also be present in other locations and times. I think the simplest and most effective (particularly cost effective) strategy is just to replicate (effectively repeat) the test. That is choose a number of different locations in which to put some number of plants of cultivar A and an equal number of plants of cultivar B and place the plants at random.

Anyone interested in testing/(looking for) some effect in daylilies should read the instructions found on a number of web-sites that explain how farmers can conduct on-farm tests/research on the topics that interest or are otherwise important to the farmers. Randomization and replication are key requirements for valid tests.

I should point out that even when plants are in pots with "processed/manufactured" growing medium/"soil" (presumably as uniform as possible) and grown in environmental chambers where light and temperature are controlled (also presumably as uniform as possible) both randomization and replication are still important for tests/comparisons. In those circumstances the positions of the pots are usually rotated at random at set intervals.

Also the plants at the edges of a block of one cultivar have neighbours that differ from the plants in the centre of the block. The plants on the edges, called guard rows or border rows are usually not included in the measurements due to the possibility that they are different due to differences in competitive effects between neighbours of different cultivars (and due to other potential edge effects).

There are nine individual clones/divisions of cultivar A (labelled A1..A9). The other plants in the location are unknown cultivars simply labelled X. Individual clone number 5 of cultivar A (A5) has eight nearest neighbours and all are of cultivar A. Individual clone 1 of cultivar A (A1) has eight nearest neighbours but only three are of cultivar A. Individual clone 2 of cultivar A (A2) has eight neighbours but five are of clone A.

In a normal test using a block design there would be more rows and columns of cultivar A and the plants of cultivar A on the outside edges of the block would not be measured due to possible edge effects.

It could not be more obvious that a single test would be scientifically and statistically invalid. A large number of repetitions would significantly reduce random differences affecting plant performance and characteristics, but they will not be eliminated. By increasing soil uniformity, mapping soil conditions (and taking them into account), and repeating the tests multiple times over many years, the accuracy of the tests can be significantly increased-- but there will always be randomness and minute variations that cast a shadow of doubt over the results. Over time analysis of an increasingly large data set should be able to generate a "preponderance of evidence" for various conclusions...

So basically .... this is trial and error to find the best place for a plant to grow in a given environment .... excluding amendments? Then of course, amendments can be added as needed.

beckygardener said:So basically .... this is trial and error to find the best place for a plant to grow in a given environment .... excluding amendments? Then of course, amendments can be added as needed.

This would not be an efficient method to find the best place for a plant to grow in a given environment. To find where a certain cultivar grows best in your location and conditions would be a trial and error case of simply planting divisions of the cultivar in different spots in your beds and seeing where it grew best.

So test first and then plant?

beckygardener said:So test first and then plant?

To test one has to plant then one would have to dig all of the plants up again except those that were in the best locations and move the ones from the poorer locations to the better locations. Not an efficient thing to do. I would just plant in a spot and only move if the growth was unacceptable.

That sounds like a smart thing to do. I don't particularly like digging up daylilies, so hoping to find the best spot on the first try is my goal.

Thanks for once again for straight up answers, Maurice. I am sure I probably make you roll your eyes ....

We have been in our 2nd house for the past 11 years and i amended in one way or another every flower bed on the property. I know for fact that some beds have gotten more tlc than others but i do have accurate records of any kind about where, when or of what they were amended with. I know i like to use peat moss, the very light fluffy peat moss (pardon the inaccurate name) and composted cow manure and twice that i know of, sand.
Every time i plant something new it is amended at least one of these things thus i figure i am always improving upon what was originally there which in my case is clay or backfill (more clay).
Last year i moved many plants but one plant outperformed itself many times over this year by being moved. It was one of the original plantings we did in a hurry to get all the plants in the ground in mid July when we moved in 11 years ago. The only amendment done then was topsoil tilled into the clay.
In any future beds i make i will amend heavily with what i like to use and never again with plain topsoil. In a perfect world i would dig up everything in my original bed but after doing the side foundation bed this year with only 20 plants involved i have decided to not do my original bed which has 80-90 plants in it.
I guess if we think of our gardens as tiny pieces of Real Estate where one spot is better than another then the old maxim of location, location, location rings true.
For the gardener it is the challenge of improving the bad spots so that there are no bad spots only better spots. A piece of real estate without any slums is valuable indeed.