How to lie with statistics

I’m attending the American Society for Horticultural Science (ASHS) meeting this week in sunny (and hot) Palm Desert, CA. Tomorrow I will be presenting some of the early results from the SOME-DED-TREES study. For those that are new to the blog, SOME-DED-TREES is the acronym for the Social Media DesigneD TREe Establishment Study. Last year, my students and I established a landscape tree study in which the treatments were suggested by Garden Professor’s blog readers. Actually, we ended up installing two studies: one to look at fertilization at planting and one to look at impacts of mulching at planting. The trees were ‘Bloodgood’ planetrees grown in 25 gallon containers. In each study we divided the trees into three root treatments: We “shaved” the outer roots to eliminate circling roots; we “teased” apart the circling roots; or we just planted the trees as is (or “Pop and drop” to use Linda’s vernacular).

I reported some of the results of the mulch study here on the blog last fall and will include some of those data in my talk tomorrow. I am also presenting some of the data from the fertilizer portion of the study. One of the ways we assessed fertilizer response is by using a device known as a SPAD meter.

SPAD meters have been around for a while and they are very useful. The device measures light transmittance through a leaf, which is highly correlated with leaf chlorophyll content. Chlorophyll content, in turn, is highly dependent on foliar nutrition so SPAD readings often provide a useful indicator of plant nutrient status.

Here at the meeting I am using my laptop, which means I am without my usual statistical and graphing software. So in order to plot some of the data for my talk I am relying on Excel. This gets me to “How to lie with statistics.” When I calculated the means for this June’s SPAD chlorophyll index values this is the chart I got. Looks like a pretty impressive response to fertilizer, doesn’t it?

SPAD chlorophyll index of Planetrees fertilized at plant (Fert) and control trees (No).

The problem is the scale. Note that by default, Excel truncated the scale on the y axis to values between 28 and 36. This is a big no-no in scientific circles. Graphs scales should include zero or show if break if there is some reason the range needs to be truncated. The reason truncating the scale is a no-no, as shown here, is that it makes differences look proportionately larger than they really are. Advertisers use this trick all the time. Next time you see a bar chart in a sales brochure or magazine ad, look that the scale – bet it doesn’t go to zero.

Let’s look at the chart after I re-scale it. It still looks like there’s still something going on with fertilizer, but it doesn’t make you go, “Whoa!” like the first chart.

SPAD chlorophyll index of Planetrees fertilized at plant (Fert) and control trees (No).

As it turns out, there is a statistically significant effect of fertilizer on the June SPAD values. However, this effect essentially disappeared by the time we re-assessed the trees three weeks later. Here’s the final version of the data as it will appear in my ASHS talk tomorrow.

This brings up another issue we face in this kind of work; statistical significance versus biological or practical significance. The ‘Fert: p<0.01’ on the slide indicates there is a 99% probability that the difference in the mean SPAD values between Fertilized and Non-fertilized trees on June 17 is not due to random chance or error. But as the leaves continue to mature this effect essentially disappears. Could the early season boost in nutrition be enough to give these trees an edge in the long run? Possibly, but I wouldn’t say likely. To date we have not seen any effects of fertilizer on growth but it’s still relatively early in the game. Tree growth is cumulative and effects that may be too subtle to detect early on may turn out to be significant later on. Stay tuned.

Neon for your garden

Was wandering through Target on Monday for the first time in months.

Helloooo!? The 80’s called and wants its neon crap back.

Didn’t care for it then and certainly don’t care for it now. Though there is the increased safety factor of being highly visible at all times, whether in sunglasses or underwear.

But never mind my lack of style.

It made me think about a few plants that, if the light is right, certainly display that glowing, saturated color, found in the “Astro-Brite” pack of copy paper usually reserved for yard sales and such.

Close to dusk, the Kniphofia uvaria ‘Echo Mango’ in our garden stands out from 100 yards away. Bred and selected by Richard Saul of Atlanta’s ItSaul Plants Inc., it is one tough perennial, taking heat and humidity with aplomb.

My experience has been one big early summer flush of blooms, with some significant reblooming until frost. Best in full sun, it’s also drought tolerant and cold hardy to USDA Zone 5. It doesn’t get whopping huge like some other Kniphofia do – stays a nice manageble size, topping out at 3′ to 4′ tall. ‘Echo Mango’ (or any other Red Hot Poker) adds a terrific bit of vertical interest to an otherwise mound-y mixed border. Best with fellow warm colors. Pink, not so much.

‘Echo Mango’ = glowsticks!
Achillea ‘Paprika’ doesn’t go so well with it. Mental note to relocate it in fall.

You can almost hear the sound of space lasers…
Eeee-yoooooo-eeeee-yoooooo…or maybe that’s just me.

Brace yourself

The photo below (graciously sent to me by former MSU Extension Educator Jennie Stanger) graphically illustrates the importance of removing ALL staking and supporting materials from trees once they are established.

Just a matter of time (Photo: Jennie Stanger)

In this case the stakes were removed but the strapping material was left around the tree. Since this is a spruce, Jennie supposes no one wanted the prickly job of wading into the center of the tree to take off the strap. Eventually the trunk was girdled and when the area recently experienced some heavy thunderstorms, the weak spot on the tree was exposed.

Stately evergreen to mangled mess (Photo Jennie Stanger)

As a general rule we recommend that all staking and support materials are removed within two years, preferably one year. This type of damage is one of the prime reasons: after two years who is likely to remember that there is still strapping left on the tree?

The heartbreak of ‘Carrot-top’ syndrome

The perk of participating in a blog is you get a platform to vent on your pet peeves. Recently I’ve seen several classic examples of ‘Carrot-top’ syndrome. No, I’m not talking about the red-headed comic; though he tends to annoy me too.

Annoying Carrot-top #1.

The ‘Carrot-top’ I’m referring to occurs when white pine trees are sheared as Christmas trees but then planted as landscape trees. The typical result is that the side and lower branches remain suppressed while the terminals go crazy. I’m not sure why syndrome occurs in white pines and not other trees; it may be related to vigor of white pines and how hard the growers have to shear them to keep them in shape.

Annoying Carrot-top #2.

I love my friends in the nursery and Christmas tree industries and they work hard to grow quality trees, but this is one practice I’d like to see end. And, to be fair, they are giving customers what they want. If we set up a survey at a garden center and placed a 7’ sheared white pine next to a 7’ white pine that had been minimally pruned, 19 out of 20 people would take the tree that had been sheared to look like a Christmas tree. However, this is truly a case where less is more.

Going off half-cocked

The good news about being a landscape extension specialist is you get to comment on a variety of landscape tree problems. The bad news? You get to comment on a variety of landscape tree problems. In an average week I probably get 10-15 tree-related calls or e-mails from homeowners, landscapers, growers, extension educators, lawyers, newspapers, and on and on. Sometimes the problem is routine and obvious like tar spot on maples; “I have these spots that look like tar on my maple leaves…” Sometimes the problem looks routine and turns into something else entirely different like the Imprelis herbicide debacle. With the number if problems I respond to, invariable I sometimes end up shooting from the hip – and miss the mark.

Red maple with sparse canopy (right foreground). Maple in left-center of picture across the road was planted at the same time.

A few weeks back I noticed some red maples on campus that were struggling as they were starting to leaf out. There were planted about three years ago following a road-widening project. They are located on a tough site between a parking lot and a busy 4-lane road. My immediate diagnosis was they were weakened due to the aftereffects of last year’s sever heat and drought. I even included a photo of them in an article I did for our on-line extension news. Then this morning I happened to talk to a friend from our campus landscape services. “Boy, those maples sure took a beating from the drought last year”, I volunteered. My Landscape services friend replied, “Well, that and we hit them with Shortstop last year…” Uh oh, now the conversation was going in a different direction. Shortstop is a plant growth retardant, the active ingredient is paclobutrazol, which inhibits GA synthesis in plants and is being widely used by arborists to reduce the need for tree pruning. In this case it worked a little too well and the trees ended up being stunted.

Leaves of paclobutrazol-treat maple (left); untreated leaves on right.

In my defense, I was partly right. Under normal weather patterns paclobutrazol applied at the labeled rates (as was done here) would have reduced shoot growth and leaf growth but clearly the drought exaggerated the effect. So, another lesson learned (actually re-learned) never assume anything about trees until you talk to the people who actually take care of them.

Sheep-eating flowers?!

I was planning to follow up on Jeff’s phosphorus post with a bit more “phun with phosphorous.” However, I was completely derailed by Ray Eckhart’s message and link left for me on our GP facebook page with this headline:

RHS ‘sheep-eating’ plant about to bloom in Surrey

“The Royal Horticultural Society (RHS) at Wisley said the Puya
chilensis, a native of Chile, would bloom in the next few days and last
about a week.

In the Andes it uses its sharp spines to snare and trap sheep and other animals, which slowly starve to death.”

“The animals then decay at the base of the plant, acting as a fertiliser.”

[Dear BBC News: “Snare” and “Eat” are not really interchangeable]

How…have I not heard of this before.

I’ve posted previously on my Puya fixation. I asked readers to bet on how long it would take for me to kill my wee Puya bertoniana, mail-ordered from Annie’s Annuals. I am happy to report that it made through the winter (greenhouse) and is now sunning itself on our deck.

Now this really ups the ante – it captures sheep! Maybe P. bertoniana isn’t as robust as P. chilensis though. Perhaps…a vole or rabbit?

Phosphorus and Big Macs

Minnesota, and I were cruising through old pictures and files and getting all sentimental about the cool stuff we used to do. A lot of it was never published just because after we were done with one thing we were just too damn excited to move on to the next. Anyway, one of the neatest experiments that we never wrote up was a phosphorus experiment. Here’s what it looked like to the casual observer.

Now let me explain the neat part to you a little. Inside those boxes, underneath three of the six plants in each container, are vials set up like this – three vials per plant (the black tubes provide air to the vials).

Each plant had one root placed into each of the three vials – one vial contained 1 ppm phosphorus, one vial contained 10 ppm phosphorus, and one vial contained 30 ppm phosphorus. The tub itself was also filled with one of these three solutions (1, 10, or 30 ppm phosphorus) as seen below.

At the end of the experiment we weighed the roots filling each vial, as well as weighing all of the roots from each plant. Here’s what we found for the individual vials.

As you can see, more phosphorus in a vial meant that the plant would devote more energy to growing roots there – but also notice that the 10 ppm solution has the greatest mass of roots overall. Here’s what we saw when we looked at the total size of all of the roots from plants for the different solutions.

As you can see, the roots from the plants in the 10 ppm solution are the largest (shoots showed the same trend). So here’s the way I see it (this is the Big Mac part). I love Big Macs. If I see a McDonald’s I want to go in there – I gravitate towards McDonald’s to get Big Macs. But too many Big Macs aren’t good for me. They might even stunt my growth! It’s the same for phosphorus. Roots do grow towards phosphorus (this isn’t technically correct, but it works for my analogy so I’m sticking with it!), but that doesn’t mean that a tremendous amount of phosphorus is actually good for them. In fact, it might even stunt their growth! This could be for a variety of reasons, but most likely because the phosphorus would interfere with the uptake of other elements.

Label GMO foods? Sure, why not?

Lots of coverage in the mainstream media these days over various initiatives to label GMO foods. I think GMO foods should be labeled; but not for the reasons you might think.

My personal opinion on GMO foods is that their benefits outweigh the potential downsides. I think GMO foods should be labeled to make the public aware of how much of our food supply depends on GMO’s and the cost of not using GMO’s. Obviously this will raise some social justice issues since wealthy people will have more opportunity to opt out of buying non-GMO products than the poor but it’s the same problem we have with organic already and it doesn’t seem to cause much of a stir. My hope is that once GMO products are labeled we could get to the point where the main discussion on GMO’s focuses on the rational and scientific questions, not the irrational and emotional. This weekend our Sunday paper included a quote from a local anti-GMO activist who tried to link the rise in obesity to increased use of GMO’s. How about we’ve become too sedentary and we’re eating too much, period? Or this from the “Health Ranger” Mike Adams: “Roundup herbicide devastates soils, rendering them contaminated and unable to produce healthy crops using traditional (or organic) farming methods. Once a farm plot is destroyed with Roundup, that farmer is forever enslaved to a chemical-based farming protocol.” Hmmm… last time I checked farms could be certified organic after three years without synthetic chemicals.

Which is not to say I don’t have concerns about GMO’s; there are issues with any technology. The largest questions I see, and the ones that are most difficult to answer, relate to unintended consequences. One of the biggest selling points of Round-up ready technology, for example, is that it enables farmers to manage weeds with glyphosate, a relatively safe product in the world of industrial-strength herbicides, in order to reduce tillage and maintain crop productivity. But as farmers use more and more glyphosate, they are also selecting for Round-up resistant weeds. How long until glyphosate is no longer effective? Difficult to predict, but glyphosate resistance has already evolved in many weeds.

On the other side of the equation, the world’s population is projected to exceed 9 billion by 2050, with a disproportionate increase in the least developed countries. As the need for food increases, land and water resources will become more limited and catastrophic droughts are likely to increase. While it is easy to demonize industrial agriculture, it’s difficult to envision feeding ourselves and the world without the technology it has developed.

Moss-tacular!

Mosses are soft, green, and tough as nails, as shown in a recent article in the Proceedings of the National Academy of Science (prestigious, high impact journal with a rather unfortunate acronym).

Dr. Catherine La Farge and associates, from the University of Alberta, visited a remote glacier on Ellesmere Island, Nunavut while studying the wild, wide world of arctic bryophyte systematics. Bryophytes are ancient, non-vascular, non-flowering plants – mosses and liverworts, mostly.

Long story short, they harvested bits of moss that had been trapped in ice for about 400 years and were now exposed. Several species were collected, taken back to the lab, ground up, placed on growing media in a growth chamber, and they soon had mosses galore.

This is fascinating on several levels, as pointed out by the authors. One is the power of totipotency – the ability of a cell to “de-differentiate into a meristematic state that can then reprogram the cell for development of the organism” a la stem cells. Another is the mosses’ ability to “shut down” when dry and “revive when conditions are favorable” (like not frozen in ice for 400 years?!)

The article also graphs the disturbingly accelerating rate of retreat of the Teardrop Glacier, where the mosses were collected. Aargh. The window of favorable conditions may not be open long for these little wonders.

Lights!

Today I want to share something that I’ve been working on recently with Fine Gardening that is really cool! So you know all of those lights you can buy to get your plants started over the winter? Did you ever wonder which of those lights really work? I’m going to leave the final answer for my article, but let me tell you, there’s a world of differences between the lights. The best seem to be some LED lights that aren’t available yet, but are made by a company called Heliospectra. Mostly they make high end lights for commercial producers and researchers, but they will be entering the home market soon. Man, these lights are SWEET!

Interestingly, the other LED lights fail miserably because their light is so columnated (in other words the light doesn’t spread out), as you can kind of see in this picture (there’s another pic on facebook that shows it a bit better):

Most of you are probably using fluorescent lights, and, in terms of bang for your buck, I’d say those are pretty good. We tested a bunch of different fluorescent, as well as incandescent lights. Be on the lookout for the article in about a year!