jmcextman: Half-baked research: multi-leader, pedestrian apple orchard

In 2022 -- which seems like ages ago -- I submitted a proposal to the New England Tree Fruit Growers Research Committee titled "A modern, pedestrian apple orchard system(s) comparison using a disease-resistant rootstock and variety to be planted in 2022." The proposal can be seen here, thanks to Steve Wood/Poverty Lane Orchards and the Committee for agreeing to fund the planting. The details:

Rootstock: G.11 (fire blight resistant)

Variety: Crimson Crisp (scab-resistant)

Planting location: UMass Cold Spring Orchard, Belchertown, MA

In-row tree spacing (3 treatments, leader spacing similar across three systems at 1.5 feet)

Super-spindle, single leader, 1.5 feet between trees
bi-leader, two leaders, 3 feet between trees
UFO, 4-leader, 5 feet between trees

Between-row tree spacing: ~10 feet

Replications: 5, with 6 trees per replicate, times 3 treatments (as above) = 90 trees total

Planted in April 2022, pretty much as above, at the UMass Orchard in Belchertown, MA. Trees were from Wafler Nursery, on the smaller side (3/8-1/2 inch caliper). Turned out to be a single row, relegated to a location which turned out to be an ideal deer buffet. The deer found Reps 4 and 5 to their particular liking, especially given the nature of the UFO being right at head-high feeding height for them. At least that is my take on the situation, a bit frustrating, and ultimately did not use Reps 4 and 5 for any of the data analysis.

Let me tell you how the trees were trained, so there are three "treatments" (trt) in this experiment:

Super-spindle, single leader (single-), 1.5 feet between trees: Pretty much planted as was, then trained and pruned to wires as hybrid super-tall-spindle. Goal was to keep them at a height that could be harvested from the ground. Turns out not-so-easy, being a tall person (or stilt capable) would help.
Bi-leader (bi-), two leaders, 3 feet between trees. Nursery trees were ideal candidates to be headed (pruning cut) low at planting, about 12-18 inches off ground, making sure there were viable buds below cut. They really took off, and the two leaders pretty-much grown as if they were single leader super-spindle, but at a slight angle. Had to so some shoot stripping while selecting the two leaders in 1st- and 2nd-leaf. Turned out very good in general. Difficulty level -- pretty easy, not perfect, but most trees have nicely balanced (same size) two leaders. That heading cut as you know stimulates growth at the cut location, and boy, sure saw that. In 4th leaf easily have filled space and are getting tall.
UFO, 4-5 leader, multi-leader (ml-) 5 feet between trees: How about I start by saying just don't do it! Lots of headaches. After planting, the nursery trees which were generally about 4-5 foot tall were bent down to horizontal and tied to single low (12 inches) wire. Upright (hence the 'U' in UFO) shoots that emerged at the top of the horizontal "leader" were selected for equal size (approximately) and subsequently trained vertical to the wires, having to use some string to tie and promote the upright shoot growth. Lots of work, had to remove all shoots early that were not going to be the UFO's. Not easy to get shoots of equal size, and in Spring 2025, had to selectively head back the too vigorous shoots to get a new shoot. Do not recommend doing this in any way, shape, or form, unless you have lots of extra time to kill. These trees, however, remain well within picking height from the ground but also have apples low which is kind of back-breaking picking. As mentioned, the deer really liked to browse these too, it's a deer buffet.

If not altogether clear, the idea was also to save money on trees when planting, given the different spacing. I am no economist so am not going to go there right now, you do the math.

See representative pics at end for the three systems.

Data collection? Well, did not go 100% as planned. No crop in 2022, a few apples in 2023 but not sure if I have any data there, all apples harvested and weighed by rep (five) and treatment (the three planting systems) in 2024 and 2025. In 2024 we (I) also counted the apples so I had an average fruit weight. Slipped my (aging) mind to do that in 2025.

Results? Of course as a "scientist" here, have to have numbers, and so some kind of statistical analysis. But the statistical "power" of this experiment is very low given only 3 reps. I have seen a colleague use a one-way analysis of means (ANOM) which I think is useful when looking at the data. So here are the ANOM's with a bit of explanation.

lbs. of apples harvested in 2024. b-, bi-leader; multi-, multi-leader (UFO); single-, single leader (tall-spindle); remember this going forward

Before I go any farther, the results are from groups of trees, the three systems having the same number of trees, but different spacing. Lbs. of apples harvested are from the trt X rep groups of trees. Always 6 trees, just different spacing. I did not think this through too well. Just consider the fact that all created more-or-less a fruiting wall. Avg is the mean of all. UDL and LDL are Upper and Lower "Decision Limits." Points within the shaded area do not differ from the Avg (= no statistical difference, I am using P<0.10). So, interpreting above, there is no statistical difference of any of the three (bl, ml, ts) treatments from the overall mean. Numerically, however, you can see multi- had less apples than bi- and single-. OK? Leaves me head scratching a bit too, but let's move on.

average weight (lbs.) of apples harvested in 2024

Bingo! multi- had bigger apples! (But remember, also less yield per above.) Followed by bi- and single (but bi- bigger apples than single-). These Crimson Crisp apples were about two to the pound. I hand thinned them.

lbs. of apples harvested in 2025

No statistical difference in yield between the systems in 2025. But look at that numerical difference! Here, let me help visualize that:

OK? Now, even though I said I was not going to dwelve into economics, I changed my mind. Keep in mind all the data above are from six-tree groupings. Single- covers approximately 10 linear feet; bi- 20 feet; multi- 34 feet. If, I, as a very small producer (urban farm market?) planted 100 linear feet with each of the three systems, my cumulative yield 2024-25 (total lbs. of apples) would be: single- 1,600 lbs.; bi- 810 lbs.; multi- 394 lbs. Trust me, it's close. If I got $2 lb. average for the apples, the dollar value would be $3,200, $1,600, and $800 respectively (single-, bi-, multi-). Interesting. But the planting cost, approximately, assuming $15 per tree would have been $1,000 (67 trees), $500 (33 trees), and $300 (20 trees), single-, bi-, and multi- respectively. Trees only. No site prep, labor, trellis, irrigation, planting, etc. But that should all be the same (approximately). What is not the same is the amount of time spent training and pruning, which in order, more time to less time, would be multi- > bi- > single-. Get it? All of the above why I don't do economics! That's your job!

Whether or not I take one more year of data on this half-baked research is TBD. Frankly, at this point I cannot recommend the multi-leader, UFO system. Simply too much labor and skill needed. Bi-leader is great if you have the right trees at planting -- smaller caliper with good, viable buds below knee height -- to make the bi-leader work. Saves 50% on tree cost of the single-leader, tall-spindle. If I were going to go with the tall-super-spindle, which ultimately I would recommend, I'd plant the trees 2-3 feet apart. Maintaining tree height so all trees could be picked from ground is a challenge, but a step-stool type "ladder" would be sufficient. I like this one. More on tall-spindle here.

One more thing, anecdotally, Crimson Crisp apples harvested in 2025 from the multi-leader were larger and more mature than what was harvested off the bi-leader and single-leader trees. Remember, they were also larger apples in 2024. If it makes a difference to you?

Thanks to the UMass Orchard farm crew for help with harvest in 2025, and once again of course, for the New England Tree Fruit Research Committee (and Steve Wood) for the monetary support to buy the trees for this half-baked research. JC