This section describes my citizen scientist experiments on precision direct seeding with predator protection, along with some confirmation by real scientist. In summary, the average seed-to-surviving-seedling rate at one year is around 30% and the best-in-class results have exceeded 60% seed-to-surviving-seedling rates at the one year mark. Of note, the best results came from trials planting more than one seed in each spot, so that the “success rate” (defined as at least one viable seedling in a planting spot at the one-year mark) has approached 80%. One goal for future studies by longleaf enthusiasts would be to isolate the variables that need to align to produce those best results.

Trials Deploying the Manufactured Prototype Seed Crowns®

Trials have been underway with varying degrees of scientific discipline at: (1) Berry College’s Lavender Mountain near Rome, Georgia, (2) the Sprewell Bluff WMA between Thomaston and Greenville, Georgia, and (3) Indian Grave Mountain between Thomaston and Zebulon, Georgia.  Berry College and Indian Grave Mountain would clearly be montane longleaf settings at elevations over 1000’ msl, with the Berry College location having a Hanceville-series sandy loam and Indian Grave Mountain having predominately Pacolet soils.  Although not very far from Indian Grave Mountain, the Sprewell Bluff site is more of a piedmont setting below 700’ elevation, with Cecil-series sandy loam soils.  

The Berry College effort has been conducted by students under the supervision of Professor Marty Cipollini who, among many other roles, oversaw Berry’s Longleaf Project at the time of the studies.  

In mid-February of 2023, the students planted 2 seeds each under 300 spots for 600 total sheltered seeds, and 2 seeds in 73 spots without shelters for 146 unsheltered seeds.  The plots were in multiple locations. They tallied the entire trial in early August 2023 and found 167 surviving seedlings in the sheltered cohort (28% survival) and 26 in the unsheltered cohort (18%), with a 47% “Success rate” for sheltered spots.  Among other hypothesis, the students sought to examine whether a predator protection device improved seedling survival.  Their results indicated that the shelters produced a roughly 56% improvement in survival rate.  In the fall of 2023, students began using the method in small scale reforestation efforts focused more on the practical process of getting seeds planted in the forest rather than on collecting data from carefully controlled conditions. Their research has been submitted for peer-reviewed publication.

I planted the five Sprewell Bluff test plots during November 2022 with 2 seeds each under 50 shelters per plot, for a total of 500 seeds under 250 shelters.  The setting was beneath recently-thinned loblolly plantation, with all on one plot and half of the remaining four plots located within skid trails, and half of four plots located under the thinned loblolly.

I used shelters made from translucent PLA for four of the five plots and opaque shelters made from a PP-wheat straw blend for the fifth plot.  I tallied all five plots at four months and found 380 vigorous seedlings for a germination-and-survival rate of 76% and a “Success Rate” of almost 88% . 

The twelve-month tally of the Sprewell Bluff trials produced astounding results, showing 317 surviving seedlings (63% germination and one-year survival) and 194 spots with at least one seedling (78% “success rate”).  The survivor/success rate for the five plots ranged from a high of 74%/86% (i.e., 74 seedlings from 100 seeds, and 43 out of 50 spots having at least one seedling) to a low of 50%/72%. Thus, even at the low end, the rates would be remarkable. As might be expected, the seedlings within the skid trail appeared more vigorous at the one-year mark but survival rates were fairly even across plots.    

In a negative observation, however, some of the pure PLA shelters showed signs of warping in the summer heat. While I could find no spots where the warping shelter killed a seedling, there were several where the shelter was obviously impeding growth. Coupled with the economics of recovering shelters as opposed to intentionally leaving them to be consumed by a prescribed fire, this observation led to favoring the less expensive, more durable PP material for the first publicly available models while work on truly biodegradable materials continues.    

I set up five plots for measurement on my property on the Pine Mountain terrane in southern Pike County, Georgia, between October and November 2022. 

The results from those five plots so far are as follows:

The seed batch for the first three plots I planted was from a frozen lot collected in 2021 and the germination test produced uncharacteristically weak results.  In addition, all three of those plots were intentionally located in especially challenging micro site conditions.  In fact, I was unable to locate PSTP 20 because the grass and other growth was chest high, and the survivors located in plots 18 and 19 appear suppressed.  The seed batch for the other two plots showed strong germination results and the micro-site conditions more favorable (even if notably rocky as the photo above shows).  Thus, it is probably unsurprising that the survival and success rates for those two plots are closer to the Sprewell Bluff results. 

Combining all the plots from the three different test sites and more than a dozen separate plots, the results show about 590 survivors from 1582 seeds at the one-year mark, or better than one viable seedling for every three seeds carefully placed in the ground.  If one throws out the 382 seeds that either received no shelter or were from a suspect seed lot, the seed-to-survivor rate at one year would be 521 out of 1200, or about 43%.  Making the same adjustment to look at success rate per two-seed spot, it would come out to 364 out of 600, or 61%.

As an aside, that 61% rate is slightly lower than the 68% per-spot rate a straight probability estimate would predict from a 43% per-seed success rate. Part of the explanation might just be the ineptitude of the citizen scientist doing the counting for two of the three trials. However, an additional explanation is that the micro-sites between individual spots even within the same test plot will differ, as will the localized calamities like whether a deer stepped on a shelter. Thus, it may be slightly more likely to have either two survivors or no survivors in a given spot than the straight math would predict, a fact that may counsel against trying to increase success rate by planting additional seeds per spot.

As noted earlier, 2023 was a pretty wet year in our area and, despite record average temperatures, we did not see any extended periods of extreme heat. These factors likely contributed to the strong survival and success rates, which were better than those seen in the earlier tests of homemade shelters described further below.  In any event, the results are encouraging.

Trials Deploying Earlier Homemade Prototypes

My first attempts to track data on test plots using precision direct seeding and some form of shelter commenced in 2020, and I am certain that my earliest experimental methods would make any serious scientist cringe.  Taken in the aggregate, however, the earlier trials summarized below involved over 1700 seeds planted in nearly a dozen separate locations. Therefore, I believe these earlier results help confirm the conceptual viability of coupling precision direct seeding with some form of predator protection in a variety of field conditions. 

A careful reader may note that the above chart starts with Test Plot 5. The first four test plots involved attempts to plant more than 5,000 seeds without shelters. The results were astonishingly bad, with total survival less than one percent. It was almost as if the seed predators were able to trail me in the woods. Who knows, maybe they can?

Reviewing some photos from the earlier trials, Plot Six was planted on November 21, 2020, in a mixed hardwood/pine forest from which most of the hardwood component had been harvested four years earlier, resulting in what would best be described as a clear-cut microsite.  The area had been burned about four months prior to planting.  The only site preparation for individual planting locations was the scrape of a boot toe to clear any surface litter, and each removable basket received two seeds.  The following photos showing seedlings at 3 weeks, 4 months (after basket retrieval), and 22 months. A re-tally of Plot Six at 22 months found 16 out of 50 planting locations with at least one survivor (32% “success” rate), with 5 spots having two survivors and 11 spots having one (21% “seed-to-survivor” rate).

Plot Thirteen used the earliest homemade prototype Seed Crowns® made from PLA drink cups and was planted on April 17, 2021, in an abandoned dove field within which planting locations were established by browning spots roughly two feet in diameter with glyphosate.  Each Seed Crown® location received one seed.  The following photos show seedlings at 4 weeks, 4 months, 15 months (after enough clearing to be able to see the Seed Crown® and seedling), and at 16 months following a low-intensity prescribed fire.  Interestingly, although only 10 of 100 seeds in this plot produced a surviving seedling at the 6-week mark (possibly due to the late planting date), 7 of those 10 remain vigorous after the burn.

Plot Fifteen used homemade prototypes with curve palisades and was established on October 30, 2021, in the understory of heavily thinned >30 year-old loblolly that had been burned the prior July.   The only site prep for individual locations was the scrape of a boot toe, and each Seed Crown® received three seeds.  The following photos show seedlings at 3 weeks and 10 months. A re-tally at 10 months found 10 out of 30 planting locations with at least one survivor (33% “success” rate), with 2 spots having three survivors, 6 spots having two survivors, and 2 spots having one (22% “seed-to-survivor” rate).

Looking Ahead

The next step for my property will be to expand the scale of precision direct seeding to cover the roughly 200 additional acres that lack a nearby cone producing longleaf.  I have learned a fair amount about collecting and processing cones at the scale of a mere citizen landowner.  More about how regular folks can secure thousands of viable seeds from a few remnant trees is available at: Cones and Seeds.  If I end up doing substantially all 200 acres myself, I am guessing it might take me another ten years, a decent spell for a 60 year-old.  With luck, I will be able to find a little help but, either way, there are worse ways to get out of the house.  

I hope that the bona fide experts attempting to replicate my results under more rigorous scientific methods will produce additional insights and refinements along the way.  Additionally, I look forward to seeing how fast a team of planters with a little practice and an ability to touch their own toes without sitting down are able to deploy Seed Crowns®.  It is still a bit of a hunch at this point, but I suspect that planting a seed in the woods will prove to be an efficient, sustainable method for regenerating longleaf in at least some settings, and perhaps other fire adapted species, as well.  

If you would like to learn more or even join the effort, please contact me.