A Closer Look to Guide Farm Use of Tree/Shrub Silages: Per-Species & Ensilement Analyses for Safe, Nutritious Rationing, plus Replicable Trial Results

NUTRITIONAL ANALYSES

Dairy One (Ithaca, NY) completed nutritional analyses on 102 leaf-samples of 20 tree, 11 shrub and 1 vine species. I organized results by species (except for one European Buckthorn sample paired with a same-day, same- site Smooth Buckthorn sample for comparison), with fresh-frozen versus ensiled paired samples listed at top of each species-specific list, and following samples generally organized consecutively by harvest-date. A later-season date = shorter ensilement; we froze most ensiled samples during 2023-’24 winter feed-out, including those harvested in fall of 2022 when our Leaf-Separator machine first arrived (therefore the 2022 barrels and buckets had the longest ensilement period, with a few warm-enough fall weeks plus one full summer).

Our Species averages of each nutritional measure use ensiled samples only, for consistency, and are only computed when multiple ensiled samples of a species were tested; we tested just one fresh sample for most species, and fresh/ensiled paired samples repeat the same batch. Averages are below species-specific lists containing multiple batches, and averages of 10 most abundant species are also compiled and averaged across species on the last 2 pages..

Dairy One NUTRITIONAL DATA from ALL 2022-'25 LEAF-SAMPLES  (For compilation of 10 most abundant species averages, go to END)                (pdf formatted to print on legal-sized, landscape-oriented paper.)  

Previous Northeastern nutritional data on tree and shrub leaves was scarse, and nonexistent for ensiled leaves; our results are therefore highly useful despite sampling limitations. My caveots describing limitations follow:

These results are specific to Waldo County, Maine, and most samples came from just 3 SARE FNE22-013 harvest-sites, with varying and sometimes low numbers of samples per species. That harvest, which produced most of our samples, was aimed at yield per lineal feet of field edge, labor-time measurements, and provision of many barrels-full for livestock trials in that project. We tested one fresh and one ensiled sample from each species that we came upon (we missed sampling a few), and tested one sample from each barrel, bucket or date-batch later once ensiled. Quaking Aspen averages for instance represent multiple samples from one stand only which produced many barrels – on retrospect I have regrets about our 3 Streams Farm Trial per-barrel test allocation decision, which left limited funds to test other species, harvest dates or sites.

We did not plan nor commit to a broad randomized sampling design for regionally generalizable results, as that would have taken dedicated driving and sample-packing time beyond our budget and availability; rather our project utilized the “bird in hand” of leaves already harvested, with just a small number of samples added to fill out our species list.

Validity of comparisons of species averages is limited by differing site qualities and differing harvest date-windows at the sites. The Leaf-Separator machine farm-trailer base required slow transport-speed, so was parked at sites consecutively with just one return-trip to Y Knot Farm. (As I write this report, Jon Thomas Jr., Thomas Bandsaw Mills, is squaring/aligning the frame and wheels of that farm-trailer base to handle normal road speeds, in future!)

Dry Matter (DM) across all individual samples 2022-2025 including trees and shrubs both fresh and ensiled averaged 40%.12 (mostly fresh) shrub samples added in 2024 and 2025 (close to half of our 26 total shrub samples) had a much earlier average harvest date than did 2022-’23 samples; earlier sampling dates seem to correlate with lower DM, and pulled our average DM down from previous calculations.Lowest among these were 6/3/25 fresh Pagoda Dogwood with 20.6% DM, and 7/31/25 fresh European Buckthorn, shaded by mature trees in a stream flood zone, with 20.8% DM. Our (mostly MOFGA-site ensiled) Honeysuckle samples ranged 25-55% DM, with unclear date correlation. Tree samples ranged less than that within species, but just as widely across all species. 10 tree-leaf samples (of 86) had DM above 50%; 6/26/23 Box Elder (the only Box Elder sample we used for nutritional testing) had the lowest %DM of our tree samples, with 23% DM fresh and 22% DM ensiled.

DM content affects how much one must harvest, to meet needs of one’s animals. Traditional harvests of pollarded trees in Europe occurred during late summer into fall, when DM is high, and tree carbohydrate stocks are also high (carbohydrates are needed for tree recovery and health, as well as for forage value).  Here are my laborious DM and Harvest Date computations, to make sense out of our less than consistent data sets (which as stated above were aimed at bulk harvest versus sampling):

% DM & Harvest Date Aves of Leaf-Sample Groups  (<--  pdf link, in case above screenshot is not clear)

 

Protein: Crude Protein (CP %DM) levels in our leaf-silages were respectable, falling between those of the 1st- and 2nd-cut hay we had that winter. Both 1st- and 2nd-cut hays were cut in late August, due to 2023 wet June and July; the bulk of our leaf-silages were harvested late June through July, and then a lesser number of barrels late September through to October 12^th. Early cuts of both grass and tree matter are known to have higher protein levels as % of DM, than later cuts. Yet I wondered t owaht degree higher moisture/lower DM at earlier dates accounted for this difference.  

2025 harvest-dates were 6/3 to 7/7 with a mid-June average date, an earlier range than 2022-’23 6/25 to 10/6 dates with mid-August average. Because DM and fiber were lower in these earlier-dated 2025 samples (by 6-7% & 5-8% respectively), I computed protein levels AS FED, below: This showed that difference in %DM protein levels between these sample-groups does not change how much volume animals must eat to get their protein.  AS FED, the protein-levels differ little.

Acid Detergent Insoluble Protein (ADICP) levels in my 2019 leaf analyses had indicated limited protein availability (as do our new results). We therefore ordered quantification of Soluble Protein %CP (SP) and Rumen-Degradable Protein %CP (RDP), to gain more clues about protein utilization. In many of these tests, the required liquid preparation gelled up and clogged the lab’s filter. After the first large batches of samples yielded much but not all requested data, I gave in to Dairy One’s gentle suggestion for us to stop ordering SP and RDP tests.

We have complete SP and RDP data for all samples of 16 species (out of 26). SP and RDP tests failures happened on all 5 Red Oak samples, and on fresh/ensiled sample-pairs of American Elm, Norway Maple, and Smooth Buckthorn, such that we have no measurements for these 4 species. Either SP or RDP failed on 4 out of 5 Honeysuckle samples; RDP failed on 4 out of 7 Quaking Aspen samples, 1 of 6 White Ash samples, and 3 of 11 Black Cherry samples. All of the first 3 SP tests and 2 of the first 3 RDP tests failed on Gray Birch, such that we have no SP and only one (curiously high) RDP result on Gray Birch; Red Maple had this problem for 3 of 6 SP and 4 of 6 RDP tests.

To compare actual quantities in the feed, I computed SP %DM and RDP %DM for both leaf- and grass-forages. Average SP %DM levels in our Black Cherry and Green Ash samples were highest among our tree samples (2.75 and 2.73 %DM respectively), slightly surpassing the level in our 1st-cut hay; certain samples of White Ash, Quaking Aspen, Basswood and Box Elder had similar levels. RDP levels were also highest in Cherry and Green Ash, but lower than that of the 1st-cut hay.

Our three Autumn Olive samples surpassed all other tree and shrub species, with CP ranging from 21 to 26.4 %DM, SP ranging from 4.8 to 9.5 %DM, and RDP ranging from 10.5 to 17.7 %DM. For comparison, Dairy One average grass hay has 11 %DM CP, 3.7 %DM SP and 7.2 %DM RDP, and average grass silage has 15.5 %DM CP, 8.2 %DM SP and 10.9 %DM RDP. Among the few shrubs with complete protein tests, a 6/3/25 Pagoda Dogwood sample also had respectable protein levels.

Jaime Garzon,UME Cooperative Extension Forage Specialist, uses a practical guideline that WSC to CP ratio should fall between .4 and 1.5 in silage, for digestive balance. Of that CP, he says SP should be 30-40%. Our average Black Cherry silage falls well within that WSC/CP range at .65, but with SP at only 20% of CP. Our Green Ash ratio is .575 with SP at 22%. Our 2 lowest-protein Autumn Olive samples come closest, with average WSC/CP ratio at .45 and SP averaging 25.5%. (We unfortunately did not obtain a WSC level to do this math on the much higher-protein early-harvested sample.)

Condensed Tannins (CT) in our leaves are likely to be increasing protein utilization of all feedstuffs consumed by as much as 25%, according to Wayne Zeller’s ongoing work on our leaves (described below) combined with studies of cattle performance when fed other sources of CT. CT at 3% DM is optimal; more is anti-nutritional and antifeedant. Our levels are similar to those in Birdsfoot Trefoil (BFT); high-CT BFT fed as 60% of dietary DM to dairy cattle falls well below the anti-nutritional level, and increased digestive efficiency and milk production (Hymes-Fecht et al. 2013).  

PDF, Birdsfoot Trefoil Silage increases Production of Lactating Cows, Hymes-Fecht et al. 2013   

According to MOFGA Grounds Director and farmer Jason Tessier, Tessier farm, many grass-based farmers in Maine like himself make early wrapped baleage, and/or grass silage in a bunker. These early cuts have a surplus of protein, making energy versus protein the limiting nutritional component on such farms. See discussion of high NFC and WSC in our leaf-silage, below (past the Fiber discussion).

Quality of Protein: Tech Advisor Karl Hallen had a thought as we edited this report: Amino Acid profiles of grass forages do not fully meet Amino Acid needs of animals; perhaps woody forages might have more complete protein. We looked and found little info on Amino acids in Northeasatern leaf-species. Mulberry leaves in China DO have a more complete array of Amino Acids than does grass (Jiang & Nie 2015, cited in Mulberry FeedValue - Trees for Graziers ).

Acid Detergent & Neutral Detergent Fiber (ADF and NDF) % of DM, averaged across averages of our 10 most abundant species of leaf-silage with August 9^th average harvest-date, were respectively 70% and 66% of Dairy One averages of levels found in grass silage. Ensiled samples of matched (fresh/ensiled) pairs that were harvested in 2025 had earlier average harvest-date of June 17^th, and had only 56% and 54% of ADF and NDF levels in grass silage (plus had more shrubs/less trees but that has less effect on %DM than does date).

Our leaf-silage leans somewhat toward the “concentrate” end of forages; low fiber correlates with high energy (see NFC, Fat EE and NE discussions below). In our SARE FNE22-013 trial, when 10 goats and a steer ate 55% and 33% (respectively) of their diets as leaf-silage (in a 100% forage diet, leaf-silage plus poor quality 2023 1^st-cut hay), animals were satiated with 97% of the total lbs. DM that they ate when given 2^nd-cut hay in place of that leaf-silage proportion. Strangely, this percentage of less DM intake with leaf-silage is the same for the steer and goats, despite that the goats were eating a much greater proportion of leaf-silage. What need was the steer meeting with his 33% leaf-silage diet, that allowed him to eat less DM?  (Or is that difference simply accounted for by higher moisture in the leaf-silage than in hay?  That seems unlikely to me.)

Leaf-silage or 2^nd-cut hay were only offered once per day for 2 hours. If I’d had enough to offer a second such free-choice meal-period per day of leaf-silage, I suspect that they might have eaten that much again (but that would bring goats above capacity – very wide and giving more milk?).

If the steer ate twice as much leaf-silage, he would be eating 67% of his lbs. DM as leaf-silage, and would probably be satiated with 86% of the total lbs. DM that he ate when given 2^nd-cut hay once per day and no leaf-silage. This level of leaf intake for a bovine would be far above modern expectations (Lindsay Whistance says cattle choose 12% of their diet as woody forages:  Whistance 2018), yet up until the mid 1700s European cattle were indeed wintered on more leaves than grass, with 100% leaf cattle-diets in winter a probable reality for thousands of years, in earlier historic times.

Non-Fiber Carbohydrates (NFC): At 38.8% of DM average across 10 most abundant leaf-silage species’ averages, our leaf-silage had well over twice the NFC of Dairy One average grass-silage (at NFC 16.8% DM), and surpassed average grass-silage level of

Water-Soluble Carbohydrates (WSC), the most quickly digested component of NFC:  WSC in our 10 most abundant leaf-silages averaged about 10% DM; Dairy One grass silage average is 8% DM.

Such concentrated energy means that leaf-silage can support reduced use of off-farm-sourced concentrates, particularly when other dietary components compensate for lower protein levels, or if the leaf-silage is Autumn Olive, or some other higher-protein species.

Fat Ether Extract (Fat EE) level in our leaf-silages averaged 6% DM, or 150% of that in average Dairy One grass-silage.

Hauge, Garmo & Austad (in Austad & Hauge 2014) wrote (my English notes from the Norwegian – thank you to Yvonne Taylor, past Black Locust Farm farmer, for translating to me a few years ago) :

“The fat content is high in leaves (5 – 7% DM), as compared to grass (1 – 3%).  The reason is that the leaves have a thick, protective layer. This layer serves as a defense against parasites, and reduces water loss as a result of transpiration in the plants. The outermost part (cuticle) is built from a cutin and wax that doesn’t have any nutritional value. Leaves from birch often have a higher content of fat than leaves from willow, vier, or, aspen, and rowan. The least fat content is found in elm, ash, linden, and hazel. Like in grass and clover (trifolium species) it is linolensyren [Linolenic acid, ALA] that is most prevalent in leaves (35 – 45% of total fatty acids) (Garmo 2012).”

In light of this, our leaves should have around 2.4% DM Alpha-Linolenic Acid.

Ensiled leaves consistently had more fat than fresh leaves (averaging 111% of fresh DM level, or 106% of fresh level As Fed).  See "Nutritional Changes from Fresh to Ensiled" below.  

Yulica Santos Ortega identified hundreds of lipids in our goats’ milk with and without tree/shrub leaf-silage, during the SARE FNE22-013 3 Streams Farm trial. (I would love to know more about health effects of those lipids on us milk drinkers – Yulica said the lipids are “all good.”) Yulica wanted to look at leaves eaten, but then moved to an all-consuming new job at University of Virginia. I have leaf-silage samples frozen, tallied by dates fed, in which to identify leaf lipids (to match our known lipids in milk samples, more milk samples are also frozen). Such further research might detail benefits of high leaf Fat content to animals.

Fats in ruminant diets increase digestive efficiency and decrease methane production, with no ill effects on digestion of up to 6-7% Fat in DM of diets (Hassan et al. 2020).  It intrigues me that these optimum levels cited match content in our leaves; as usual, thousands of years of ruminant consumption of leaves can't be very wrong.       : )

Metabolizable Energy (ME) signifies how much energy is available to animals for all bodily functions, and is used to compute Net Energy figures below.

Net Energy (NEL, NEM and NEG) across averages of our 10 most abundant species of leaf-silage averaged scores of .66 for NELactation, .64 for NEMaintenance, and .38 for NEGrowth. These were higher than levels in either of our hay samples or in Dairy One-tested average grass hay (across better years than 2023). Our NE figures even surpassed those of average Dairy one-tested grass silage levels (.58, .59 & .33). Our leaf-silage .66 NEL fell squarely halfway between Dairy One’s NEL averages for grass silage and corn silage!  The steer Angelo, who was 2 years old and growing, skipped to his leaf-silage offerings. Despite NEM still being low in his forage-only diet (though better with the leaf-silage), he seemed to grow fine – perhaps the Condensed Tannin magnification of protein utilization (see Protein discussion above) helped.

Minerals varied widely in our leaf-silages, as did those in my 2019 samples*, with a trend of higher levels of key elements than in grass forages.  (*2020 testing thanks to a Vt Grass Farmers Network Mini-Grant; that report, Hanson 2020^b, linked here & fully referenced in "Introduction" above, includes a more in-depth look at mineral levels in leaves and livestock needs.)

Calcium was high in our leaf-silage species-averages, ranging from similar DM levels to both our hay samples, to almost 3x as much as in the hay. Across our 10 species-averages, Calcium DM level averaged to be 1.75% of that in average Dairy One-tested grass silage.  I suspect that my lactating goats (and possibly my growing steer as well) preferentially choose forages that are high in calcium.

Zinc was especially high in both Aspen species, with DM levels in the two ensiled Y Knot Farm Big-toothed Aspen samples averaging 273.5 ppm, and 5 of 6 ensiled MOFGA Quaking Aspen samples averaging 143.2 ppm (6 of 6 samples averaged 126.7, but I am tempted to discredit the sample with 28 ppm, harvested amidst all those samples from one stand of trees).  Across our 10 species-averages, Zinc averaged 74.6 ppm, just about twice the 37.8 ppm average level in Dairy One-tested grass silage.

During our SARE FNE22-013 trial, I did not feed animals any ensiled rockweed that I harvest (into barrels, from rocks along the shore of Penobscott Bay at low tide) and usually offer intermittently. That seaweed and a salt block (which I did provide during our trial) are my animals’ only mineral supplementations, excepting dirt that especially the steer eats from upturned tree root-masses (no such root-masses were present in their Winter Trial paddock). I have at times in past tried offering mineral mixes to my goats, but they showed no interest, perhaps because of their high browse and tree matter diet.

If rations include a broad species-range of leaf-silage, free-choice intake of mineral supplements may decrease, or such suplements may become unnecessary.

Note when looking at Ash and Mineral figures of our 2^nd-cut hay that it was full of dirt from that rainy 2023 season, with minerals that my animals left in the offering-sled.  (Our leaf-silages include no dirt.)

Fermentation & pH:  In past I was confused by low 3.7 pH of spring-harvested Beech leaf-silage with almost no fermentation acids detected in any of 5 samples tested (Hanson 2020^b, linked above).  I thought that perhaps different acids were produced (possibly that acidity was from Vitamin C, already present when leaves were fresh?). So we decided last summer to send 4 samples ensiled for more than a year; the usual fermentation acids did appear.  PH went down from fresh to long-ensiled (over 1 year) 1.1 points in Big Toothed Aspen (going from 5.5 to 4.4), zero in July-harvested Beech (which stayed at 5.4 with very low .4% total acids), and 1.5 in Winterberry (going from 5.7 down to 4.2, the acceptable upper limit for grass silage. In the first winter as fermentation slowed, pH was 5.4). Across these three sample-pairs (Beech, Big-Toothed Aspen & Winterberry), fresh pH averaged 5.53, and 1+year ensiled pH averaged 4.67.

The animals consistently mobbed our pleasant-smelling 1+ year-old leaf-silages of palatable species and ate them, as they did in the first winter.

Dairy One FERMENTATION PROFILES & Nutrition, comparing 2018 short fermentation and 2023-'24 1 yr+ fermentation                                              (pdf formatted to print on legal-sized, landscape-oriented paper)

Nutritional Changes from Fresh to Ensiled

We had Dairy One analyse nutrition in 24 matched fresh/ensiled sample-pairs of 23 species = 6 shrub plus 17 tree species (we included two pairs of White Ash, with much healthier-looking leaves from the site with later harvest-date, hence duplication), to look at nutritional changes when ensiled. My summary of most notable changes follows:

Dry Matter (DM) decreased slightly with ensilement, such that both types of

Acid Detergent and Neutral Detergent Fiber (ADF & NDF) both went slightly up as % of DM when ensiled, but since DM decreased, fiber % AS FED stayed stable.

Crude Protein (CP) went up slightly (whether considered as % DM or % AS FED). Within CP,

Rumen-Degradable Protein (RDP) %DM decreased to 95% of Fresh average, which indicates that the increase in CP consisted of rumen-escaping protected proteins. Within decreasing RDP,

Soluble Protein (SP) %DM increased to 103% of the Fresh average. So Rumen-Degradable Insoluble Protein was what decreased, as CP went up as a whole with more rumen-protected protein plus more soluble protein. (I realize I am be-laboring this, and also that my use of SP %DM versus %CP is unconventional, but I am trying to visualize complexity of the changing amounts).

Water-Soluble Carbohydrates (WSC) %DM dropped by 34.5% of fresh level on average when ensiled, across all paired samples,and

Non-Fiber Carbohydrates (NFC) %DM as a whole dropped by 10% of fresh level; this is mosty accounted for by the drop in WSC, within NFC(so mostly sugars, and not much simple starches, probably got used up by acid-producing microbes).

Fat Ether Extract (Fat EE) went up.  Fat content in ensiled samples averaged 111% of % DM levels (or 106% of % AS FED levels) in matched fresh leaf-samples (across 20 sample-pairs with full Fat EE data).  Higher Fat level was fairly consistent across samples; only 2 pairs had higher Fat % DM fresh, or 3 pairs had higher Fat % AS FED fresh, versus ensiled, out of 20 pairs.

This rise of Fat when leaves are ensiled intrigues me; tree physiologist Kevin Smith, UNH Extension, Durham, said it can only happen in aerobic conditions (does it happen immediately upon sealing the barrel?)  Perhaps the rise is in wax coatings, as the leaves suffocate? Do the wax coatings eventually break down through fermentation to become digestible fats?

My animals like most ensiled leaves as well as they do the same fresh, accepting even aged offerings even in summertime, when they wander plus graze pasture. Retired UMaine Cooperative Extension forage specialist Rick Kersbergen once told me “It’s all downhill, from fresh.” Maybe so, but it’s not a steep hill.

Dairy One 2023-2025 data comparing PAIRED FRESH & ENSILED leaf-samples                                                                                                                           (pdf formatted to print on legal-sized, landscape-oriented paper)

Harvest Date Effects

I divided our Fresh/Ensiled sample-pairs into 3 harvest-date categories, to look at the interaction of said-to-be lower protein at later harvest-dates, less warm weather left for fermentation in those harvested later, and any other observations related to date.  Both comparison spreadsheets below have limited validity due to differing harvest-sites per date-period.  The 1st and 2nd pages of the original "Dairy One data comparing PAIRED FRESH & ENSILED leaf-samples" spreadsheet ABOVE are actually as informative, with differing harvest-year groupings averaged across samples.  (Different harvest-year groupings of Fresh samples especially have different average harvest dates, with 2025 having the earl;iest.  Please refer to the "% DM & Harvest-Date Averages" sheet further above, included as a photo and also a link to pdf, to cross-refererence harvest-date averages for different harvest years.)

Dairy One PAIRED FRESH & ENSILED samples, HARVEST-DATE categories COMPARED                                                                                                              (pdfs above & below are formatted to print on legal-sized, landscape-oriented paper.)                                                                                                      Selected leaf-samples comparing 3 HARVEST-DATE categories

 

TOXIN ANALYSES                                  

Cyanide in Cherry Leaves

Initial results: Cyanide (HCN) tests on 4 fresh & ensiled sample-pairs, 3 of Black Cherry and 1 of Pin Cherry, from our 2022-2023 harvests showed very safe reduction of Hydrogen Cyanide with ensiling. Even fresh levels were below the toxic threshold.

2022-'23 ISU VDL data, Cyanide in Cherry Leaves, 4 Fresh & Ensiled Pairs    (<--  pdf link, in case above screenshot is not clear)

Struggles for a Cherry-leaf set with all treatments: In 2024 I made a full set of Y Knot Farm black Cherry samples with all combinations of fresh, wilted, and ensiled, but used double zip-lock quart freezer bags to ensile. This had worked for late-harvested 2023 samples in a basement, but these early-harvested ones in my warm house leaked and molded (the bags also seemed thinner with different texture – did they change design?).

In 2025 I collected samples on May 21^st and June 3^rd from mature Black Cherry trees at Y Knot Farm that were similar to those we harvested there in 2022 and ‘23. Such early-cut cherry leaves are reported to be more toxic, as are wilted leaves. June 3^rd to 4^th, I packed my 3nd try at a full matched array of 12 samples: fresh, ensiled 30 days, ensiled 60 days, and ensiled 90 days, then after branches wilted for 4 hrs in the trailer, 2 similar sets of 4 but first leaving one set to wilt for 4 hours and the other for 24 hours, before packing and sealing, this time in plastic wide-mouth laboratory containers with foam lid-liners (that worked). I also made a Dried sample from same harvest, but unfortunately packed that in bags (why that mattered becomes evident, if you read far enough into my typing).

All spring I was also collecting sample-pairs for Wayne Zeller’s C Tannin work, and the ensiling ones for both purposes cluttered my cupboard, such that I did not notice my oversight of the 4 hrs wilted “fresh” Black Cherry sample, that had gotten in there instead of directly into the frreezer (my farm gremlins are interested to participate in anything I do here).

So on July 3^rd I harvested my own stand of small Black Cherry pollards, selected new growth sprouts for higher Cyanide content (since the date was late), and made my 3^rd try at a complete set of 12 samples that day and the next (again in plastic laboratory jars, and this time all made it into the freezer on designated days as planned!). I ran brush with remaining leaves of older growth through the Chain-Flail Leaf-Separator, to almost fill one barrel (a winter treat for my animals).

I did not have a protocol of weighing samples for equivalent air inclusion, and on retrospect should have done this. A few samples had a piece of bubble-wrap inserted at top to take up a bit of extra space (when leaves from each treatment were running out) - instead of a piece of paper that I use to make sure leaves don’t get under the sealing edge - and I failed to note which ones had that extra space with bubble-wrap.

It is a daunting and pressured task for one farmer (whose animals rely on me wandering with them most of every day, and/or climbing trees) to get all those samples harvested then hand-stripped and packed, at right stages of wilt across two days. Choosing to up Cyanide potential by using new growth only precluded use of the SARE FNE22-013 Chain-Flail Leaf Separator, and even when I do use it, I remove sticks and twigs further by hand. Each quart jar packed tightly with leaves and labelled took me about a half hour, so the “wilted 4 hrs” and “wilted 24 hrs” are approximate.

A barrel of Black Cherry leaves from my sampling harvest, & Observations:  On December 24^th and December 25^th, I fed that barrel of older-growth leaves from the July 3^rd harvest to 7 happy goats and the steer, Angelo.

Angelo ate a good quantity from a sled, for about ½ hour, then took a hay and water break, came back to eat more and I’d passed it to the goats. So he asked for me to open the barrel, which I did. The second day he had a sled-full and took the same break – can he sense Cyanide?  I sent in no test of those older leaves (I wished to, but was already sending more tests than the grant funds were covering). The 7 goats did not have enough between them, so ate all straight up.

An animal can eat the toxic dose in a day and be fine, if they do not eat it all at once; Cyanide acts quickly, but then also exits the body quickly (at least when not fatal).

During our SARE FNE22-013 winter 2023-’24 trial, my animals ate free choice amounts of ensiled Black Cherry during 2 hr offering-periods, but with 2 other leaf-silage species simultaneously offered – except on one day (February 7^th, the 5^th measurement day of 7 in the last trial-rotation, when barrels were almost gone). That day, goats got just 2 species, Quaking Aspen and Black Cherry. Angelo was entirely refusing Quaking Aspen, so he ONLY got Cherry; he’d had one previous day like that (also in last rotation, when goats got Cherry along with Quaking Aspen and least palatable Gray Birch). Both those days Angelo was offered 14+ lbs as fed; on the earlier date, he ate 11.5 lbs and on the latter 9.25 lbs as fed; the 1^st barrel (#9, harvested 6/30/23 and packed 2 days later in gray weather at MOFGA) had 40.3%DM, and the 2^nd (#46, harvested and packed on 9/2/23 at Y Knot Farm) had 41.9%DM, so he ate 4.63 lbs DM and 3.88 lbs DM of ensiled Black Cherry leaves, respectively.

2 to 2.5 mg Cyanide per kg body weight is considered a fatal dose. Some Cyanide levels in my 2025 samples below, ensiled in small jars versus barrels, hand-stripped rather than machine-separated, frozen at lower temperatures than barrels outdoors, and with new growth selected separately from old, would have been sufficient to kill Angelo twice over, if levels were as high in Black Cherry leaf-silage that he ate so well on each of the above-discussed winter 2023-’24 trial days.

I saw my animals for the first time sparsely eat (versus gobble) fresh-cut Black Cherry from a lightly pollarded tree in their pasture this past early summer (2025). They did finished all leaves within a day. I really do conjecture that they know when to stop eating. Susan Littlefield, Y Knot Farm, has bountiful Black Cherry pollards and observes similar behavior in her milking sheep.

2025 (strange) results: After a busy summer, I finally sent the July 3^rd, 2025 frozen multiple-treatment 12-sample set, 2 Fresh samples with 5/21 and 6/3 harvest-dates, a matching Dried 6/3/25 sample, all Black Cherry, plus 1 Fresh 7/13-harvested sample of Choke Cherry, to Iowa State University Veterinary Diagnostic Laboratory (ISU VDL), and received results in October.  Those results were not what either Scott Radke, ISU VDL Toxicologist and Director, nor I expected.

2025 ISU VDL data, Cyanide in Cherry Leaves, updated   (<--pdf link, if screenshot above is not clear)

2025 ISU VDL Cyanide (HCN) results had one trend opposite of expected, and two sample results that were extreme exceptions to expected patterns. The fresh 7/3/25 sample, with leaves selected from new growth sprouts only, had higher Cyanide than did the same “wilted 4 hrs,” and then “wilted 24 hrs” had even less.  

Perhaps the samples dried somewhat versus wilting; wilting is supposed to increase Cyanide, and drying is supposed to reduce Cyanide. But the 6/3/25 pair had 2x the fatal threshold level in the Dried sample versus the Fresh, which was well below toxic level.  The dried sample somehow had gotten just as wet as the fresh; later a 2^nd run of the moisture test registered half as much, but still 3x as wet as dried 2018 samples which I sent for nutritional testing in 2020. ( They started out crisp; I regret double-bagging versus using a plastic lab jar with better seal.)

The commonly understood trend of Cyanide reduction when ensiled loosely held but with one frightening exception: 7/3/25 “Ensiled 90 days” sample had a bit more Cyanide than did the matched Fresh sample, jumping it to just above the fatal threshold. In the same matched set, 30- and 60-day ensiled samples had less than the Fresh sample, with a less-than-measurable amount in the “Ensiled 60 days” sample.  

On 1/6/26, he lab kindly re-tested both the 6/3/25 Dried and the 7/3/25 4 hrs wilted ensiled 90 days samples above, that had such unexpectedly high levels.  New levels were even slightly higher (new and old figures are included in the above spreadsheet).  

I opened an unsent 6/3/25 ensiled then frozen sample, and it seemed properly fermented, despite the small quart container. Was there some way that Cyanide released by leaves got trapped in a non-gaseous form in my sealed containers, and so didn’t dissipate when opened?

What was happening?  Scott Radke, Toxicologist and Director at ISU VDL, took generous amounts of time, intermittently for the next three months, consulting his past toxicology mentor, talking to me, and authorizing bits of re-testing at the lab.

Here are some details we pawed over, beyond my confession above about differing packing densities: All 2025 samples were frozen (some previous 2022-’23 ensiled samples were not frozen); 2025 samples were ensiled in containers that were not opened prior to testing (2022-’23 samples were drawn from larger containers into double bags); the 2025 dried sample which got so wet was double-bagged & frozen, but in ULine bags with imperfect zippers. All 2025 samples were packed and the whole inner box went back in the freezer, before wrapping in a plastic bag and additional outer box at time of sending, a day or two later. Then all were sent in that one package by FEDEX overnight service (2022-’23 samples went piecemeal by UPS 2-day or slower ground service).

None of this got us much closer to understanding, except to recommend future studies using more separated packing with additional layers of plastic (that freezer does have a lot of frozen condensation, and perhaps condensation on cold samples once laid out to thaw at the lab might have gotten into the imperfect zipper-seals of that Dried sample?), possibly more real-life large-batch ensiling, and at least more careful weighing for equivalent packing density, and note-taking… if ever someone takes this up where I’ve left off.        : )

Scott Radke sent me an article (linked below), about similarly irregular high Cyanide (a.k.a. Prussic Acid) results from drying regrowth of Sorghum. The authors indicate that if leaves are intact versus damaged, they can re-activate & make more HCN, even when dried.

https://enewsletters.k-state.edu/beeftips/2025/09/01/good-news-and-bad-news-on-prussic-acid/

I machine-stripped the 5/21 Fresh sample, & then hand-stripped to remove twigs; 2022-’23 samples were all machine-stripped then sorted similarly. But I used no machine & only hand-stripped the 7/3/25 samples, by running my hand down new growth only (hoping to get the highest Cyanide level in this young growth). I may have also hand-stripped the 6/3/25 dried sample. Perhaps doing that was less damaging? Yet our machine-separated leaves look mostly intact. And I would have thought that drying for many days in a sunny room & then freezing would cause sufficient damage to release all Cyanide potential, in that fatal dried sample.

As I type on January 11^th, I await re-testing of Cyanide levels in those 2 worst samples, which Scott Radke has kindly gotten authorized for no additional charge.

Please if you know more about chemistry which may have happened to cause such results, CALL me! & leave Voicemail: (207) 338-3301

 

Gallic, Ellagic, & 4 non-proteinogenic Amino, Acid Toxins in Maples, Box Elder & Staghorn Sumac

A goat at Locust Grove Woodworks, Unity ME, died a few years ago of bloat from Box Elder, when fed in a stall with limited choices. Leaves of large, vigorous coppiced Box Elder from Diane and Kevin Weisner’s sunny roadside, Hubbard Brook Farm, Hunt Rd. Unity, ME, were eaten without issue by my steer and goats at June 28^th, 2024 time of sampling for metabolomics work, but were refused when harvested from the same site in 2025, on May 1^st, May 23^rd, and June 21^st (my palatability ratings were .5, .5 and 0 respectively, on a scale of 0 to 3, with 0=“refused without tasting,” 1=“tasted,” 2=“eventually consumed) and 3=“immediately consumed”).

Huge healthy growth of a pollarded tree in Unity at MOFGA South Orchard harvested on July 20, 2024 was also refused, when fresh and when ensiled. Yet at least 2 farmers, one in Windsor ME and one in NY State, have told me that their animals use significant amounts of Box Elder without issue.

Diane Weisner (Hubbard Brook Farm, Hunt Rd source of my samples) suggests feeding more repeatedly, to develop familiarity, or in case my animals’ rumen microbes need some shift specific to Box Elder. My animals can count on long daily to twice-daily attended free wanders, or fresh Red Maple in winter, so I’m not sure this would work – though my animals have exhibited evidence of a rumen-adjustment process with more palatable White Birch, when felled in late June.

The Kitchen Box Elder (also previously pollarded) in Unity at MOFGA had a 2x+ greater spike of Hypoglycin B than our other 2 Box Elder samples in 2024. I re-pollarded it on June 21, 2025, and dumped it in shade by my driveway, where my animals passed by but (not surprisingly) took no interest.

(Before sampling, I spent most of a day driving and hiking, looking for Box Elders nearer than Unity. The only nearer Box Elder I found is along the downtown Belfast Rail Trail, in a narrow patch of well-disturbed soil between the rocky bay-front and the parking lot of the old potatoe factory. Karl Hallen, my Tech Advisor, sees lots of Box Elder in NY State and has observed them to thrive in toxic places.)

As noted in Westermann (2016), regarding levels of Hypoglycin A in Maple species), toxin levels can vary greatly from tree to tree. Our species with highest spikes of each toxin had broadly ranging levels among (3) same-species samples, mostly* taken within a day of each other. My animals will browse individual trees or shrubs selectively.

* Jack Kertez froze 2 samples of Box Elder from MOFGA, Unity, for me on the right day, while I was sampling in Belfast. When I fetched them, the trip home to my freezer had other stops; the samples thawed and I worried that their condition was too deteriorated, or at least dissimilar to other samples. So I collected 2 fresh Unity re-dos 4 days after the Belfast sampling.

The animals’ discernment is much more affordable than metabolomic analyses (plus the lab did not have the ability to quantify above toxins other than Gallic and Ellagic Acids), but sometimes the animals aren’t present for my forage harvest. Scandinavian farmers of antiquity were said to taste tree leaves themselves, to know when to cut; this pertained to timing for best nutrition and tree carbohydrate stocking (probably not to toxin levels?). Short of that skill (though my NOSE DOES say Box Elder is not tasty when wilted or ensiled), and hoping that your animals are discerning enough to also guide you, I proceed to report comparison of our 3 samples-per-species average measurements (Gallic & Ellagic Acids) and comparative spikes (Hypoglicins and homologues):

Box Elder metabolomic analyses showed spikes of Hypoglycin B (HGB) and of related γ-glutamyl-MCPrG in each of 3 samples to be drastically higher than in our other species, with HGB barely present at all in other species. Box Elder was 2^nd highest in HGA (to Sugar Maple, though one Sugar Maple sample was much lower), but lower in average spike-level of HGA homologue Methylenecyclpropylglycine (MCPrG) than all other species tested, with range similar to that of Staghorn Sumac and Norway Maple.  

Our spikes of HGB in Box Elder leaves were consistently much higher than spikes of HGA; this was in contrast to findings of El-Khatib et al. (2022), who reported spike-average of HGB in leaves to be 26% of HGA spike-average.  They quantified HGA in their leaves at 535 mg/kg DM. 

Red Maple leaves are used by my livestock if eaten alternately with other forages over time.  In our SARE FNE22-013 3 Streams Farm trial, goats averaged a low 1 lb As Fed ( lbs DM) per goat in the 2 hr offering-periods, with 1.5 lbs As Fed ( lbs DM) maximum average per goat (across 10 goats), and steer Angelo averaged 1/2  lb As Fed ( lbs DM) with 1 lb maximum ( lbs DM); 2 other leaf-silage species usually were offered just after interest in Red Maple slowed, during the 2 hr periods.  Animals then finished last leaves plus all the twigs overnight.  Winter twigs and bark are an immediately-consumed staple for us, with seemingly less digestive limitation than leaves. When I have brought home Sugar Maple, animal response has been similar (but with slightly less choice twigs and bark).

Red Maple leaf results show Free* Gallic Acid (GA) and Methylenecyclpropylglycine (MCPrG – homologue of Hypoglycin A) to have much higher spike averages than in the other species tested. All species tested showed HGA; Red Maple had the 2^nd lowest level (Norway Maple had the lowest).

Both Gallic and Ellagic Acids have been batted around in research as potential additive to cattle feed to reduce methane emissions, or other benefits.  Here's the latest quite detailed study I found, with conclusion that "more research is needed," before feeding these acids in vivo to cattle:  Manoni et al. 2024:  Gallic & Ellagic Acids Differentially Affect Microbial Community Structures...

More surely, Gallic Acid does have positive antioxidant health benefits for humans (Wianowska & Olszowy-Tomczyk 2023:  A Concise Profile of Gallic Acid...).  I live on Goats' milk, and the goats (and steer) do eat some Red Maple leaves regularly.  I did not find anything about whether Gallic Acid makes it into milk (clearly not enough to curdle the milk, thank goodness).

*A discrepancy seemed to exist in the Free and Hydrolyzed (which was supposed to indicate Total GA, including free GA plus GA in more complex molecules such as toxic hydrolyzable Gallotannin) GA figures. In 7 of 12 samples (across species and including all 3 samples of Red Maple), Free GA level was higher than Hydrolyzed.  Ellagic acid (EA) figures less strongly had the same issue.  

Zhentian Lei explained (1/16/25 email): “ The hydrolyzable data (hydrolyzable gallic acid and hydrolyzable ellagic acid) contain the free data. Because the hydrolysis was performed at high temperatures in strong acid, some free gallic acid could be lost due to degradation. Thus, the free gallic acid is a better indicator of its content in the samples while hydrolyzable ellagic is a better indicator of ellagic content in the samples.”

Sugar Maple had contrastingly low levels of the 2 toxins in Red Maple, while Hydrolyzed Ellagic Acid and Hypoglycin A were much higher on average in Sugar Maple than in ant other species.

It interests me that Hypoglycin A (HGA) showed a much higher spike in our Sugar Maple leaves than in our Box Elder leaves, and (homologue of HGA) Methylenecyclpropylglycine (MCPrG) showed a much higher spike in our Red Maple leaves than in those of Box Elder. I had found these toxins mentioned in literature as present in Box Elder and Sycamore leaves (plus maples in the Netherlands that we do not have: Westermann et al. 2016).  On reviewing saved sources of info, I did find two studies confirming HGA in Sugar Maple (Novotná et al. 2023; Fowden & Pratt 1973).

El-Khatib et al. wrote that Hypoglycin B (HGB) and its homologue γ-glutamyl-MCPrG had been found in maple species 50 years before (Fowden & Pratt 1973, mentioned above) but were subsequently overlooked until their 2022 study, and all that time HGA was assumed to be the sole cause of poisoning from both Box Elder and Sycamore. (As mentioned above, we found higher HGA in Sugar Maple leaves than in those of Box Elder).

Norway Maple is a highly edible fodder tree from Europe, with leaves that my cattle and goats consistently devour. The leaves stay bright green when ensiled. Our results confirmed low presence of toxins indicated by other researchers (El-Khatib et al. 2022, Westermann et al. 2016).

Striped Maple bark is stripped from young trees by my goats, before it gets established here; I therefore thought they might like leaves from my blueberry property, but they didn’t like them as well as I expected.  Wayne Zeller screened relative level of Condensed Tannin in Striped Maple at “10” (on a 1 to 10 scale; Norway Maple was scored 3.5, and Red & Sugar Maple scored 5); I'm guessing that some of the toxins in other maples are there as well, and I'm sad that due to budgetary limitatiions, we did not include Striped maple in metabolomic analyses.

My palatability ratings for their spring of 2025 responses were 1.5= “tasted and ate a bit,” when offered out on a walk, and 2= “eventually consumed” to a rare 3= “immediately consumed” when small qiantities were given in their yard. On June 1^st , 2025 Angelo ate some, then left the rest and went to point to Gray Birch in a sled outside the fence. June 24^th was no better. September 6^th through leaf-drop we lived on the blueberry parcel, and I never saw Striped Maple selected.

Staghorn Sumac is considered by Susan Littlefield to be a choice forage for her Y Knot Farm sheep; my goats refused senescing leaves but ate a few berries, in October of 2023 (the only time we were near some, that year). In 2025 they accepted some in their yard on May 28^th, then we stayed on my blueberry field where it is a young weed, for most of June and then September 6^th through December 6^th. They hardly ate any leaves of that young sumac, but on the way there on September 6^th, they briefly gobbled some leaves of a taller specimen along a shady road.  Once leaves fell, they started stripping bark of the young field specimens quite eagerly.

Staghorn Sumac results showed second highest average Free GA level (well behind Red Maple), but higher Hydrolyzed GA average than Red Maple, with one sample twice as high as the highest Red Maple sample.  The Staghorn sumac sample with lowest GA figures had HGA spike in range of that in our Box Elder samples. 

As above described, my animals have given preliminary indication that as with Red and Sugar Maple, the bark has less or no toxins as compared to leaves, so fresh cutting and feeding in winter may be better use than ensiling summer leaves.

MU Metabolomic Toxin Results on Maple Species & Sumac, Final                                                                                                                                                                (pdfs above & below are formatted to print on legal-sized, landscape-oriented paper.)                                                                                                                      MU Metabolomics Charts & Graphs made by Gentian Lei (edited slightly by Shana) 

 

Antifeedants in Birches:  I have in past looked for chemical info explaining my animals intermittent use of Gray Birch with a lot of refusals once stored, and somewhat less limited use of White Birch, to no avail.  Yet just now I Brave-searched "Antifeedants in Gray Birch that inhibit browsing by animals" and got this:  "plant secondary metabolites, including alkaloids, tannins, and terpenes, are known to contribute to antifeedant properties in birch species. These compounds are often bitter or toxic, reducing palatability and deterring feeding. For example, bark phenolics and volatile terpenes in birch trees have been shown to reduce herbivore feeding behavior."  Yet still I found no studies specific to birch, and cannot (at this late date of report submission) chase the references in general antifeedant studies.  Please call me if you do so, with palatable morsels of info for me to ruminate on, while I watch my herd devour Gray Birch fresh, next spring or fall: (207) 338-3301.

Yet I included this section, to at least summarize my animals' limited intake amounts of Gray Birch leaf-silage from our SARE FNE22-013 3 Streams Farm trial, and tell you that our quite eager fresh eating of Gray Birch in spring or fall is more predictable and worthwhile than our storage efforts, either ensiled or dried. 

White Birch leaves are preferred to Gray Birch when fresh or ensiled, with spring or fall harvest still preferred, and hybrids of the two seem to retain benefits of White Birch (wild hybrids seem to abound on our blueberry land, but as yet least palatable but best soil pioneer Gray Birch dominates). 

Yellow Birch leaves are well-received all season whether fresh or ensiled. 

If leaf-species mixes are to include Gray Birch, perhaps intake levels during the once-per-day 2-hr offering-periods in our SARE FNE22-013 3 Streams Farm trial can give some guidance; note that once interest waned for these species, I then stopped adding more and started offering other leaf-species (with aim for them to eat leaf-silage to full capacity, during those 2 hrs).   

Gray Birch intake levels are offered here above levels of next most intake-limited Red Maple leaf-silage (discussed under Red maple in Maple Toxin section above). 

 

CONDENSED TANNINS (CT) *

Wayne Zeller, US Dairy Forage Research Center in Madison, WI, screened and continues to screen tiny fresh leaf-samples for comparative levels of Condensed Tannins (initially we used one ensiled: Quaking Aspen, as I had no fresh saved when we started collaborating). Some samples came from our 2023 harvests, plus I collected many more samples, including fresh/ensiled matched pairs of large samples, in 2025. Wayne is isolating, purifying, and identifying tannins from large-sample pairs of all species rating “5” or higher (and he is right now grinding leaves in order to screen the additional 2025 species). So far, CT in 4 species, including that in fresh and that in ensiled leaves, has been isolated, purified and identified. Wayne has been short of help at the lab, but a student will join him in this work very soon.  They have fresh/ensiled sample-pairs of 22 to 31 species left to do! We’re not sure how many because 9 depend on unfinished screening results.

Wayne Zeller's C Tannin Screening Method & Results with Photos, my slides and 3 of Wayne's used in NOFA MA 2024

To explore change in relative CT level across different harvest-dates, Wayne and his student will also be grinding and screening lots more tiny samples of a few species, just for my own curiosity. It’s my fault that 9 species-screens are not done, as Wayne was waiting for all my date-screen samples to be there, to dry, grind and then screen all at once. If his results on my date-screens are complete soon after SARE’s report deadline, I will update this report (per NESARE permission), to include those results here.

Wayne’s CT identifications are breaking ground for other researchers to look at effects of woody forages on cattle protein utilization, methane emission and digestive efficiency, and internal parasites. Wayne is already collaborating with another USDA ARS researcher who will be running methane and digestive tests, in a continuous-feed rumen-simulator. (When I called him this week, he was in midst of dividing out and packing portions of my leaf-samples to send to her.)

Watch the scholarly journals for their articles, hopefully to be published by end of 2026 or start of 2027.

Wayne will soon retire, but he is committed to finishing this woody chunk of work with me, first. We need a younger scientist to take up his baton of CT work with woody forages in particular; most CT study is aimed at leguminous field crops (Alfalfa, White Clover) bio-engineered to produce CT, or naturally CT-producing field legume Birdsfoot Trefoil, versus historic and ubiquitous woody tree/shrub CT sources with as yet unidentified CT structures (until Wayne completes these).

When Wayne retires, the scientific community will also experience a gap in CT expertise in general, as Wayne currently performs expert CT services for other researchers with a full spectrum of CT-related studies (he has those bioengineered crops in his freezer, along with my leaves).

Thank you, Wayne, and also a big thank you to Andrea Clemensen, Environmental Biologist, USDA ARS Northern Great Plains Agricultural Research Center, for connecting us. Andrea co-authored an excellent article on Tanniferous Forages. I found her through that article, and she immediately agreed to a phone conversation, then organized a one-time multi-researcher zoom meeting where Karl Hallen and I described what we were doing, to foster collaborations. It worked!

* All of this CT work was/is above and beyond our SARE committed research plans.  

 

 

                                        RATION-PRODUCING STRATEGIES, & SETTING RATION PROPORTIONS                                                                        with regard to toxic leaves, tannins, & other intake-limiting antifeedants                               

Intake Summary per 9 Species of Leaf-Silage, offered as Rationing Guidance:  I made a new spreadsheet using data from the 2023-'24 SARE FNE22-013 livestock trial at my farm, to make best use of animal guidance.  Toxin and antifeedant information are spotty at best, with complexity of numerous compounds that the animals do seem to be better able to sort, than are researchers and laboratories at this point.  At end of this spreadsheet, I've ordered leaf-silage species by intake of goats (most to least), and then of the steer.  Intake levels were measured during once-per-day 2-hr offering-periods, with late-cut 1st-cut hay offered 24 hrs per day (no concentrates).  Note that 3 leaf-species were generally offered side by side. These intake amounts are therefore NOT indicative of one leaf-species limits (excepting for Gray Birch and Red Maple, which I tended to offer first, in order to get animals to eat any at all).

3 Streams Trial Goat & Steer Ave & Max Intake Lbs, & Preferences Alongside Nutritional Data

(For full report on that trial, look for Hanson 2020^a full reference for SARE FNE22-013, at end of "Introduction" section above.)

 

Efficient Production of Leaf-Silage, & Species-Mixing:  Karl Hallen, my Tech Advisor and machine creator, is well-educated in cattle nutrition and keeps abreast of new info, used to dairy farm, and is a farm and forestry consultant on the side of his SUNY ESF Willow Biomass Project job. He thinks that for broad farm use of wild woody forage, woody harvests cannot efficiently descriminate by species, so mixing leaves in smaller pieces should be tried. Most* toxin-levels in Northeastern tree/shrub species can be sufficiently diluted through such mixing, to negate danger. (* I say “most” because I am not sure about Lambs Kill, our wild laurel. Other well-known Northeastern toxins are in herbs such as Helebore and Hemlocks in the carrot family, not in wild woody forages.)

I’m not sure that such mixing will work. My animals’ input is that when species are palatable but poisonous, as in the case of Cherry species, mixing might be a perfect way to remove risk. But when toxic species are unpalatable (my animals indicate that’s the case with most Box Elder batches), I fear that mixing will lead to poor utilization, thus wasting the tasty species along with the unpalatable. If you don’t like Lima beans and asparagus, will you happily eat them if I mash them into a soup of other vegetables?

Very few species are unpalatable at all times, nor toxic at all times. Many species become unpalatable during certain parts of the growing season (Gray Birch being the most troublesome and hard-to-guess example of this on my farm). Also, animal groups differ in what they find palatable; this may reflect differing acculturation of their rumen biome, differing soils on their farms, different dietary array available, specific content that they are lacking, simple habituation, or other factors that the animals will have to tell us.         : )

When/if wholesale industrial harvests of woody matter become available to farms, some indication of species content proportions or at least species array should be provided to receiving farms (and possibly Box Elder should be left in the woods – but I am biased on behalf of my own animals).

When farms develop their own wild woody forage edges, or develop fodder trees in fields (using existing or planted trees), or make whole areas of pollarded woodland on-farm, they can certainly respond to literal feedback from their animals, in decisions on area-specific and tree/shrub species-specific treatments. Excepting loads with known potentially fatal toxicity, bring a load of brush to animals whenever you are cutting, and if not well-utilized, try again for 4 days if you can. Sometimes animals’ digestive microbes just need to populate in response to a new feed.

In such on-farm design, be alerted that stage of growth is yet another factor affecting palatability and utilization; my animals usually prefer tree leaves from well-developed growth out of reach versus from young coppice. I suspect that this is due to appropriate tree allocation of defensive chemicals (known as antifeedants).

On-farm woody forage designs can provide industrial-style harvests but with optimal species array with optimal years of rest between harvests, and optimal harvest-season timing (for best balance of palatability, nutritional value, and ongoing tree energy and health). When woody resources are developed on-farm or in leased areas that farmers control, all these factors can be fine-tuned to be optimally farm-specific, This includes herd specific, tree/shrub-growth specific, and yes, farmer-work-rhythm, farmer-method-preferences and any other farmer happiness-related specific. Susan’s pollards are shorter than mine.        : )

Some mechanical harvest equipment is readily available and in the price-range of other farm implements; for instance a sturdy sickle-bar on jointed arm or even one that simply tips to cut vertically can work when tree rows are trained as hedges for frequent (annual?) harvest. These small cuttings can be baled with normal hay equipment. Purchase of high-priced best industrial equipment to develop and harvest full-height side-canopies along fields will probably require farmer-cooperative arrangements for shared use and repair.

Karl and I suspect that aromatic compounds and countless other tree/shrub species-specific antifeedants (Hassan et al. 2020) that are not measured in nutritional testing, nor explored in my studies (except for unmeasured enjoyment of aromas when I open containers or barrels), are a huge factor affecting animal intake of tree and shrub matter.  These compounds have positive medicinal qualities at correct levels, and also have intake limits, with anti-nutritiional effects when too much is eaten. These compounds differ per tree/shrub species, and animals balance them against each other with healthful effect. Again, offering multiple species separately is ideal to support animal selection, and mixing of species may work.

Once a significant quantity of woody forage can be stored, with broad mix of regionally prevalent species or better selection, in a form compatible with farm feed-out and waste-management methods, on-farm and researcher trials with greater numbers of animals can give better rationing recommendations than can I. I suspect that farmers and researchers will find that free-choice intake levels range, across different herds with differing diets, from minimal supplement to substantial staple, when offered along with usual feedstuffs (my steer ate leaf-silage as 1/3 his diet, while Nathan and Beth Zimmermans’ cattle receiving optimal grass silage at Palmer Hill Farm only nibbled on pretty choice willow leaf-silage).

I suspect that farmers and researchers will have great animal performance results when mixed woody species or optimal species ration percentage is set based upon such herd-specific supplementary free-choice intake rates. I suspect that at much higher % of rations (in case of grass-based forage sources becoming limited), such as 30-60% of forages for cattle, 60-100% for goats, and perhaps 40-80% for sheep, performance will drop mininmally, or be stable or even improved, once animals' rumen microbiomes adjust.  Stable to improved is most likely when protein level is supplemented by that in remaining grass-forage proportion or concentrates, and of course larger farms and researchers would balance all nutrients within TMR (Total Mixed Ration). Based upon high energy levels in tree/shrub leaves (as discussed under NFC above), animals will require less total amounts of forage.

As on-farm need for shade and environmental cooling (through evapotransporation of leaf-surfaces) increases, and agroforestry and silvopastural systems take off, impetus is rising for others to follow my small farmer studies with more solid next steps, to fill out and firm up our still-spotty knowledge bank supporting use of temperate woody forages.

 

REFERENCES CITED, but NOT LINKED within text above:

Austad, I. & Hauge, L. (2014). Trær og tradisjon. Bruk av lauvtrær i kulturlandscapet. (Trees and tradition: Use of leaf-trees in the cultural landscape.) Fagbokforlaget. ISBN: 978-82-11-01905-9.

 

El-Khatib, Ahmed H., Anna Maria Engel, and Stefan Weigel (2022). Co-Occurrence of Hypoglycin A and Hypoglycin B in Sycamore and Box Elder Maple Proved by LC-MS/MS and LC-HR-MS Toxins (Basel). 2022 Sep; 14(9): 608. Published online 2022 Sep 1.  doi:10.3390/toxins14090608

 

Fowden, Leslie & Helen M. Pratt (1973). Cyclopropylamino acids of the genus Acer: Distribution and biosynthesis. Phytochemistry, Volume 12, Issue 7, 1973, pp 1677-1681, ISSN 0031-9422, https://doi.org/10.1016/0031-9422(73)80387-5

 

Hassan, F., Arshad, M.A., Ebeid, H.M., Rehman, M.S., Khan, M.S., Shahid, S. and Yang, C. (2020). Phytogenic Additives Can Modulate Rumen Microbiome to Mediate Fermentation Kinetics and Methanogenesis through Exploiting Diet–Microbe Interaction. Fronteir Veterinary Science, 11 November 2020, Sec. Animal Nutrition and Metabolism, Volume 7 - 2020.  https://doi.org/10.3389/fvets.2020.575801

 

Manoni M, Gschwend F, Amelchanka S, Terranova M, Pinotti L, Widmer F, Silacci P, Tretola M. Galliv and Ellagic Acids Differentially Affect Microbial Community Structures and Methane Emission When Using a Rumen Simulation Technique. Journal of Agricultural and Food Chemistry. 2024 Dec 11;72(49):27163-27176. https://doi.org/10.1021/acs.jafc.4c06214

Novotná T, Jahn P, Šamonilová E, Kabešová M, Pospíšilová S, Maršálek P. (2003). Hypoglycin A in Acer Genus Plants. Toxicon. 2023 Oct;234:107271. https://doi.org/10.1016/j.toxicon.2023.107271

 

Westermann CM, van Leeuwen R, van Raamsdonk LW, Mol HG (2016).  Hypoglycin A Concentrations in Maple Tree Species in the Netherlands and the Occurrence of Atypical Myopathy in Horses .  J Vet Intern Med. 2016 May;30(3):880-4. https://doi.org/10.1111/jvim.13927

 

Whistance, Lindsay (2018). Could Trees Supplement Forage? July 27, 2018 , https://www.soilassociation.org/farmers-growers/farming-news/2018/july/27/could-trees-supplement-forage/