Efficient Leaf-dense Tree/Shrub Silage Production from Field Edges: Climate-resilient Winter Forage Supplement for Cattle, Sheep, and Goats

CHAIN-FLAIL LEAF-SEPARATOR:

To see all sides of this functional creation, open the pdf below:

Chain-Flail Leaf-Separator Machine Photo-Diagram 2025

Following day-dreams: I had thought and consulted (with Doak’s machine, and with AMS, UMaine) about how to design a leaf-separating machine, on and off for years. It came to mind every time I hand-stripped. One day I decided to look at what the logging world had for de-limbers. I decided that a chain-flail de-limber would handle feed-stock variability better than a pull-through de-limber. Both are huge dangerous machines that one approaches from within the protective cab of a loader. (Nothing that comes off the logs is collected.)

I “went” to current Tech Advisor Brett Chedzoy’s Silvopasture webinar, for a virtual look at his farm; I also wanted us to know each other better, to approach him to be my Tech Advisor for a next step (this project!). In the discussion period, I put out a verbal query: “I’m looking for someone to build me a mini chain-flail de-limber, that will just take the leaves off.” Someone fron SUNY was there, and sent me to someone else at SUNY ESF Willow Biomass.

Next day, that person said to call Karl Hallen. Karl answered, and said ”I’ve built one of those before, but it was to make clean wood pellets, so it didn’t collect the leaves.” Then he said he used to dairy farm, cattle then goats.

That fall, our pre-proposal for Novel approaches was rejected. I emailed “We didn’t get it...but I need this machine! You wouldn’t by chance make it in exchange for cheese?” (Y Knot farmer Susan Littlefield had been practicing making aged cheeses with my cow’s and goats’ milk all summer, as some of my usual milk-drinkers were on vacation). He replied that he might do it for cheese : ) I borrowed Susan’s car for the interstate trip, took a wrong turn and arrived at ESF Tully willow Biomss Project in the wee hours, but got Karl to immediately help me finish a Farmer Grant application regardless. Karl returned once there was daylight, to receive a large box of cheeses, jars of milk, box of apples, and bucket of cider. He gave me a quick tour of hundreds of willow varieties, and had me visually show him what size and shape of tree matter needed to run through the hypothetical machine. I had to leave that same morning, to get back and milk goats.

We got that Farmer Grant award (this project!).

Basic machine description: 18 flails=9 per each of 2 vertical rotating square rotors, are attached in spiral pattern; the spiral pattern helps to blow leaves downward into a bin. (Square rotors are not ideal; see “Recommended Machine Changes” below.) Each flail consists of 5 or 6 links (5 on lowest ones so they don’t hit the frame) of 3/8” high-tensile steel chain, large, heavy and durable, such as used to tie a bulldozer to a trailer.

The main feed system consists of ingoing and outgoing pairs of rollers each made by stacking 8 small tires vertically onto 4threaded rods. Branches and trunks must be straight and stiff enough to manually aim across the flails toward center of the rear rollers, as the front rollers pull them in. A wavy interlocking gathering belt assembly for a corn head on a John Deer forage harvester alternatively drags short or floppy branches through the chains from above. Two flow control valves set variable speeds for feed system and flails.

A gas engine powers the hydraulic pump and system. All components are mounted on a 2 axle farm trailer. A large bin, cut to fit, slides underneath once parked, to catch the leaves.

***Machine Demonstration: We hope to have Bailey Riordan’s Video: “Chain-Flail Leaf-Separator Machine Demo, Common Ground Country Fair 2024” under Information Products above near top of this report, soon.  ***

Here is a short video, recorded by Colin Yarnell, of myself feeding the Leaf-Separator a long Gray Birch tree-top, with branches "braided" over each other so as not to tangle:

Colin's best video of Shana feeding a very long trunk of gray birch to the leaf separator, and then walking around to catch it.

...and a very short video I made single-handedly, that shows how messy our machine was before I added hopper panels and a fitted bin:

Black cherry going into the leaf separator June 27th, set to 9 on flail flow valve 

Machine development, technical changes: Karl Hallen broke five ribs, punctured his diaphram, and collapsed one lung falling onto a different machine as he climbed off the fork-lift, in late winter 2022 (he’d have done more before that, but the garage at SUNY ESF, Tully NY is un-insulated). After convalescence, he configured, reconfigured 3 times, and completed our machine fabrication, delivered on September 29, 2022 (we’d originally planned to start harvest in June 2022).

Karl harvested with myself and Susan Littlefield at her Y Knot Farm that day of machine delivery, and the next. After hard use, almost every imported hydraulic hose fitting started to leak. Karl tightened all, and later sent new US-made fittings, for Pat Scribner at Doak’s Machine to switch out (that happened once new leaves came in spring, for Pat to be able to run brush and re-find elusive leaks). Meanwhile that first fall, Susan supplied/applied a lot of rags, to keep oil leakage away from our prized leafy animal forage, and I added plastic bags to wrappings above the flail-chamber.

Post-delivery, Karl also located faster motors to run the flails, which (like many parts) had been unavailable, probably related to COVID19 economic disruptions. He’d actually had to compose the first flail motors from two each, as the correct shank size had been available only on even slower motors. The new flail motors offered an rpm gain of 25%+ (1,500 rpms continuous, at 14.53 gal./min., versus previous 1,150 rpms continuous at 15.5 gal./min.), allowing us fuller use of the flow control valve for much wider choice of flail speed. We could then set the feed-speed flow control valve higher, put brush through fast, and still get all the leaves off.

We were dissatisfied with gap-width between the 2 (John Deere Corn Head) gathering-belts on the top-feed system, designed to feed branches too small or floppy to make it across moving flails from front to rear tire-rollers. Karl wanted branch butts to drop from those belts to the rear tire-rollers, to exit neatly, but a helping hand (usually Susan’s) on the back side was necessary, as branches were dropped too early. Small branches had to be fed in bunches, to be grabbed at all. Shaft gap was unchangeable due to shared use with front/rear horizontal-feed tire-rollers, and excerbated by imperfect function of spring-loading within the left-side frame (iron members probably warped during welding).

[Separate shafts/motors per feed system would have required a larger iron frame over all, and upscaled hydraulic power-pack (= larger engine and probably also pump) to drive the increased number of motors. Such change would have been beyond both our budget and the hauling capacity of my Ford Ranger.]

I found that Jon Thomas, Thomas Bandsaw Mills, Brooks ME, could laser-cut larger sprockets (15 teeth, instead of 11 on the John Deere sprockets) for one side, to close the gathering-belt top-feed gap. Karl thought correctly that we could have the 2 new larger non-drive-side sprockets free-wheel (shafts continue to turn due to shared use with tire-rollers, and differently sized sprocket-pairs must turn at different speeds in order to mesh).

In June 2023, Jon supplied new sprockets, I bought bearings for them (1 pair/sprocket, sandwiching sprockets between to disable the bearings’ “self-aligning” wobble) and Pat Scribner finished all 2023 changes: larger sprockets (including lengthening the gathering belt – we bought a spare one to cut up and patch in), faster motors, and only one slow leak left (below the pump, well away from leaves – 3 people have reduced but not sealed that leak).

In June 2024, I enlisted Jon to make even larger free-wheeling sprockets, with 2 more teeth =17 now, leaving a flexible wavy-belt minimum gap of about 1/3”. Pat again installed (labor was as usual the largest expense). This change improved processing of all small stock, and was required in order to fulfill my commitment to try Multi-flora Rose (it worked!).

As delivered, leaves were supposed to drop directly into a barrel, but too many escaped. If we raised the machine up to allow a barrel hopper, everything else would be too awkward for me to reach (the engine starter cord is already awkwardly high for me, but thankfully starts on 1^st pull). For awhile, we used a tarp beneath the barrels. Andres Veles briefly worked as Intern for our 2023 MOFGA harvest, and his competent perspective inspired my additions of metal flashing and plywood beneath and adjacent to the flails, to minimize leaf loss (I can’t really call these surfaces a hopper, as they are interrupted by frame cross-members). I still must reach in intermittently with a garden tool, to free up gluts of leaves in side-portions of patchwork hopper (especially the left side, which has less room between cross-members).I also scoop leaves out of many nooks in the frame before sliding the bin out.See “Recommended Machine Changes” below.

I subsequently cut the large new bin (on-site), fitted closely on all sides. I now slide the bin out, and armload leaves into barrels, putting the barrel inside a corner of the bin to work once there’s room. 8 minutes to move leaves from bin to 2 barrels, while removing stray sticks and packing leaves more tightly (which we did either way), is less than timewrestling with a tarp.

An exhaust chimney was very necessary, and absent until I bent some 2’ flashing lengthwise, crimping a loose French seam by hand on-site. I cut a hole and rolled a short smaller pipe to meet Honda engine exhaust hole, and wired all together. I wrapped another piece of flashing loosely around the large hydraulic pump hose below the base of my chimney, as heat protection. Amazingly, this has all lasted multiple road trips! It saves us from walking directly across the stream of exhaust in order to reach the engine shut-off, plus reduces our intake of exhaust overall while running the machine.

Expense to make such a Leaf-Separator : $16,000 of our grant went to Karl for machine fabrication. Taking into account repeated or multiple purchases of fittings, motors, metal etc. (once tried, or bought for parts to add to another, they cannot be returned – a pitfall of fabrication), and reimbursements for parts Karl had on hand, includingthe Honda engine, and a farm trailer he modifiedand contributed (an after-thought, for instant set-up with barrel/bin access), Karl estimates that all funds went into materials (Karl’s receipts list items for multiple projects; accounting is not his strong-suit, and his sawmill customers were backed up, while he made our machine).

Bad deisel on the way home from machine delivery, with the big truck and trailer, caused extra stay-overs mid-trip, for repair. Karl rented a car for the 2^nd trip. His budgeted compensation covered just a standard direct mileage amount. I’d already re-allocated my own stipends to purchase 64 new barrels (which had become unavailable used), and paid out of pocket for some of the post-delivery machine changes, and more intern stipends than planned. We both wanted to make this project happen, and did so beyond budget. Karl ate the cheese, milk, apples, cider, and later blueberries I provided; I at least will enjoy continued use of the Chain-Flail Leaf-Separator and barrels (and send Karl more blueberries).

If YOU were to duplicate this machine, here’s a

Parts & Materials List, with today’s (rounded-off) prices:

$900 Engine: Honda 13 horse GX390

$480 Hydraulic Pump: 22 gpm 2-stage (7 gpm high/22 gpm low stage), 3000 psi max @ 3600 rpm max

$100 Hydraulic Tank: Re-puposed 50 gal propane tank

$400 =$200x2 Hydraulic Motors, for Flail Rotors: Dynamic BMPH32H2KP 2.20 cu.in/rev, 1831 psi cont./2393 psi int., 1500 rpm cont./1650 rpm int @ 14.53 gpm cont./15.85 gpm int.

$230 Hydraulic Motor for Feed Systems: (Dynamic BMPA-315-112-K-P-H-20220130-215 does not seem to exist – discontinued?)

$160 =$80x2 Flow Control Valves

$50 Hydraulic Shut-off Button

$290 Hydraulic Hose, 50 ft.

$500 Hydraulic Fittings (various; get US made to avoid leakage)

$600 Steel for Frame: Cut from (2”x2”x20’)x2 Angle Iron + (2”x20’)x1 Square Tubing + 20’ I-Beam

$200 Welding Supplies

$600 =$200x3 John Deere Corn Head Gathering-Belts, for top-feed system (one is to piece one side longer, to accommodate custom sprockets)

$90 =($35+$10 hub)x2 John Deere Corn Head (11-tooth) Sprockets AH103303; one must cut out the center bearinsg to replace with weldable hubs

$150 =2x 17-tooth Custom Laser-cut Sprockets (Each is 2 layers of ¼” steel welded together; tips of teeth were then bevelled to mimic John Deere tooth shape, for smooth chain entry.)

$640 =$20x 8 (approx. 2” treads, would fit 6” rims) Tires x 4 Feed- & Exit-Rollers

$320 =$20x16x3/4“ Threaded Rods, on which to mount tires above

$120 =$20x6 Shafts: Approx. 1”x3’x6 pieces Round Bar

$640 =$40x16 Bearings: 2 each/6 shafts + 1 pair each/2 free-wheeling new sprockets

$20 Flail Housing: Used 55 gal plastic barrel

$40 Exhaust Pipe (3x4”x2’ Stove-pipe ideally, but ours is aluminum flashing with hand-bent French seam)

$60 Leaf-Catchment Hopper: Some better flat flexible material than the auminum flashing that I used, to line area beneath flails & barrel housing, and funnel leaves into bin

$100 Leaf-Catchment Bin: Re-purposed sturdy plastic 250 gal. tank from road salt, custom-cut/fit (chain-saw)

$100 Nuts & Bolts, Wire, Guerilla Tape

$780 =$130x6 buckets =30 gal. Hydraulic Oil

$500 Shipping (on all parts not near you – total guess – depends where you are)

_____________

$8,070 Total Materials & Components, without Trailer & without Labor

$1,000 Used Farm Trailer with Good Tires

_____________

$9,070 Estimated Outlay without Labor (also with no mistakes nor changing out parts)

$13,500 estimated labor of a Farm Mechanic, figured as 300 to 600 hrs (so we used 450 hrs) at $30/hr.  (A machine shop would probably take half the time, but at 4 times the rate/hr, so twice as much $$.)  

Recommended Machine Changes:

Cylindrical Flail Rotors: The current square rotors bend, catch and wind up flexible branches of strong-fibered species that do not break (from worst to not-as bad: Hawthorn, Birches, Arrowwood). We suspect that such branches might be more easily pulled through, or off of, cylindrical rotors.

Larger (Pipe?) Frame: As mentioned above, the frame as built does not leave room for any continuous smooth lining nor hopper (so leaves get caught multiple places). Supports for lower flail shaft bearings should protrude inward from cross-members, so that chamber housing or lining does not include cross-members.

Alternatively or additionally, Karl suggests a Pipe Frame, which would have more strength per weight, and less leaf-catching surface.

High-Tensile Chains with Teeth for top-feeding: Karl recommends fabricating these in place of BOTH the gathering-belts and front tire-rollers(the belts were made for corn, are already cracking from our hard use; the tires on front and rear feed-rollers will soon get chewed up by the threaded rods on which they are mounted). One would feed ALL branches and trunks by handing butts up to that chain-feed, rather than having to aim stiff long pieces across moving flails.

The above new Chain Top-Feed System would be strongly Spring-Loaded, so as to grab ALL sizes of stock firmly, and drop exiting butts down to the…

Improved Exit-Rollers: Current rollers are stacks of 8 small tires, mounted on 4 vertical threaded rods per roller (no wheels, so tires squish effectively to grab irregular diameters). We are open to suggestions for a more long-lasting squishy design, but more likely the rollers can be more rigid and share Improved Spring-Loading with the feed-system as a whole.

Control Bar = Safety Shut-Off Bar, on the ingoing feed side of the machine, so one’s body can stop the machine by falling toward it (or being drawn into it – unlikely, but possible). Right now, we just have a small left-hand shut-off button.

Rear Shut-Off Button: With above Control Bar in front, the current Shut-Off Button can then be moved to the back side. Often after walking around, one sees reason to shut off the machine there. (Without chipper, brush piles up, and occasionally tries to re-enter from the exposed sides of feed-rollers).

Accessible Nuts on Tensioning Rods for tightening top-feed belts: As built, the left-hand one is especially directly above the flail-rotor top-plate, without space to engage and turn a wrench. (I succeeded very slowly at tightening by using a 1” wrench sideways on a 7/8” nut.) Karl of course built the tightening system before he placed the rotors.

All in all, these are relatively minor changes. I am quite satisfied with the effectiveness of Karl’s creation.

Leaf results: We had expected the flails would rip up leaves, so we promised textural photos per species. As you can see from these photos of 25 species, most leaves remain intact. Twig inclusion varies per speed settings and species (Cherry and Red Maple twigs are eaten by my animals; Susan’s sheep ate ALL twigs).

Mechanically separated leaves:

Leaf-silage proportion to woody refuse (woodchips):

Leaf/Woodchip Proportions: Chipping branch-wood from 5 gallons of leaves/species separated 2022

Red Maple had the lightest wood, and leaf percentages by weight and volume were similar (chips blew notably farther). This particular Red Maple had large young upright growth with leaves almost directly attached – hardly any leafy twigs, so lots of wood per leaves (it took a large armload of branches to produce 5 gallons of leaves). Red Oak had the heaviest wood and large rigid leaves, so leaf percentage by volume was much greater than by weight – a 5x8’ trailer load yielded almost two 30 gallon barrels of leaves in 1 ½ hrs., (more leaves per branch made up in time for the slowed feed speed to remove these tightly attached leaves). Just 8 small Oak branches produced our targeted 5 gallons of leaves. This Oak is a large and vibrantly leafy pollard, and oaks have high carbohydrate storage ability for re-sprouting (Furze et al/. 2019). The Birch (species undetermined) was our least attractive tree with limited new growth, and yet branches yielded higher leaf percentages than one would expect, probably due to having small short branches containing less wood. The Birch was impractical for our machine, as the very flexible branches do not break when they wrap.

In 2023 we fed the Leaf-Separator all diameters up to 3½“ (which is too large for neighbor Jim’s chipper). These large pieces have lower leaf/woodchip proportion, but are more efficient to feed into the Leaf-Separator. We observed that proportions varied more per specific stand than per species.

2022 samples had butt diameters 2” or less, and provided approximately 40% leaf-matter by weight. Means from this small data set indicate a yield of 810 lbs. leaves per ton of leafy branches, and about 8 yards of leaves for every 9 yards of bedding chips produced.

HARVEST:

Mustafa Gedo at MOFGA, farmer Susan Littlefield at her Y Knot Farm, other excellent occasional helpers (Seamus Blake, Andres Velez, Colin Yarnell, Matthew Crisp, Merissa & Justin MacDonald and their children, Gray, Connor Breen) and I (Shana Hanson) harvested tree/shrub matter and used Karl Hallen’s creation, the functional and efficient Chain-Flail Leaf-Separator machine prototype, to produce leaf-silage along field edges of 3 main sites (plus a few targetted species elsewhere).

Coppicing & Pollarding: We coppiced many shrubs, leaving short stools to re-sprout. Of those, Honeysuckle was most numerous; Arrowwood, Winterberry, and in less quantity Smooth Buckthorn (just 2 small individuals total) we also coppiced. Pagoda Dogwood and Leatherwood we pruned for higher re-growth. We felled trunks of both Quaking and Big Toothed Aspen, and of Gray Birch, to also re-sprout at ground level. When pollarded, Aspens and Birches easily abandon their trunks in favor of root sprouts (Aspens) or stump sprouts (Birches), unless significant leaf-bearing growth is retained to keep the trunk dominant. Dense spacing increased height and decreased foliage, such that we were only able to pollard a few Aspens and Birches at the sunny edge of stands. But as planned, we pollarded choice trunks of many other species (mostly Cherries, Maples, and Ashes) , one every 10 or so lineal feet, to re-sprout at heights above reach of browsing animals, but mostly reachable (for this and future harvests) with my 9 ft pole chainsaw.

Tools Used: At MOFGA we almost solely used hand-held power-tools while walking on the ground, pollarding tops with the 9 ft pole chainsaw. At Y knot Farm, we had pollarded some trunks much higher, leaving multiple sprouting locations, five years or so previous; these trees I climbed for branch removal, before using the pole chainsaw from the ground. At Faithful Venture Farm I climbed with an arborist hand saw, making cuts always in young wood less than 5” in diameter, leaving tall climbable tree structures. One green and one white ash there remain about 35 feet tall; all other trees are now 12 to 18 feet tall.

Labor Efficiency: Cutting and collecting branch matter took ¾ to 2 person-hours per 30-gallon barrel of leaf-silage, each containing about 50 lbs. = 22 lbs. Dry Matter (DM). Pollardng tall fully branched edge trees was most efficient, despite climbing and hand-tool use; as tall leafy faces hold more leaves per lineal foot of field edge (and large wood was left standing, so required no handling). Coppicing tall Quaking Aspen trees was next most efficient, but we did not count stacking or loading trunkwood. Leafiness and vigor of growth also increased efficiency; previously harvested but well-rested black cherry at Y knot Farm in full sun, and mature ashes in their prime at Faithful Venture Farm, were most bountiful; mature gray birch at both MOFGA and Y Knot Farm took longest per yield, as leaves and branches were small, and tendacy to tangle necessitated prep-time of “braiding” branches, as described below.

Machine leaf-separation took one person 10 to 15 minutes per 30-gallon barrel, varying by tree/shrub species. Ashes were fastest.

Gray Birch was difficult; branches do not break and easily wound and tangled in our machine. We successfully proccessed 15- to 20-foot top-pieces of gray birch by “braiding” branches inward , laying them across each other working from bottom to top. My goats and steer received a truckload of lower branches of that Gray Birch for fresh eating, after each Gray Birch harvest day (the animals remain our most reliable leaf-separators).

Flexible strong Arrowwood was also very difficult. Over-mature Hawthorn, with angular unbreakable branches and long thorns, tangled consistently (see center photo above) and was impossible to leaf-separate with our machine (the animals were very eager to separate such tasty leaves despite thorns).

Yields varied from about one 30-gallon barrel (each about 50 lbs as fed, or 22 lbs Dry Matter) every 10½ lineal feet (LFT) at both farm sites, to one 30-gallon barrel every 15½ LFT at the MOFGA site (low mostly due to shorter growth, and machine-tangling inseparable over-mature hawthorn which the animals instead ate fresh); 15 ft deep field-intrusion of 18 yr-old Quaking Aspen and Gray Birch there actually yielded one barrel per 4½ LFT and per 7 LFT respectively, or averaging 68 and 70 square feet/barrel.

I can fit 12 packed barrels of leaves inside this ¼ ton pick-up truck, but only filled 1 to 4 each 4-6 hr half-day; we or I also untarped and re-tarped the machine, made improvements, took notes and photos, filled sample-bags, dealt with fire-wood and spent brush, etc. (My home herd gets only fresh greenery in summer; the other half of each day I sorted photos and data on attended browse walks.)

Costs: We used about one gallon of gas (for Leaf-Separator plus power tools) per four 30-gallon barrels ($.95/barrel). Barrels were almost $70 each new, but can sometimes be obtained more affordably used, and can be re-used indefinitely. It is possible to set up less costly bulk silage containment for leaf-silage; the barrels offered us transportability, harvest-timing flexibility, and separate sampling of species.

I had huge mileage costs; my farm is mostly wooded, hence our 3 off-my-farm sites, and I can only get away from my home herd parts of each day, so made long trips for low working hours. Hopefully those with on-farm field edges will more affordably pursue such harvest.

LIVESTOCK TRIALS:

My 3 Streams Farm herd and farmers/animals at four other farms participated in winter trials to determine free-choice leaf-silage intake rates of cattle, sheep and goats, and effects on milk production (5 goats, 1 cow) and milk components (5 goats) when leaf-silage is offered in addition to their usual feedstuffs. Animals consumed leaf-silage faster than we expected, and weight of our yield was less than expected. We had filled all available barrels, but our new Leaf-Separator left leaves almost perfectly intact; we had expected leaves to be shredded and dense. The animals tend to PREFER intact leaves, so we were excited about that result. Full barrels at 60-70 lbs (versus the 110 lbs I had experienced with chipped leaf-silage) were easy to load and transport to animals. In consultation with animal research advisors Juan Alvez and Karl Hallen, I shortened and re-designed “short” trials, to supply enough leaf-silage to2 “long” trials.

Our harvest supplied for these trials 1 barrel each of undetermined tastierBirch (leaves looked like Red Birch, but bark is like Gray or White Birch), Sugar Maple, American Elm and Black Locust, and multiple 30-gallon barrels-worth of Quaking (12 barrels) and Big-Toothed (3) Aspen, Gray Birch (12), Black Cherry (12), White (6½) and Green (4) Ash, Red Oak (6), Red Maple (4½), and Honeysuckle (5),plus5 barrels-worth including 3 more tree and 8 more shrub species in buckets and sample-bags.

Favorite species: Black Cherry, White and Green Ash, American Elm, Black Locust and Red Oak were eaten with no limitation except satition. Honeysuckle was prized but left a lot of twig waste; Quaking and Big-Toothed Aspen were prized by sheep, goats and cow, but less than 50% consumed - or later (end of 2^nd trial leaf-silage period) rejected entirely - by my steer.

Birch (and possibly Aspen) antifeedant issue: Of 10 species most prevalent at our harvest sites, summer-harvested Gray Birch was least eaten, with animals at both my 3 Streams Farm and Susan Littlefield’s Y Knot Farm leaving ½ or more in our offering-periods, or about 1/3 when left overnight (and it was most laborious to harvest). This low intake may have been a harvest-timing (leaf antifeedant) issue, exacerbated by abundant inedible shattered catkins. Spring- and fall-cut Gray Birch both fresh and ensiled are often well-eaten. Availability of tastier species also magnified animal reluctance. Strangely, and as mentioned above, my animals thoroughly ate leaves off the July-harvested side-branches fresh (and could easily leave catkins hanging).

* The preferred 2022 fall-harvested Birch, mobbed by goaats in above photo, had leaves similarly shaped to those of Gray Birch, yet not so shiny.  I'm wondering if that tree at Susan's might be European White Birch, which would magnify animals' preference beyond the preferred fall harvest date.

Susan’s sheep at Y Knot Farm similarly loved that 2022 fall-cut mystery Birch, and left so much of the summer-cut Gray Birch that she gave me those 5 barrels back.

Meadowsweet Farm beef cattle and sheep accepted that less palatable summer-cut Gray Birch when fed in lower quantity per head. Farmer Eliot Van Peski used up leaf-silage from 2 barrels by sprinkling it on top of the Icelandic sheeps’ hay each day, for almost 2 weeks; he said thay seemed to value it.  Certain Angus beef cattle ate well from another barrel, eating 19 lbs of 39.5 in 1 hour (this is same proportion my and Susan's anmals finished; did the catkins get too concentrated, at that point?).  In 4 days they completely finished the left-overs (my goats and Susan's sheep never finished theirs). It probably helped that neither sheep nor cattle were right then receiving tastier tree/shrub forages.

I saved 3 barrels of that 2023 MOFGA July-harvested Gray Birch over and fed in winter 2024-’25, to see if longer fermentation had caused preference; this winter they have little leaf-silage (but lots of fresh Red Maple tops) so were appreciative, but still ate it much slower than they did that 9/30/22-harvested (and possibly species-different) Birch.

Based upon experience with my herd, White Birch has date-related palatability variance, similar to but much less pronounced than Gray Birch; Yellow Birch is eaten all season. Within past SARE FNE18-897 I found Quaking Aspen to be preferred similarly (but even less strongly) in spring and fall.

Also my animals tend to prefer foliage of mature trees to that from young stems, which seem appropriately more defended; Jogia et al., 1989 state that hares (also) tend to prefer mature twigs to new sprouts in general (citing Klein 1977 & Sinclair & Smith 1984). This is consistent with the “‘plant defense guild’” hypothesis (discussed by Jogi et al., pp. 191-192) of evolutionary interaction. We observe pollarded (young, vigorous) growth from mature tree structures (harvested after an approprite 3 to 8 year rest-period) to be most choice. [An aside: Interestingly, pollarded growth on Holly in England is nearly thornless, while lower over-browsed branches become fiercely thorny.]

Gray Birch can become especially abundant when blueberry fields are not well weeded. Within related SARE FNE24-083, I’ve added full-season 2025 plans for palatability testing of young coppiced Gray Birch of various ages from my blueberry field at multiple harvest dates, to better understand seasonal and age-related browse dynamics of this forage resource. As yet I have found no literature regarding what chemical defenses might cause animals’ mid-summer Gray Birch refusals.

Maple Species Toxins:

In my home trial, animals ate small amounts of Red Maple eventually, if I left it overnight. In the Y Knot Farm trial, Susan’s sheep ate our only barrel of Sugar Maple, plus 2½ barrels of Red maple, but she fed each barrel over multiple days, as did I. Our small herds (11 and 16 head respectively) ate whole barrels of preferred species in less than 2 hours.

MU Metabolomics Center, in our tests funded by additional SARE FNE24-083, quantifiedGallic and Elagic Acid levels, and relatively quantified 4 non-proteinogenic Amino Acid toxins, for us to better understand what might be limiting consumption of Red and Sugar Maple leaves, plus “Ash-Leafed Maple” AKA Box Elder* (which caused death of a goat at a farm near me) and Staghorn Sumac* (not a maple, but said to have a similar Gallic Acid issue). We included Norway Maple* for comparison (*these last 3 species were not present in quantity at our 3 harvest-sites, so were not fed in trials, but do appear in our nutritional spreadseets).

 ***Above Metabolomics spreadsheet is re-printed from SARE FNE24-083 January 2025 Project Update. See that report for more detailed discussion of these specific toxins.***

Norway Maple was confirmed to be non-toxic, and animal intake in past has seemed non-date-sensitive and unlimited (but we did not have enough in our 3 sites to fill a barrel). Norway Maple bark is less palatable than Red Maple bark (so young trees are more likely to grow up past goats), and twigs are easy to snap for hand-harvest; these attributes and nutritional appeal all make sense, in light of long European history of domestic ruminants and tree forage use.

Toxins above (Gallic & Ellagic Acids, and Hypoglycines A & B with their respective “homologues” MCPrG and y-glutamyl MCPrG) are said to increase throughout the growing season. We sampled rather early (June 24-25); my goats and steer accepted Box Elder at that time. On July 20^th, they refused Box Elder from a different tree (my animals have many choices on long free browse-walks daily; the goat that died was in a stall with limited choices). Note variance between same-species samples above (3 samples/species), even with same harvest-date.

Hydrogen-Cyanide (HCN or Prussic Acid) in Cherry leaves: We fed many barrels of Black Cherry, in both Susan’s (Y Knot Farm) and my (3 Streams Farm) trials. Cherry was a top choice for sheep, goats and steer. Iowa State Veterinary Diagnostic Laboratory (ISVDL) analysed 4 fresh/ensiled sample-pairs for us, 3 Black Cherry, 1 Pin Cherry. Spring-harvested and wilted Cherry leaves are said to have highest levels; our earliest-harvested (June 27^th) fresh Black Cherry indeed had highest Cyanide level, but barely crossed into the “should not be the only feed” category. Cyanide dissipated during ensilement, in all pairs.

Dr. Radke at ISVDL regularly tests Bromegrass for Cyanide, and says 60 days ensilement of that is a commonly accepted safety measure. We will be testing Black Cherry at more harvest-dates and stages of wilt, this coming growing season for SARE FNE24-083.

Free-choice intake rates:  

The Jersey steer at my 3 Streams Farm in 1x 2 hour offering period/day (during 3 leaf-silage periods of 4 transition days + 7 measured leaf-silage days = 11 days each, within our 66-day trial) ate 2 to 3 species/day of leaf-silage as 1/3 the Dry Matter (DM) of his grass-/browse-fed (no concentrates) diet, with late-cut 1st-crop hay offered 24 hrs/day. He may have chosen more leaf-silage; I slightly limited offerings to him, to assure sufficient supply for the goat milking trial. 10 grass/browse-fed Saanen dairy goats here in same once/day 2 hr. offering periods chose leaf-silage as over half (55.5%) of their dietary DM. I suspect all animals would have consumed that much again, if I had been able to provide offerings 2x 2 hrs/day.

3 Streams Trial Chart, Data Entry Copy

3 Streams Trial, Leaf-Silage & Hay Summary Chart

At Faithful Venture Farm, I offered a bit less than one 30-gallon barrel per each of 4 consecutive cold January days (less because I took samples), to 14 mostly Holstein heifers (+ or -; started with 16, but on various days one escaped and got moved). Each day I fed out just 1 leaf-silage species: I timed their consumption, and noted individuals’ participation. One submissive young heifer tried but never got any; all others ate shoulder-to-shoulder along 7 plastic utility sleds used as troughs.

See my mini-presentation of the Faithful Venture Farm trial below:

Faithful Venture Farm trial, Jan 2024

We experienced eager animal intakes, yet I must mention caveats:

Quality of 2023 hay that animals in our trials had as their staple forage was poor; even Meadowsweet Farm's purchased baleage that year was poor.  Also, my animals and Susan's are very familiar with tree forages.   

Outside of our Project's trials, Nathan Zimmerman, Palmer Hill Farm,  offered a barrel of hybrid willow that I harvested in October 2024 to his 3 groups of grass-fed cattle (milking Jerseys, Jersey heifers, and Angus x Jersey beef cattle).  They were completely satisfied with his excellent 2024 spring-cut grass silage, kept in a tll silo.  They were not interested in my late-cut willow, that MY animals all devoured. This may have been due to habit, but more likely they simply had high-quality feed.  Nathan and I will be offering them some willow from the same stand but cut in June, NEXT season (I'm curious).  They do have a browse-line on the trees, there. 

]Hay cost saved: With just that one 2hr offering/day, the steer and 10 goats together consumed 44% (24.13 lbs DM or 30.16 lbs/day as fed) less hay on average during leaf-silage measurement periods than during 2nd-cut hay periods. The difference was most of a rare 2^nd-cut bale/day @$10/bale at harvest(I was lucky to get any, in wet 2023), minus a bit more total 1^st-cut hay eaten, when leaf-silage was offered (that was the steer, who filled up faster on 2^nd-cut hay than on leaf-silage; the goats actually ate less 1^st-cut hay in leaf-silage periods). I saved $8.70*/day on hay x 30 days of leaf-silage in our trial = $261 (minus $25 gas for the Leaf-Separator = $236). [*Priced at harvest; prices doubled that winter.]

Susan Littlefield at Y Knot Farm had hoped to do a summer milking trial with our leaf-silage, but the State did not arrive for re-licensing. Then she got too busy. With short hay crops State-wide including her own, her butcher had no openings until February. She bought $50/100 lb bale NY timothy and alfalfa hay, to hold her lambs ($12/45 lb bale local hay was too poor quality for growing stock). Then Susan re-licensed and refurbished her store and hot-dog stand, and perfecteda Gyro Sandwich recipe to recoup costs of her lamb meat; I was grateful for her to run the trial at all.

Susan saved 1x 45 lb bale/day of hay x $12 for 30+ days = $360+ (she missed recording some of the days – we think it was 33 days), by rationing out 15½ barrels of leaf-silage to her 15 Freisan x Dorset ewes. She judged them to be too fat, and said if she followed my original plan to determine how much they would eat, the sheep would simply eat everything offered. “There’s a thing called over-eating disease,” she told me. I didn’t argue; we moved on to a taste-test of the Gyro samdwich.

Susan didn’t identify each barrel from the typed list I provided(her barn is quite dark), so we may have misidentified some species, as noted. The sheep ate almost all twigs of all species except July-harvested Gray Birch, which Susan stopped offering early on, due to her ewes’ lack of interest. See spreadsheet below, which I created from her paper notes.

Y Knot Farm Sheep Trial, March 12 thru April 21, 2024

Our hay dollar savings did not compensate for tree/shrub harvest labor, nor for our purchase of reusable barrels (as previously mentioned, silage bunkers would be more economical). Controlled field edges, firewood, and reliability of deep-rooted multi-year growth, harvestable in short flexible time-frames regardless of moisture, are incentives. One farmer said that he’s cutting brush all the time, anyway (NOFA MA webinar, May 2, 2024).

Animals indicate that nutrition, health and happiness are also incentives to feed leaf-silage; details await more research (see “Condensed Tannins” and “Fat” under “Leaf-Silage Nutrition” below).

Winter milk yields of 5 grass/browse-fed (no concentrates) Saanen goats at my 3 Streams Farm rose generally over the 66-day trial, correlating loosely with day-length. Over all, longer day lengths should correlate with rising DM intake throughout (and hence more milk). Exceptions were the 2^nd (BL) Leaf-Silage period, in which everyone ate less DM than in any other period (possibly weather-related), and 3^rd (CL & CH) rotation, when goats ate more DM in the Leaf-Silage period than in the Hay-Only period. [Perhaps this change was because I had taken more (very choice) barrels from Faithful Venture Farm at the end, and so added leaf-silage more readily, when they slowed down picking through it. I was finally sure that we had enough to finish the milk trial.]

Since each of 3 (11 days each = 4 transition + 7 measured days) leaf-silage periods preceded each 2^nd-cut hay period, I am choosing to compare milk yields during leaf-silage periods B & C to 2^nd-cut hay periods A, B & C, in order to equalize day-length.

When thus discounting the 1^st leaf-silage period to control for day-length, my does gave 6% less milk in leaf-silage periods than in 2nd-cut hay periods. But they also ate 4.9% less 1st-cut hay (offered 24 hrs/day) in leaf-silage than 2nd-cut hay periods. Average leaf-silage and 2^nd-cut hay Dry Matter (DM) intakes (during the 2-hour daily offering) were equivalent; the 4.9% difference in 1^st-cut hay intake caused 3% less total DM eaten in leaf-silage periods than in 2nd-cut hay periods. Goats’ milk yield per lb DM intake in leaf-silage periods was 98% of that in 2^nd-cut hay periods (0.0088 lbs milk difference/lb DM eaten = 14 oz milk difference/100 lbs DM eaten).

3 Streams Trial,, Milk & Butter-Fat Yields per goat per day, plus Summary Charts

3 Streams Trial, leaf-silage, hay & milk, summary computations

Our leaf-silage had much higher energy/lower fiber than hay, leading to less dietary volume needed, but had less protein (and less protein availability) than the hay (see “Leaf-Silage Nutrition” below). In my goats’ forage-only (no concentrates) diet, protein level limits winter milk production, but does not seem to affect Butterfat yield.

Goat Butterfat lbs yield differed insignificantly between leaf-silage and 2^nd-cut hay periods; significantly higher Bfats percentages in leaf-silage periods compensated for slightly lower milk yields.

Goats’ milk Component Analyses: Outside of original project commitments but out of curiosity, I sent* 1 milk sample/goat from each of the 6 measured leaf-silage or 2^nd-cut hay periods to Dairy One for analysis. This low number of samples, especially with variability of individual goats, does not offer conclusive data, but gives an indication of possible leaf-silage (versus 2^nd-cut hay) effects.

(*Glendon Mehuren at Faithful Venture Farm asked the Organic Valley driver to transport my box of frozen samples with their own, directly to Dairy One, as there is no same-day shipping from our area.)

Butterfat (Bfat) percentage averaged 16% higher in leaf-silage periods than 2^nd -cut hay periods (x lower milk yield in leaf-silage periods = same Bfat lbs yield across periods, as mentioned above). Lactose was slightly lower in leaf-silage periods. Total milk-solids in leaf-silage and hay periods were equivalent. Somatic Cell Counts (SSC) varied broadly, with higher leaf-period average; these are of concern for dairy licensing. Other milk components did not differ significantly between leaf-silage or 2^nd-cut hay periods.

3 Streams Trial Milk Data, Dairy One

Lipid Identification in goats' milk, with/without leaves, and comparison to cow Betsy (below)'s milk without leaves :  Yulica Santos Ortega, previously at Maine Health Institute for Research, now at U VA, showed up with her family to harvest blueberries with us, and later offered to do an immense amount of pro-bono laboratory work.  the whole family plus child-friends came up for milk sampling (plus music and steer riding), then Yulica identified hundreds of lipids in our goats’ milk with  leaf-silage and with 2^nd-cut hay.  She graphed those with greatest difference between tretments (leaf-silage versus 2^nd-cut hay):

Yulica Saantos Ortega's graphs comparing levels of 10 lipids in goats' milk with leaf-silage vs 2nd-cut hay

Here are ALL the lipids she identified:

Yulica Santos Ortega's Lipid results from 2 samples per each goat, 1 with leaf-silage, 1 with 2nd-cut hay

Yulica also compared one goat's samples, with leaf-silage and with 2^nd-cut hay, to the Tilden Pond cow's milk post-trial (without leaf-silage):

Yulica Santos Ortega's comparison of 1 goat's milk with leaf-silage or 2nd-cut hay, to 1 cow's milk with no leaf-silage

(*Some of these results are over my head.  I await further explanation from Yulica, or other knowledgeable scientist.)

Milk yield of the mini-Jersey cow at Tilden Pond Farm was higher by 8% or 1.625 lbs (about 3½ cups) per day cross 10 days, when John Morse fed cow Betsy about 7 clean* lbs/day of Honeysuckle (with some Black Cherry also on 1 day) or Quaking Aspen leaf-silage, in addition to her usual diet of late 1st-cut hay plus 8 lbs raw grain/day. Her average daily yields were 21.81 lbs with mostly Honeysuckle (some Cherry), 21.74 lbs with Quaking Aspen, and 20.15 lbs across the following* 10 dayswithout leaf-silage. (These were decent mini-cow yields; Betsy was milking on just 2 quarters). Betsy’shigher milk yield with leaf-silage was despite of very loose stool (from the dietary change?), and in spite of shorter days of the Winter Solstice during the leaf-silage period.

(*No twigs; John spent much untimed prep-time, as the cow convinced him to remove them all.)

(**John took a break on 1 especially rainy day when he had too much else to do.)

Tilden Pond Farm Cow Trial, leaf-silage & milk amounts

LEAF-SILAGE NUTRITION:

Leaf nutritional analyses indicate low Soluble Protein (<2% of DM on averge, versus 14.5 and 33% in our 1^st- and 2^nd-cut hay samples respectively), and low Rumen-Degradable Protein (2.5% DM averagae, versus 5.8 and 13% in 1^st- and 2^nd-cut hay samples), with exception of our one Autumn Olive sample cut July 5^th, 2023 (). Hauge, Garmo and Austad (in Austad & Hauge 2014) reported that young leaves in general are as high in protein as grasses (25-30% CP), but that protein decreases as the season progresses. They cite Garcia Plazaola et al. (2003), who found that shaded leaves are often higher in protein than those in full sun. We harvested no north-facing tree-lines, nor did we cut spring leaves. One can surmise that, as with grass, higher protein is a trade-off with lower yield of an early harvest, and also may tax plant/tree energy reserves. Cutting in shade may also diminish yield, or jeopardize regrowth. Many farmers feed early-cut grass silage, which provides a surplus of protein to their cattle, so can supplement with energy-rich leaf-siilage cut anytime, regardless of leaf protein content.

There was about ½ as much %DM of Digestible Fiber in our leaf-silages as in our grass hay, and more than twice the %DM of Non-Fiber Carbohydrates as the grass hay (39% versus 14 and 16%), slightly higher % DM of Water Soluble Carbohydrates (9.9% versus 9.7 and 7.5%), much higher Crude Fat Ether Extract (6% DM; grass forages average 2- 3%; see further discussion under “More about Fats” below), and almost 2x the Calcium as our hay samples, plus other high mineral levels. Total Dietary Nutrition scores across our 10 most plentiful leaf-silage species averaged 62.3; our 1^st- and 2^nd-cut hay samples scored 55 and 57 respectively.

Net Energy projections were 22% to 36% higher for leaf-silage than for our hay samples. The 36% was NE Growth, which may be inaccurate due to use of high Crude Protein figures without taking into account low protein availability.

Dairy One NUTRITIONAL DATA from ALL 2022-'24 LEAF-SAMPLES analysed

Dairy One data comparing PAIRED FRESH & ENSILED leaf-samples       (I appologize for imperfect layout on this one, as PDF.)  

Most information in above spreadsheets was funded through our related add-on SARE Farmer Grant FNE24-083. SARE is funded by the National Institute of Food & Agriculture.)

***For more in-depth discussion, and additional spreadsheets comparing harvest dates of our leaf-silage, and fermentation acids, see the SARE FNE24-083 January 2025 Project Update.***

More about Fats: Fat EE level in our leaf-silages averaged 6% DM = 150% of that in average Dairy One grass-silage.

Hauge, Garmo & Austad (pp 74-75 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 covering. This layer serves as a defense against parasites, and reduces water loss... The outermost layer (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 acc to Google] 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.

Our ensiled leaves consistently had more fat than fresh leaves (averaging 111% of fresh level). Perhaps the rise is in wax coatings, as the leaves suffocate? Or do the wax coatings break down to become digestible fats?

As mentioned above, Yulica Santos Ortega (previously at Maine Health Institute for Research, but now at U VA) identified hundreds of fats in our goats’ milk with and without leaf-silage.

Now Andrea Clemensen, Research Biologist at Northern Great Plains Research Laboratory, S.W. Mandan, North Dakota, is talking to those at their partnering USDA laboratories to see if they can identify lipids, fats or both, in leaf-samples fed in my 3 Streams Farm trial on dates of Yulica’s milk sampling.

***See my summary of Yulica’s findings, & Yulica’s graphs, in the NOFA MA Presentation Slides with Text Added, under “Informational Products” below.***

Condensed Tannins (CT): Wayne Zeller, US Dairy Forage Research Center in Madison, WI, screened tiny fresh (and one ensiled: Quaking Aspen) leaf-samples from our 2023 harvest plus additional 2024 samples. Now he has started isolating and identifying tannins from larger samples of 20 species rating “5” or higher for CT content, on a scale of 1 to 10. (This collaboration goes beyond our SARE committed research plans; thank you, Wayne.)

The above list does not include 2024 samples just screened this January, as specific ratings were not assigned in that 2^nd screening, though all but Staghorn Sumac clearly scored >5:  Below is a look at tht Jaunary 2025 screening:

***For slides of Wayne’s 1^st CT Screening, see SARE FNE24-083 January 2025 Project Update.***

Citations:

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.

Jogia, Madhu K., Sinclair, A.R.E., & Andersen, Raymond J. (1989). An Antifeedant in Balsam Poplar Inhibits Borowsing by Snowshoe Hares. Oecologia 79: pp 189-192