Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

fe United States

p> Department of
Y11 (Sy) ‘i

Agriculture
4 o> Forest Service

Pacific Northwest
Forest and Range
Experiment Station

Research Note
PNW-406
July 1983

Abstract

Introduction

hepeeinesintneenineeentettenteiiennainiiiaiemntsteaaminiameenteenl
PSW FOREST AND RANGE

EXPERIMENT STATION

MAR 2 5 1986

STATION LIBRARY COPY

Tree Shaking Machine Aids Cone
Collection in a Douglas-Fir Seed
Orchard

Donald L. Copes and William K. Randall

A boom-type tree shaker was used in a Douglas-fir seed orchard to remove cones from
7- to 9-meter tall grafted Douglas-fir trees. An average of 55 percent of the cones were
removed by shaking, while damage inflicted to the upper crown was confined primarily to
branch and leader breakage in the top three internodes. Damage to the lower bole, where
the shaker head attached to the tree, occurred only once. Shaking a significantly greater
proportion of cones from 7- to 9-m tall Douglas-fir trees without drastically increasing top
damage is not likely. Much of the cone crop on small trees is found on the difficult-to-shake
lower half of the trees; energy levels sufficient to remove those cones would cause severe
breakage in the upper crown. The shaking procedures outlined in this report were
relatively gentle, yet they resulted in the harvest of over half the cone crop in a rapid and
efficient manner. It should be possible to machine-harvest cones from approximately one
tree per minute when a tree shaker and mechanical catch frame are used in seed
orchards that have cones on a majority of the trees. Handpicking the cones remaining on
the trees after shaking will still be necessary, but that job should be much easier than
picking cones from unshaken trees; more than half the cones will have already been
removed, including most of the difficult-to-reach cones from high in the trees. A
combination of machine harvesting and hand picking of cones should considerably
shorten the time required to complete the cone harvest and the cost per bushel should be
reduced because of increased efficiency.

Keywords: Cone collection, field equipment, Douglas-fir, Pseudotsuga menziesii.

It is difficult to complete cone harvest in Douglas-fir seed orchards prior to natural
seedfall. Orchards today contain several thousand trees of various sizes, and picking
cones by hand becomes more difficult as tree size increases. Physically climbing and
using ladders or lift devices are slow and inefficient methods of harvest. Such methods
also involve an element of risk to cone pickers. Slow production, high labor costs, and the
purchase or rental of expensive equipment combine to make low-cost cone collection
difficult. Surprisingly, hand picking is still the only method used in today’s Douglas-fir
orchards.

DONALD L. COPES is a plant geneticist at the Forestry Sciences
Laboratory, Pacific Northwest Forest and Range Experiment Station,
3200 Jefferson Way, Corvallis, OR 97331. WILLIAM K. RANDALL is
an area geneticist at the Suislaw National Forest, Box 1148, Corvallis,
OR 97330.

Methods

Cone harvest would probably be safer and less expensive if existing commercial nut and
fruit harvesting machines could be used to harvest Douglas-fir cones. Trials with other
conifers indicate that many species do not readily drop their cones when shaken because
no abcission zones are formed in the woody stems that attach the cones to the
branches.’/ Tree shaking machines have been used successfully in the southern United
States to harvest cones in slash pine orchards. Shakers have not been used in
Douglas-fir orchards of the West, but trials were conducted in Idaho and in southwest
Oregon on large forest trees.2/ In the tests, Douglas-fir trees were shaken with
encouraging results for wildland collections, but the lower bole and the upper crown of the
shaken trees were often severely damaged. Extensive top damage resulted from the
extremely vigorous shaking, which was needed to effectively remove tightly held cones.
Such severe crown damage is not acceptable in seed orchard trees because they must
repeatedly be harvested.

Orchard managers are reluctant about using shakers until it can be demonstrated that
such mechanized harvesting methods are nondestructive. We decided to conduct a
shaker trial in a Douglas-fir seed orchard during the 1982 cone harvest to determine if
mature cones could be harvested from 7- to 9-m tall orchard trees without causing
excessive tree damage. Specific evaluations were percentage of cone removal, amount
of top breakage, the relationship of cone removal with amount of top breakage, and effect
on cone removal and top breakage of balanced versus unbalanced weights in the shaker
head.

Tree shaking was carried out at the Beaver Creek seed orchard of the USDA Forest
Service near Corvallis, Oregon, during the last week of August 1982. Ninety-two trees
were shaken. Evaluation of percent cone removal and top damage is reported on 10- to
14-year-old grafted Douglas-fir trees. Trees ranged in height from 7- to 9-m and had stem
diameters that ranged from 9 to 25 cm as measured 30 cm above the graft unions. The
orchard consisted of ten 2-ha blocks, but trees in only four of the blocks were shaken.
Each block contained approximately 300 trees, or 10 to 15 ramets of 25 clones. Some
trees produced good cone crops, but the proportion of producing trees were small. The
same clones were not present in more than one block, so comparisons of treatments
between blocks are limited.

1/ McLemore, B. F. Chemicals fail to induce abcission of loblolly and
slash pines. Res. Note SO-155. New Orleans, LA: U.S. Department of
Agriculture, Forest Service, Southern Forest Experiment Station;
1973.3 p.

2/ Anon. An evaluation of two tree shakers for harvesting cones.
Equipment development and test report. Missoula, MT: U.S.
Department of Agriculture, Forest Service, Missoula Equipment
Development Center; 1972. 27 p.

sala Picea:
he CH a

fea A to inniingee)
bliin sie

i” He “4 B

Results

The machine used in the trial was a Kilby Company, four-wheel drive, boom-type shaker
with a scissors clamp shaker head. 3/ Maximum opening width of the head was 53 cm
(21 in.). The head contained two 41-cm (16-in.) energy wheels adjusted to produce a
six-sided, star-shaped shaking pattern. Blocks 2, 3, and 4 were shaken with balanced
weights attached to each energy wheel (30 kg), whereas block 5 was shaken with
unbalanced weights: 26 kg weights on the upper energy wheel and 30 kg on the lower
wheel. Clamping pressure between the shaker head and the tree was measured at 750
psi. The amount of shake energy or force applied to a tree was a function of the weights
used on the energy wheels of the shaker head and the number of times the shaking
pattern repeated each second. Any combination of heavier weights and more shake
cycles per second resulted in increased shake force or energy.

One person on foot helped the machine operator determine where to attach the shaker
head on each tree, and how much power (cycles per second) to use and for how long. The
duration and frequency of the shake were varied according to the physical characteristics
(angle and length of branch, and size of tree), number of cones, and ease with which
the cones were shaken from each tree. Proper shaking methods were perfected as work
progressed the day of shaking. Only limited prior information existed on proper
techniques for small conifers. The shaking procedures we tested were limited to those
which were not likely to cause extensive crown damage.

Efficiency of cone removal was determined both from visual estimates and from actual
measurements of cone weight. One person visually estimated the percentage of cones
shaken from each tree. To check the accuracy of the visual estimates, weight of cones
shaken and weight of cones remaining on each tree after shaking were determined by
hand picking the cones remaining on 28 of the 92 trees. Weight data for all 92 trees could
not be made because of limited staff and time. The cones were mature and picking had to
proceed rapidly. The amount of top breakage was determined by measuring tree height
before and after shaking. The number of growth internodes broken off was also recorded.

Data on percent cone removal were subjected to arc sine transformation before
regression or analysis of variance. Regression analysis of data for estimated and actual
percentages from 28 trees in blocks 4 and 5 was made so that the estimated values for all
92 trees would be more accurate. Corrected values were calculated from the formula

y = a+ bx. Inthe formula, values for y, a, b, and x stand for corrected percentage,
intercept, slope, and original estimated percentage, respectively. These estimates were
then used for analysis of variance tests of relationship of top breakage by percent of cone
removal and top breakage by orchard blocks. A polynomial comparison was used to
evaluate linear effects. It should be noted that statistical tests based on adjusted values
that do not contain sampling error are not exact, but in this case are thought to illustrate
the differences fairly accurately.

Average shake time was 1.5 minutes; individual trees were shaken for as short a period
as 25 seconds and for as long as 3 minutes. The average time for the whole operation of
clamping, shaking, and moving to a clamping position on the next tree was 3 minutes.

3/ The use of company or brand names is for the convenience of the
reader and does not constitute endorsement of the product by the
U.S. Department of Agriculture.

fers “ jt a on a ott bts
siglo eh} Hi ie Pea Fe pe a
snes sit vse nO Ca

esgshneeiin aii
pale fear f ‘avai : ‘a 4 Ap on,
priitoivn My iw iid TPA

“nites * well bavi
bead " at Ai ; ibd

“we a wane j
sai any nORS aL a

gn as iota Th
zt. = fom 3 Tig hy, ba .

Y 7

"pessoal council rm Hie ce
hades Rite Eat
Ne i Suan
eee ee
8 inaoreg gees

aa on ion

te he Mea:
ie i nan Hy Uae Sh i je th Bay

aR wale emuiy wit i CO a
VARRENT agi ‘Me i i hae vem pA

AGS
tng “i G5 a A is, Y i "
(ite. eke Ne ab png lars eile Pari sail heat
Le EER YH are aphiwit tf
yO vont au an

An average of 55 percent of the cones were shaken from each tree (fig. 1). considerable
tree-to-tree variation existed; individual trees varied from almost O- to 95-percent cone
removal. Percentages of removal among all orchard blocks were similar.

80 !

70
60
50

40

30

Percent of cones removed

20

1 2 Sivas 4
(29) (37) (24) (2)

Number of whorls broken

(Number of trees shaken)

‘Figure1.—Percentage of cones
removed by shaking.

Estimates of cone removal in blocks 4 and 5 were found to closely correlate with actual
percentages derived from weighing cones (r = 0.80). The uncorrected visual estimates
averaged 5 percent higher (table 1). Values presented in figure 1 are for all 92 trees in
blocks 2, 3, 4, and 5 have been corrected by the regression formula derived from the
relationship between estimated and actual data in table 1. The regression formula

y = a + bx provided the corrections, where y = the corrected percentage value, a =
intercept value of 9.127, b = slope 0.771, and x = the original estimated percentage
value.

When the percentage of cone removal is compared to amount of top damage, the law of
diminishing returns is evident (fig. 1). As more of the top is broken by increased shaker
energy (more cycles per second), lesser percentage of increase in cone removal

(8, 4, and 2 percent) is obtained with each increase in breakage (whorls 2, 3, and 4). Even
though the correlation (r = 0.27) between cone removal and degree of top breakage was
significant at the 0.05-level, there were large differences in cone removal between trees
in the same damage class. Some trees with low percentages of cone removal occurred
independently in all four damage classes.

Table 1—Estimated and actual percentages of cones removed by shaking 28 trees ina
Douglas-fir seed orchard by number of whorls broken

Number of growth whorls broken —_ Total or

Orchard block and item ,; ele oi ee cea: 3 4 average
Block 4:
Number of trees shaken 2 9 3 0 14
Estimated percentage of cones removed 1/ 35 60 82 0 61
Actual percentage of cones removed 2/ 31 53 TAs) 0 56
Block 5:
Number of trees shaken 2 6 6 0 14
Estimated percentage of cones removed !/ 45 52 70 0 59
Actual percentage of cones removed 2/ 45 47 69 0 56

Total or average for both blocks:

Number of trees shaken 4 15 9 0 28
Estimated percentage of cones

removed !/ 40 57 74 0 60
Actual percentage of cones removed 2/ 38 51 71 0 55

1/ Uncorrected visual estimates.
2/ Derived from weighing the cones.

Some damage to the upper crown occurred at any shake frequency powerful enough to
remove cones. Minimal damage resulted when the number of shake cycles per second
was kept low. By either halting, reducing, or sustaining the energy applied to the shaker
head when leader breakage first occurred, it was possible to confine breakage within the
top three whorls in all but 2 of the 92 trees (fig. 2). Damage thought to be unacceptably
severe for seed orchard trees (breaking 4 years or more of growth from the top) occurred
in only two trees when they were intentionally shaken excessively hard to see if additional
cones could be removed. An average of 2 years of growth was broken from the top of the
trees (table 2). Approximately one-third of the tops broke in each of the first, second, or
third internodes.

Significantly more severe physical damage occurred to the trees in blocks 4 and 5 where
larger segments of top were broken than in blocks 2 and 3 (table 2). Unbalanced weights
in the shaker head were used for trees in block 5 but the same balanced weights used in
blocks 2 and 3 were also used in block 4. Little difference in damage occurred between
trees of blocks 4 and 5, so the unbalanced weights were not responsible for increased
breakage. There was no apparent advantage to using unequal weights.

Other observations were made when a number of shaking techniques were tried in an
effort to perfect the shaking process. Results from those observations cannot be
analyzed statistically but are presented simply as observations to aid others who may
wish to use shakers for cone collection in Douglas-fir seed orchards.

npaa to .

£ 5 » 4 ry i
Pan iam) gy ‘
! Yoiaky
’ ry
pina itn 1 isetane bein.
; si MAD lit a |

an #ey. a ‘abe

ee ae

Figure 2.—Top breakage in seed
orchard trees 7-9 meters tall:

A, acceptable amount (confined
within the top three whorls); B,
unacceptable amount (4 whorls or
more).

Table 2—Characteristics of and damage to trees shaken for cone removal in a Douglas-fir

seed orchard by orchard block number

Characteristic of and
damage to trees

Number of trees shaken
Average diameter (cm)
Average height (m)

Before shaking
After shaking

Percentage growth whorls broken:

1 whorl

2 whorls
3 whorls
4 whorls

Average amount of top breakage
(number of whorls)

19.3

1.6

Orchard block number

3 A
28 21
16.8 19.1
9:2 9.3
7.3 7.4
46 14
32 52
18 29
a 5
1.8 22

5

23

24:0

2.3

Total or
average

92

19.3

2.0

Observations on methods to reduce top damage indicated that the least damage
occurred when shaking was started with the machine at idle speed for 5-10 seconds, and
then the power was gradually increased until no more top breakage could be tolerated.
Shaking was sustained at the maximum acceptable power level until no more cones were
removed. Low energy at the start of the shake removed cones from the upper whorls
where, otherwise, a combination of heavy cone weight and strong shake force could
break the top unacceptably low in the tree.

Most trees were shaken twice: the machine was repositioned 90 degrees after the first
shake and reshaken for a few seconds from the new position. The change in direction
yielded extra cones which would not have fallen if the trees were shaken from only one
side. A slight burst of power at the end of each shake was often beneficial in removing
additional cones, but it was difficult to control crown damage when power bursts were
applied; crown damage was easier to control when power was gradually increased to the
highest acceptable level and then sustained until no more cones were removed.

Shaking action in the upper crown was dampened when the shaker head was attached
less than 0.5 m above the ground. Such low attachment also caused an undesirable
amount of root disturbance or movement. The best shaking action and most time-efficient
procedure was to clamp the shaker head 1 to 2 m above the ground; that height at the
Beaver Creek seed orchard was at or just above the graft union. Attaching the head

3 to 5 m above the ground did not aid in removing any more cones from the 7- to 9-m tall
trees than when the trees were shaken lower on the bole. Shaking high in the crown was
also a more time-consuming procedure because of the necessity of positioning the head
between the whorls of branches.

Only one tree sustained trunk damage where the shaker head attached to the lower bole.
In that case the bark was removed from two-thirds of the tree’s circumference because
the rubber pads on the shaker head were not adequately greased. Extremely low
clamping pressure of 750 psi was used because the shaker head often attached to the
tree where the bark was alive and green. The lower pressure appeared to eliminate
pressure damage, yet it was high enough to allow proper shaker head operation on the
small trees. No other external bole damage was seen.

Some trees had physical characteristics that either hindered or promoted removal of
cones by shaking. Trees with long, pendant branches and dense foliage did not yield a
high percentage of cones. Long branches in the lower crown which had a heavy mass of
cones at their tips were nearly impossible to effectively shake. The easiest trees to shake
were those with short, stiff horizontal branches. Ramets of the same clone tended to have
more similar percentages of cone removal than did comparisons made between ramets
of different clones. It was also possible to shake some incompatible trees more vigorously
than was possible with healthy trees and still not break the crowns below the top three
internodes. No problems occurred with breakage in incompatible graft unions.

id

' i Ae 1 ‘ ot aie seh en ee
1s Phi Pe ae ae vay @ rey 3

seh anh
Vora rant a
hun a Qe

A 4 ! ‘ '
aS ar uf
) i. P !
; yyy iS }
Meno * j Ua
‘ t iad 4 4 i
itd es:
yf
Beirne ial Wi cwy é
b aunt if 4 Weha
Cy ee u. 2 Ll j
z
: =:
1, 7 4
’ ‘ )
Nf + } at
ry fa
i M 1
i i kg
+ P se
Fy 4 7 nit i ‘ .
es i 3 a my fe)
i” : a
‘
-
bs j { {
{ Mig t } \ i { } /
es ’ ‘ 7 en At if 4 tev) Oh. date Jiu) le i
ry ae Bie ty Seon OAbe WwW 8 OOD Te rR TF
\ ‘ ; ) ae
sp" ow) & ‘ee 1 oy ; Lt ieee © AVL) OMS FRAT
a : wen al aa
5 i N f " oe sian BOW
I if Ms Siro r alee . \ (i 7 THT 4
\
af ’
i
{ \
j
!
\

Discussion

English Equivalents

Acommercially developed nut and fruit tree shaker appears suitable for harvesting cones
in Douglas-fir seed orchards where trees are 7- to 9-m tall, but some upper crown damage
is unavoidable if significant nuymbers of cones are to be removed. It was evident that a
portion of each tree’s cone crop will remain on the tree even if sufficient power is used to
cause extensive breakage to the upper crown. When low energy shaking (low cycles per
second) was used, good cone removal occurred from the upper third of the crown,
whereas results were varied for the middle third and were poor for the lower third of the
crown. Yet an average of more than 50 percent of the cones were shaken from each tree.
It is thought that an acceptable balance between gain from cone removal and loss from
top breakage was obtained in this trial.

Tree size and shape had a definite effect on how many cones were removed by shaking.
Future research is needed on the effects on crown shaping or branch pruning in order to
produce trees that can be shaken more effectively. Trees less than 10 m tall are probably
more difficult to effectively shake than larger Douglas-fir trees because a greater
proportion of the cone crop in small trees is located in the hard-to-shake, lower half of
each tree. Large trees tend to have a greater proportion of cones in the upper third of the
crown, where they can be most easily shaken; thus, it is likely that the percentage of
cones that can be removed will increase as the orchard trees get larger.

Shaking removed the highest, most difficult-to-reach cones from each tree. Cone pickers
in the Beaver Creek seed orchard found that hand picking cones remaining on shaken
trees was easier and faster; consequently, shaking should be less expensive than when
ladders or lift devices are used to collect cones from unshaken trees. Gentle shaking
removed a significant portion of the cones, yet the shaking caused only minimal upper
crown damage. Our results indicate more than half the cone crop could be machine-
harvested from 7- to 9-m tall trees. The 3-minute average per tree obtained in this shaking
test was much slower than could be done in a production operation. We spent
considerable time experimenting with different heights at which to attach the shaker head,
positions around the trees from which to shake , and often used longer and slower shakes
than were really necessary. We estimate that about one tree per minute could be
harvested during good cone crops if a shaker and mechanized catch frame are used. An
additional advantage of tree shakers is that they are not restricted to gentle terrain as are
most lift devices. A combination of machine harvesting and hand picking will speed up
cone harvest considerably and will greatly reduce collection costs per bushel.

1 centimeter (cm) = 0.39 inch
1 meter (m) = 3.28 feet

1 kilogram (kg) = 2.20 pounds
1 hectare (ha) = 2.47 acres

Wiz) snshutaebe I vlan la aa haw)
- CPR NMR Aerial atic Hanh Ry teh fee mat anaton
he = oe ea i heb Co vi ee nee pine AIT

| ares rt ero i ian avr ae soap svete bh 008; penne |
ae 6 bac bcthin i et ign eae Kins Paint fbb wrth? Getae Brgw prin

a ea a |
wie vl hn av ation mad ecole © Sad adiiqenos rs Met | i
} | a esiudbehcader

bliss pitberw Spares ponent i cetin Fi a aaa
aI. A elite TTR we IDL iti: ANGI Ne: erate, esting meee
_ eben k at Fe et et Agewtiie Tie exerts mamaria ;
| tng #oene tad inlet @. miguel wey at nani
gh MeV aeariay,oteoetaencnt Anil tart) i Betagie ) doer Nanivesn
| vit. Le rapes! SFY NL RON Ne ANON OND Ss inte DIEM
Ma iyi Be ned agi A) ARTI eR Rie ROT
oe mo a ene Tae

7 et no et ily UA oie |
apatites ge crane era scien |

| pole wie inten pero ‘ deci his
lane ane ln oi date SVC

i Ben) \ tem et ane Rete se baw Beast pubs

ete ¢ aueeatint BI Gt raion baa alt ernie 4) es

+ ge) Kaeeigpey iw: (em lly formant tira Bold gvnuciy i prnirrenaner bite a Me
weita@nid So alata ey pec ave

. aein