Historic, archived document
Do not assume content reflects current
scientific knowledge, policies, or practices.
OCCASIONAL PAPER NO- 69 OCTOBER 11, 1937
SOUTHERN FOREST EXPERIMENT STATION
E. L. Demraon, Director
New Orleans, La.
A METHOD OF STUDYING THE FACTORS
AFFECTING INITIAL SURVIVAL IN FOREST PLANTATIONS
by
Philip C. Wakeley, Silviculturist,
and
R. A. Chapman, Assistant Silviculturist
The Occasional Papers of the Southern Forest Experiment
Station present information on current southern forestry prob--
lems under investigation at the Station, In some cases these
contributions were first presented a& addresses to a limited
group of people, and as "occasional papers" they can reach a
much v/ider audience. In other cases, they are summaries of
investigations prepared especially to give a report of the
progress made in a particular field of research. In any case,
the statements herein contained should be considered subject
to correction or modification as further data are obtained.
A METHOD OF STUDYING THE FACTORS
AFFECTING INITIAL SURVIVAL IN FOREST PLANTATIONS
by
Philip C« Wakeley^ Silviculturist,
and R® A® Chapman, Assistant Silviculturist,
Southern Forest Experiment Station®
At one time or another almost every man in charge of a planting opera-
tion faces the necessity of studying the effect of various factors upon ini-
tial survival ^ of the planted trees® Such studies are particularly likely
to be needed when changes in nursery practice, planting tools, or available
labor cause a change in the general method of planting ^ when a new species is
added to the list of those planted; when planting is extended to a new terri-
tory; or when a review of accounts shows an unduly high cost per thousand
trees established® Often, however, the demand for such studies arises at the
height of the planting season, when the men in charge have scant time to de-
vise experiments® In addition, the men may have had relatively little ex-
perience in investigative planting® As a result, the numbers of trees used
in the studies may be either too small for reliability or too large for effi-
ciency, or flaws in arrangement of the plots may cause variations in soil to
make insignificant factors appear important or significant ones appear negli-
gible ®
This paper describes a method of laying out plantations and of recording
survival counts which the Southern Station has found particularly well adapted
to the study of most of the factors affecting survival. The method^ which re-
quires that the plantations be laid out in what are known as "balanced, random-
ized blocks", is based on one developed largely by R. A. Fisher. Althoxjgh at
first glance it appears considerably more complex than the old-fashioned "plot
and check-plot" studies, or than plantations of two treatments in alternate
rows, it is actually very easy to apply if the directions given here are fol-
lowed step by step. The little extra trouble involved is more than offset by
the greater economy of planting and increased dependability of results. These
advantages are attained by the internal arrangement of the experimental plan-
tations, which largely eliminates the disturbing effects of irregular moisture
supply, variations in soil, and differences in competing vegetation, and which
makes it possible, by planting a relatively small number of trees, to measure
accurately tne net effect of the different treatments being studied.
1/ Initial survival may be defined as that shown by the planted stock after the
occurrence of the mortality more or less directly brought about by the whole
process of planting, including the environment chosen for the planted trees.
It is usually best measured at the end of the first growing season, but in spe-
cial cases supplementary measurements may be desirable about the first week in
June of the first growing season and at the end of the second or third growing
seasons. A study of initial survival should not include effect of epidemics,
insect infestations, or droughts after the planted trees are well established,
and of course never includes later losses arising from competition as the
crov\Tis close.
Many modifications of the method are possible, , but unless the planter
is an experienced statistician, it is suggested that he follow strictly the
application described here, lest he weaken his results or make them entirely
unreliable •
Material and Equipment Required
In addition to the necessary planting tools and planting stock, the
following materials and equipment will be required to carry out the study;
much time and annoyance (and possibly serious errors in the work) can be
avoided by preparing them in advance ^
1, Office equipment;
a« A set of wooden "lotto" or "bingo" numbers
running from 1 to 60 or higher, for randomizing® They
can be obtained at almost any 5- and 10-cent store®
Other numbered objects, such as slips of paper or flat
metal tags, can not be mixed thoroughly in a reasonable
time and should not be used,
b« A tin can with straight sides, or a cylindrical
oatmeal box, for shaking up the numbers®
c® One establishment report diagram (form 1) for
each block® This form is more convenient if typewritten*
d® One establishment report summary (form 2) for
every four blocks® This also should be typed if possible.
e® One field examination record (form 3) for each
examination of each treatment ; as a relatively large
supply of these forms is involved, mimeographing them is
recommended rather than typewriting®
2® Field equipment;
f® Large, durable corner posts sufficient to mark
"block" corners®
g. Enough small, durable stakes — 2 feet is a good
length — to mark each end of every row®
h® At least one measuring cord, 26 times as long as
the distance between one tree and the next In the planta-
tion® From two to eight cords will be needed in large
studies® Each cord must have a loop at each end to be
slipped over the stake at each end of the row, and must be
marked at equal intervals, with paint or tags, to show
where 25 trees are to be planted in the row, at the spacing
chosen. (A spacing of 5 feet between trees and between
rows is suggested, as it reduces the size of the plot
and makes it easier to find trees on re-examination, but
6x6 feet or any other moderately close spacing will serve.)
- 2 -
Size and Arrangement of the Plantation
At any given spacing, the area required for an experimental plantation
depends upon the number of treatments to be tested* The method described
here is flexible enough to permit from 10 to 60 different treatments in one
experiment* It is best adapted, however, to testing l6 to 30 treatments at
once* Actual practice has shown that studies of less than 10 or more than 60
treatments made as described in this paper are generally inefficient and lack
desirable precision.
The plantation layout has been developed to test each treatment by
means of 100 trees. The total number of trees in a study, therefore, equals
100 times the number of treatments. These trees are planted in four blocks
of identical size and shape, laid out as follows?
The length of each block is 26 times the distance between two consecu-
tive trees in the row, that is, long enough to accommodate rows of 25 trees
each. The width of each block is obtained by adding 1 to the number of treat-
ments to be tested, and multiplying this sum by the distance between two ad-
jacent rows, which should be the same as that between two trees in a row. For
20 treatments spaced 5x5 feet, there would consequently be required 4 blocks
each 130 x 105 feet.
These four blocks should be laid out on as uniform a site as possible.
Furthermore > it is especially important that conditions within each block be
uniform; if the site as a whole shows differences that cannot be avoided, the
blocks should be laid out so that there is less difference between two parts
of any one block than betvifeen tv/o blocks each taken as a whole.
On a fairly uniform site the blocks should be laid out end to end in a
rectangle; this arrangement has the advantages of simplifying work and saving
stakes. On a site marred by small drainage ways, areas of strikingly differ-
ent soil, or patches of heavy brush, these advantages should be sacrificed and
the blocks separated to exclude the disturbing factors; this precaution is
necessary to maintain the comparability of rows within each block. The great-
est-gap between blocks ordinarily should not exceed 100 feet.
When the corners of the four blocks have been established, the next
step is to drive stakes at the ends of the blocks, at the spacing decided
upon, to mark the beginning and end of each row. The blocks should then be
given the designations A, B, C, and D (or other convenient designations),
beginning with the block farthest from the road or other avenue of approach.
Each series of stakes marking the ends of the rows should be numbered from 1
to 20 (or whatever the number of tests is), with the numbers running from
left to right, (Paint, metal tags, or a good, permanent grade of lumber
keel may be used to mark the row stakes.)
This system of lettering and numbering is suggested because, as one
approaches the plantation from the road gate, it makes an orderly arrangement
of block letters, row numbers, and tree numbers, running in the same direction
as the writing on a sheet of paper (fig. 1). As will be seen in the sections
on planting and on re-examination, ease in finding any particular row is es-
sential to efficient use of the method.
~ 3 -
Arrangement of Tests Within Blocks
The blocks can be laid out in the field, staked, and numbered as soon
as it has been decided how many methods are to be tested and what spacing is
to be used. Before the actual planting can be done, however, the arrangement
of tests within blocks must be determined. As the method of experimental
planting under discussion depends for its success entirely upon this arrange- *,
ment, the following directions must be followed exactly;
1, List the treatments to be tested, for example, 20, in any conve- \
nient order on form 2, and number them from 1 to 20, as in the example on j
page 15. j
2, Take lotto numbers 1 to 20 and place them in a cup.
3, Take a copy of form 1 and head it for block A, as in the example
on page 13.
4, Shake the lotto numbers thoroughly and blindly draw one from the
cup without exercising any choice in any way. Copy this number in the second
column on form 1, opposite row number J., Lay the drawn number aside; do not
return it to the cup.
5, Draw the rest of the numbers and copy them in the second column of
form 1, opposite row numbers _2, _3, etc,, laying each aside as soon as
copied. This process gives the order in which the various treatments (20
rov/s of 25 trees each) will be arranged in block A, and in block A only.
6, Head a second form 1 for block B, replace the lotto numbers in the
can, shake thoroughly, draw the numbers at random as before, and copy, as
drawn, in the second column, to give the order of treatments in block B.
7, Repeat the process for block C; also for block D,
8, Now turn again to form 2, which gives the nature of the treatment
corresponding to each treatment number. Using this as a guide, enter the
nature of the corresponding trea.tment opposite each treatment number that has
been entered on form 1 for block A, for block B, etc. Form 1 for each block
thus becomes a diagram, from which the actual method under test in any row
can be read at any time; the row number can of course be read instantly from
the stake at the beginning or end of the row.
9, While entering on form 1 the nature of treatment read from form 2,
enter also on form 2, in the appropriately lettered column, the number of the
row in which that treatment is being tested in the block. This makes of form
2 a cross-index to the location of treatments — an Index which is invaluable
not only in supervising and checking the work of planting but also in making
out the forms for later survival counts,
A word of caution is necessary concerning the process of selecting at
random the rows to which the various treatment numbers are assigned. Each
block in each experiment must be randomized afresh. Except in the rare event
that an identical arrangement appears by chance, the same arrangement must
not be used twice, either in the same study or in different studies. If it
does appear by chance, it must be used, because the numbers must be accepted
- 4 -
as they fall« The person arranging the layout may no more reject a series
of drawn numbers "because he doesn’t like the looks of it" or because it re-
sembles or duplicates a previous series ^ than he may substitute a systematic
arrangement of tests for the random arrangement upon which the ultimate suc-
cess of the method depends* If lotto numbers of uniform size are used, and
if they are well shaken in a cup that is large enough to permit free movement,
the experimenter may feel confidence in such randomizations, which are abso-
lutely essential to the measurement of the precision of results obtained in
the study.
Establishment of the Plantation
The following precautions must be observed in the actual establishment
of the plantation, lest disturbing factors enter in and obscure the effects
of the treatments to be tested*
1* Except in a test of season-of-planting , or of weather-at-time-of-
planting , all four blocks should be planted in one day. If they are not,
some of the stock will be subjected to one set of weather conditions (risk of
drying out, etc.) and some to another. In fact, a week of bad weather inter-
vening between two planting days may seriously derange the experiment. In
the example cited on forms 1 and 2, involving method and duration of storage,
the stock should be lifted at 5-day intervals and planted all in one day,
instead of being lifted all in one day and held for progressively longer
periods.
2. It is best if one man (or one pair of men working together) plants
all four blocks. If the number of treatments to be tested makes this impos-
sible within the limit of one day, two men (or two pairs) should each plant
two blocks, or four men (or pairs) should each plant one block. Two men
should never divide the work on any one block.
With the treatments given on forms 1 and 2, for example, one man should
not plant all the heeled-in stock and another all the baled stock, because, if
the baled stock showed poorer survival, it would be impossible to tell whether
losses were the result of baling or of poor planting by one man. Similarly,
if one man plants both heeled-in and baled stock on blocks A and B, and a
poorer planter plants both stocks on blocks C and D, and if blocks C and D
show a low average survival for all treatments, it will be impossible to tell
whether the losses on blocks C and D were caused by poor planting or by poor
Soil on those blocks. This, however, is not important, as the study is of
treatments . not of planters or plots of ground, and those particular blocks
(and probably also the particular planters) will never be used again. The
important thing is that stock from each storage method (heel-in and bale) be
equally exposed to both good and bad soils and to good and bad planting so
that the true differences between the storage methods, if any, will be brought
out.
3. Even at the expense of extra walking and loss of time, it is usual-
ly better to plant treatments in numerical order, and to complete each on all,
four blocks before going to the next treatment, than to plant row 1, row 2,
etc., of each block consecutively. The example on forms 1 and 2 illustrates
this point clearly. If the bale for treatment no« 2 is opened to plant row 7
of block A, and left open during the planting of rows 8-20 in block A, all of
- 5 -
blocks B and C, and several rows in block D, the stock drawn from this bale
for the appropriate row in block D may have been very seriously affected by
changes in temperature and moistui*e» Figure 1 shows the logical progress of
planting, through the planting of the first 175 of a total of 2,000 trees*
With these precautions in mind, the simplest way to supervise a small
experimental planting is to give one man a measuring string (or, preferably,
four strings) and the form 1 for all foUf blocks, while a second man (or
pair), equipped with the necessary tools, gets the stock for each treatment
in turn, under the direction of the man in charge. It is advisable to count
the. stock into four groups of 25 trees each before distributing to the plant-j
er^ labeling each group with block, row number, and treatment number, read
from form 2 by the man in charge.
The man with the marking string glances down the second column of form
1 for block A, sees that treatment 1 is to be planted on row 2, and accord-
ingly stretches a measuring string from the no* 2 stake on the far side of
block A to the no, 2 stake on the near side. Then the planter, or pair of |
planters , plants one tree at each mark on the string and tucks the label from
the bundle of trees into the grass at the foot of the stake at the far end of ^
the row. Meanwhile the "string -man” goes to block B and stretches a string
between the appropriate stakes (marking row 17), shown by the position of
treatment no. 1 in the second column of form 1 for block B,
Even on the largest studies, with four blocks of 60 rov;s each, the
same system works* A study of this size, however, requires four planters
(or pairs); each block then requires one planter, a string -man, two strings,
and, if possible, a carrier to relieve the planters from walking back and
forth with supplies of stock. Under this arrangement, the string-men carry
the copies of form 1 for their respective blocks.
The Southern Forest Experiment Station, working on blocks laid out as
described with a crew of 13 experienced GCC enrollees (4 string-men, 4 plant-
ers, 4 carriers, and 1 tree-counter to assist the officer in charge) and us- ;
ing 8 strings, has planted 7,200 trees in 5 working hours, without undue haste J
The stakes, of course, had all been driven and numbered in advance.
Each string -man, as he finishes stretching a string, should check off
both row number and treatment number, in pencil, on form 1. When the plant-
ing has been completed, the man in charge of the work should go along each
row of stakes, picking up the labels removed from the groups of trees and
checking the stake number and label against form 2, If any error has oc-
curred, forms 1 and 2 should be corrected to show the actual location of the
treatments on the ground, as these forms are an essential part of the record
and constitute the guide to re-examination.
Forms 1 and 2, completely filled in, together with a map showing the
location of all the blocks, and with descriptions of the stock (including
seed soui'ce), planting site, weather during planting, and treatments being
tested, constitute all the establishment report needed.
- 6 -
Type of Survival Study to Which the
Foregoing Layout is Particularly Adapted
The layout just described is especially adapted to the study of
factors affecting the individual tree rather than the entire planting site.
Such factors are the species, age, class, grade, health, or size of the
planting stock; its treatment by cultivation, fertilizing, shading, or water-
ing in the nursery; root pruning during the growing season or at lifting
time; mechanical injury or exposure during lifting or planting; method or
diiration of storage of planting stock; season of planting; depth of planting;
tool used; localized site preparation, such as plowing of furrows or hoeing
of spots; control of pests by various sprays; and control of drought injury
by shading, mulching, coating with oil, or pruning the needles. Most sur-
vival studies involve just such factors.
The layout of four balanced blocks, each with treatments randomized by
25-tree rows, is not adapted to studies of the effect of treatments applied
to the site as a whole rather than to the individual tree. Such studies in-
clude eradication or control of animals, insects, or diseases by scattered
baits, poison applied to burrows, or removal of alternate hosts; control of
competing brush large enough to exert its influence beyond the width of a
row; preparation of the entire site, as by burning or by complete plowing and
harrowing; and effect of soil type or topography. It is obviously impossible
to get typical burns on single rows, or to arrange ridges, slopes, and draws
according to the chance order of the lotto numbers. For studies of this kind
special techniques must be developed.
In choosing the treatments to include in one study, that is, in one
set of four blocks, care must be taken to compare similar things. Except as
noted in the following section, any one study should be confined to varia-
tions of one factor (such as root length, storage, or planting tool), or
should test several complete planting systems such as have been developed and
found effective in different territories, A study should not include compar-
isons between "date of planting" and "2~year-old stock", or between "seed-
lings fertilized in nursery" and 'teeedlings heeled-in for 40 days"; such mis-
cellaneous comparisons are meaningless.
Special Cases
A modification of the general method very greatly increases the effec-
tiveness of the layout in many instances, especially if the number of avail-
able treatments (for example, tools, or root lengths) falls below 15. This
modification involves applying each of several original treatments (perhaps
10, as in the case of tools) to each of two species, or to each of two ages
or grades of planting stock. The combination of several original treatments
with two species is especially effective; in fact, three or even four species
may sometimes be used to advantage.
In applying this modification, each combination of original treatment
(such as tool) with species constitutes a separate final treatment. For ex-
ample, 10 tools tried simultaneously with longleaf and slash pine would give
a total of 20 treatments to randomize over each block, with each tool repre-
sented by 50 trees (25 longleaf and 25 slash) on each block, and each species
by 250 trees (25 for each tool) on each block.
- 7 -
There are two advantages of this special modification; One is that it
makes it possible to test a limited number of original treatments (such as 10
tools) and still bring the number of rows in each block within the range (16
to 30) which appears most effective; the other is that it brings to light any
differential response of species (or age or grade of stock) to treatment.
One tool, for example, may work well with longleaf and poorly with slash pine,
or the shortest permissible root length for one species may differ by 2 or 3
inches from that for the other. Such information is extremely valuable in
determining planting policies and in supervising routine commercial planting
of either species.
Re-examination of the Plantation to Record Survival
Re-examination of the plantation, apparently complicated by the random
arrangement of treatments in each block, is actually very simple, provided
only that the numbers on the stakes marking the rov/s are distinct and that
form 2 is co:^ectly made out. The ease of examination results from the use
of form 3, -2/ one copy of which is needed for each treatment being tested.
To prepare the forms for re-examination, enter at the head of the -j
first copy of form 3 the study, species, and nature of treatment for treat-
ment no. 1, taken from form 2 as made out when the plantation was established -4
(see pages 4 and 15). At the tops of the four main columns on form 3 enter
the block designations, A, B, C, and D, as shown. Then, from form 2 take the '-‘I
number of the row in which treatment no. 1 was planted in each block, and
enter it under the designation of that block on form 3. In like manner pre- ^
pare forms for treatments 2, 3? etc. (see page 17).
For most efficient field work in re-examination, go first to row 1 in i
block A. Take the sheets (form 3) for all 20 treatments (or whatever the
number is) and arrange them in numerical order according to the row numbers |
entered in the first main column, under the heading "Block A." Go down row 1, %
entering the condition, etc., of each tree opposite the tree number on form 3. |
Tree 1 in row 1 should be in the far, left-hand corner of the block as it is
approached from the road or gate, as suggested on page 3 and in figure 1. A 1
measuring stick may help to speed up the work if survival is low or if the ^
trees are hard to find in the grass; in extreme cases it may save time to
stretch the original measuring string along the row. Since exactly 25 trees
were planted in the row, the list of trees living, dead, or missing must total
25 in the corresponding column on form 3*
When the main column for row 1 of block A has been filled in, replace
the form 3 for that row with the one for row 2, move over one space in the
plantation to row 2, and enter the record for that row on its form. It is
not necessary to walk back to the no. 1 tree in row 2; the row can be run in
reverse by starting at the bottom of the main column on form 3j opposite tree
no. 25^ and working upward on the form.
2/ Form 3, originated by C. F. Olsen, Assistant Silviculturist, Southern
Forest Experiment Station, is essentially a mechanical means of "unrandomiz-
ing” the randomized rows for purposes of re-examination and averaging. The
reduction in cost and in errors resulting from its use in an important factor
in making the method of randomized balanced blocks feasible for extensive
studies of survival in forest planting.
- 8 -
When the last (for example, the 20th) row in block A has been examined,
in each of the 20 forms the first, or left-hand, main column will be filled
in, and the other three main columns will be blank , The next step is to ar-
range the 20 forms in numerical order, based on the row numbers in the second
main column, for block V/hen this has been done, proceed with the examina^
tion of block B as with block A, and repeat for blocks C and D^
For the sake of uniformity in accuracy and judgment, it is best to have
one man re-examine all the trees in one study » In the case of very large
studies, or of great numbers of studies, it may be necessary to work with sev-
eral assistants. If four assistants are used, one might be given the first
five rows on each block, the second the next five, and so on» By giving the
same man the same portion of each successive block, slight differences in ac-
curacy and judgment are "averaged out" by the random arrangement of treat-
ments, just as are slight differences in soil within each block. Gross dif-
ferences in accuracy and judgment, however, can not be tolerated.
Form 3, as given, provides for records of vigor and injury^ as well as
of survival. Such records, if carefully taken, may add greatly to the value
of the study, and their inclusion tends to increase the accuracy of the sur-
vival counts. At the bottom of form 3 are shown symbols (capital letters for
condition and vigor, small letters for injuries) found useful by the Southern
Forest Experiment Station; other symbols may be added as needed. It must be
remembered, however, that the plantation has been established primarily to
study survival. All other information is of secondary importance, and form 3
may therefore be simplified by omitting all "vigor" and "injury" columns and
retaining merely a "condition" column for each block.
Analysis and Summary of Results
Since each treatment has been tested by planting 100 trees (25 per
treatment per block), the total number of living trees recorded on form 3
for that treatment constitutes the survival percentage.
The survival percentages for the various treatments may be entered
along the right-hand margin of form 2, or summarized or tabulated in any oth-
er manner desired. The ideal procedure is to make an analysis of variance
and to test the significance of / the differences betv/een the survival percent-
ages of different treatments, ^ but this requires both extra time and
j/ This method of analysis was first used by Fisher and is presented in;
Fisher, R„ A. Statistical methods for research workers. Ed« 5# rev/ and
enl. 319 PP«? illus. London 1934 »
For an elementary description of the analyses, see;
Bruce, Donald, and Schumacher, Francis X. Forest mensuration.- 360 pp-,,
illus. IlZ] New York and London. 1935.^
Snedecor, George W. Calculation and interpretation of analysis of variance
and covariance. 96 pp., 35 tables.. Monograph No. !<, Division of Industrial
Science, Iowa State College. Collegiate Press, Inc., Ames, Iowa. 1934«
Ti^ett, L« H. C. The methods of statistics, 222 pp., illus. [chs^ VI and
X.J London. 1931.
- 9 -
considerable familiarity with statistical technique.
It should be emphasized, however, that these elaborate statistical
analyses merely measure and describe the precision and reliability of the
results, and are useful only because they show what the odds are that the
observed differences in survival are the result of the treatments and not
of chance. The precision and reliability have already been attained by the
randomization of treatments within balanced blocks, and are there whether
measured and described or not. In other words, the man establishing an ex-
perimental plantation may be assured that differences in survival shown by
several treatments tested according to the methods here described reveal the
relative value of the treatments much more clearly than they would if each
treatment had been confined to one lOO-tree row running across the entire
planting site .
The principle involved can be illustrated very simply. Suppose 20
treatments are being tested, by means of 100 trees each, planted in a rectan-
gular block in which there is an evident decrease in soil quality from north
to south. The careful experimenter presumably would have his rows run north
and south, so that some of the trees in each row would fall on good soil and
some on poor. Suppose, however, that for the width of one or two rows along
the west edge of the plot there is a strip of soil that appears to the eye to
be like the rest, but is actually much more favorable to initial survival.
If each treatment is confined to a single row running the length of the plot,
the one or two rows falling on the strip of good soil will show an initial
survival out of all proportion to the merits of the treatments tested in
those rows. But if the plot is divided into four equal blocks, each 25 trees
long, and the treatments are rearranged at random in each block, the odds are
very small that any one treatment will fall on the strip of good soil in all
four blocks. If it falls elsewhere, even in only one block, the average sur-
vival of that treatment is decreased, and the average survival of the treat-
ment that replaces it on the good soil is increased, thus reducing the dif-
ferences in average survival arising solely from soil fertility. The princi-
ple that applies in the case of such a narrow strip of favorable soil applies
equally to soil or other variations of any other size or shape.
Summary of Precautions
To aid in avoiding delay or errors in establishing survival studies,
the following precautions are summarized in the order in which the need for
them is likely to occur;
1. Have clearly in mind the kind and number of treatments to be
tested, being sure that all those in one study may be legitimately compared
or contrasted.
2, Record the description of each treatment in writing, so that the
successful treatments can be unquestionably duplicated and the unsuccessful
ones avoided in later practice.
3* Before beginning actual installation, obtain lotto numbers, a re-
ceptacle for shaking, a supply of forms, the required number of marking
strings, and the necessary corner posts and row stakes.
- 10 -
4. In laying out the plantation, take every reasonable precaution to
have uniform conditions throughout each individual block*
5e If unavoidable differences occur over the planting area as a whole,
lay out the blocks so that the greatest differences occur between one block
and another, and the least possible differences within any one block*
6, Keep blocks fairly close together, each preferably v;ithin 100 feet
of the next, if not actually adjoining*
7* Mark the blocks with durable posts, plainly labeled, before plant-
ing; the blocks should also be mapped for the office record*
8. Stake the rows before planting*
9. Number the rows systematically, preferably as shown on page 3 and
in figure 1*
10. In randomizing treatments over block diagrams (form 1), mix the
lotto numbers very thoroughly, and draw, list, and retain the numbers as they
come, without bias, laying each number aside as drawn.
11* Check forms 1 and 2 against each other before beginning planting*
12. Plant the entire study in one day (unless it is a study of weather-
on-day-of-planting, or of date-of-planting ) * Do not start work if rain
threatens to prevent completion*
13 e Use one planter or pair of planters for the entire study, if the
study is small enough for them to complete in one day. If it is too large
for this and more men must be used, be sure that each block is planted by the
same man (or pair) throughout* Within any one block never have one species
or treatment planted by one man, and another by another man*
14. Be careful not to introduce any disturbing factors, such as using
seedlings of one seed source or date of sowing for one treatment and of a
different source or date for another; exposing part of the stock in treatment
no. 1 to a much longer period of drying than the rest; or dipping part of the
stock in water and planting the rest without dipping . In the example cited
on pages 4? 13, and 15, the "fresh checks" for the storage study must also
be baled and heeled-in, if only for a minute, lest they benefit from less
handling as well as from prompter planting than the other stock.
15 « Before leaving the planting site, be absolutely sure that the
records on forms 1 and 2 agree with the location of the treatments on the
ground.
16. Before beginning re-examinations, check all copies of form 3 for
accuracy of species, treatment, and block and row designations*
17. In re-examining any one block, be particularly careful to avoid
entering data for two rows on the same sheet of paper. Take a new sheet for
each new row of trees.
11 -
18a In running from the high-numbered to the low-numbered end of
the row, remember to begin at the bottom of the corresponding column on
form 3 .
19. Allow plenty of time for re-examination and thus avoid careless
inaccuracies. Search thoroughly before calling a tree (=missing, or
blank). If this work is entrusted to subordinates, be sure that they are
adequately trained and instructed, and check their work by examining a num-
ber of rows without letting them know which rows you are going to check,
20. Be sure that every form (1, 2, or 3) includes the name or de-
scription of the study (including species), the season of establishment, the
date the form was filled in, and the name or initials of the man responsible
for the accuracy of the data.
SAMPLE FORiM 1„ ESTABLISH?4ENT REPORT DIAGRAM
Block A* Study; Test of method and duration of
storage of slash pine planting stock
Planted by on 193_.
Supervised by_
Row No.
Treatment No.
Nature of treatment
1
19
Heeled-in, then taken up and planted immediately
2
1
Heel-in 5 days
3
15
Heel-in 40 days
4
17
Heel-in 45 days
5
16
Bale 40 days
6
3
Heel-in 10 days
7
2
Bale 5 days
8
18
Bale 45 days
9
6
Bale 15 days
10
13
Heel-in 35 days
11
10
Bale 25 days
12
9
Heel-in 25 days
13
5
Heel-in 15 days
14
11
Heel-in 30 days
15
12
Bale 30 days
16
8
Bale 20 days
17
4
Bale 10 days
18
7
Heel-in 20 days
19
14
Bale 35 days
20
20
Baled, then unpacked and planted immediately
- 13 -
SAMPLE FORM 2, ESTABLISHMENT REPORT SUMT^ARY
Study; Test of method and dixration of
storage of slash pine nursery stock
Established by on , 193_*
Treatment No.
Nature of treatment
Row occupied by treatment in block;
A B C D
1
Heel-in 5 days
2 17 18 11
2
Bale 5 days
7 4 17 etc.
3
Heel-in 10 days
6 20 etc.
4
Bale 10 days
17 etc. etc.
5
Heel-in 15 days
13 etc.
6
Bale 15 days
9
7
Heel-in 20 days
18
8
Bale 20 days
16
9
Heel-in 25 days
12
10
Bale 25 days
11
11
Heel-in 30 days
14
12
Bale 30 days
15
13
Heel-in 35 days
10
14
Bale 35 days
19
15
Heel-in 40 days
3
16
Bale 40 days
5
17
Heel-in 45 days
4
18
Bale 45 days
8
19
Heeled-in 5 then
1
taken up and planted
immediately
20
Baled, then impacked
20
and planted immedi-
ately
- 15 -
I
>
1
■j
■ I
A
SAMPLE FORf'/l 3'. FIELD EXAiMINATION RECORD
Study; Test of method and duration of
storage of slash pine planting stock
Established by on , 193_*
Re-examined by_ on , 193_«
Species; Slash pine Nature of treatment; No, 1 Heel-in 5 days
Tree
No.
Block A
Block B
Block C
Block D
Notes
Row No« 2
Row No. 17
Row NO'e 18
Row No* 11
Con-
di-
tion
Vigor
In-
Con-
di-
tion
Vigor
In-
jury
by
Con-
di-
tion
Vigor
In-
jury
Con-
di-
tion
Vigor
In-
jury
by
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
'
20
21
22
23
24
25
To-
tals
Condition
L = living
D = dead
B = blank
(missing)
SYMBOLS
Vigor
very thrifty
T2 = thrifty
Gj = growing
G2 = growing poorly
F^^ = failing
F2 = almost dead
- 17 -
Cause of injury
a = ants
e = erosion
d = drought
f = fungi
g = gophers
i = insects (not ants)
p = poor planting
r = rabbits
s - silting
i
\
I
V
Figure L Diagram of storage siady ploni-aHon described in iex-f,
shor/mg locoiion of firsi /75 -frees p /an fed.
/?o»v N urr? her
- 19 -
I
?