DENDROPEDAGOGY:
H O W- T O - D O - I T
Teaching Botany!
Ecology & Statistical
Principles Through
Tree"Ring Studies
D A R R I N L. R U B I N O B R I A N C. M C C A R T H Y
I ntroduction
When we meet people in the woods, we are
often asked if it is true that one can tell the age of a
various research arenas: forest dynamics, recon-
struction of past climate, dating archeological sites,
dating forest fires, and even forensics. Needless to
say, dendrochronology is an exciting field with
tree by counting its rings. We tell them that not many subdisciplines sure to interest many students.
only can age be determined but also that a glimpse
Tree-ring studies offer an excellent opportunity
at the tree’s history can be found in the rings. In
for field- and lab-based teaching exercises. First,
temperate regions, trees deposit a single layer of
general public interest in forests and forest ecology
wood around their circumference every year.
can be used to grab the students’ attention; we have
Dendrochronology (dendron = tree; chronos = time;
yet to meet a student who doesn’t pick up a dated
logos = the study of) is the science that uses tree
piece of wood, briefly study it, and begin to ask
rings, dated to their exact year of formation, to ana-
numerous questions. Secondly, tree-ring studies are
lyze temporal and spatial patterns of various
truly multidisciplinary and require the integration
processes (e.g., biological, physical, or cultural).
of information from several fields: botany, ecology,
The width of a tree ring in a given year represents
physiology, geography, climatology, and statistics.
the radial-growth response of the tree to prevailing
Tree-ring laboratory lessons can be designed at var-
environmental conditions. The width of a tree ring
ious levels (introductory to advanced) and are easi-
can be determined by noting the pattern of early-
ly tailored to programmatic needs and equipment
wood and latewood cells that are produced during
availability. Finally, projects can be easily applied in
secondary or radial growth. During the early part of
urban, suburban, or rural schools (wherever trees
the growing season, trees often produce large, thin-
or forests are available) and can be conducted with
walled cells, and during the later period of the
either minimal or large budgets.
growing season, they produce smaller, thicker-
walled cells. Hence, an annual ring will contain It has been our experience that most students
both an earlywood and latewood portion. By meas- in the basic and applied plant and ecological sci-
uring the distance from the earlywood boundary of ences appear to be well trained in their disciplinary
one ring to the earlywood boundary of another areas. However, these same students are often
ring, the amount of wood produced during a given underprepared to actually perform a scientific
year can be quantified. Dendrochronologists have analysis using the scientific method by forming
used accurately dated and measured tree rings in hypotheses, collecting and analyzing data, inter-
preting results, forming conclusions, and reporting
their findings. Our primary goal was to design a
DARRIN L. RUBINO is an Assistant Professor in the Biology range of class experiments that could be used to
Department at Hanover College, Hanover, IN 47243; e-mail: strengthen student understanding of the scientific
rubino@hanover.edu and BRIAN C. MCCARTHY is an Associate method and hone their field data collection, labo-
Professor in the Department of Environmental and Plant ratory analysis, and report writing skills – all of
Biology, Ohio University, Athens, OH 45701; e-mail: which are necessary for students pursuing careers
mccarthy@ohio.edu. in the biological sciences.
TREE-RING STUDIES 689
Here we develop a simple tree-ring laboratory to Basic Dendrochronology
demonstrate the basics of dendrochronology.
Additionally, we provide two upper-level laboratory Investigation
exercises that are primarily intended to demonstrate the
specific dendrochronology subdisciplines of dendrocli-
Field Sampling
matology and dendroecology. These exercises can be
used at various instructional levels and can be expand- Minimally, several pieces of equipment will be need-
ed or condensed to accommodate resource limitations. ed in the field: increment borer (for collecting cores),
We have successfully applied these labs at the introduc- drinking straws (for core storage), indelible marker pen,
tory college level and tested the exercises over a period masking tape (for joining two or more straws together
of four years. We further suggest the possible use of the for long cores and sealing ends of the straws), beeswax
exercises separately or in unison as part of a multidisci- (to rub on borer tip to facilitate boring), and two wood-
plinary capstone approach to an undergraduate pro- en dowel rods and a rubber mallet (to free pieces of
gram. Below are instructions and recommendations for wood that may become entrapped in the borer bit — this
implementing the lab exercises. Additional information is a pretty common occurrence).
and a pictorial narrative of the various steps can
Field sampling begins with the selection of appro-
be found at our dendrochronology web page:
priate trees for analysis. We have found that choosing
http://www.plantbio.ohiou.edu/epb/instruct/ecology/
species with ring-porous wood (trees that form rings
dendro.htm (McCarthy & Rubino, 2001). A primer on
with very distinct earlywood and latewood) are much
plant secondary growth is also available at the web site.
easier for beginners to date. We have had excellent suc-
We suggest either requiring the students to read over
cess with most species of oaks, but other ring-porous
this very basic primer or presenting the material to the
species such as elm, ash, catalpa, black locust, and
students to refresh their botany skills.
hickory are suitable for beginners also. We have had
Also provided at the web site are detailed methods, variable success with conifers. Once a study species has
worked statistical examples, and useful links for both been chosen, carefully select individual trees to core. A
instructors and students wishing to know more about suitable tree should have a single stem at breast height
dendrochronology. and should not have large scars or deformities — these
690 THE AMERICAN BIOLOGY TEACHER, VOLUME 64, NO. 9, NOVEMBER/DECEMBER 2002
the threads have engaged, step back from the increment
Figure 1. borer and turn the handle in a clockwise motion (Figure
A. Disassembled increment borer; from top to bottom: handle, 1B). The bit will proceed into the tree. If you hit a rot
pocket (you will know immediately because of the ease
bit, extractor.
of turning), back out immediately or else your bit may be
B. A student coring an elm tree. incredibly difficult to remove. Continue turning until the
C. Removing the core from the bit using the extractor. borer has slightly gone past the pith. To gauge your depth
at any given time, you can hold the extractor up to the
side of the tree (it is the same length as the bit and will
inform you of your progress).
When the proper depth has been achieved, turn the
handle counter-clockwise one full turn and insert the
extractor into the bit while placing slight up-pressure on
the back of the extractor to ensure the leading tip stays
under your sample. Insert the extractor to its full length
and slowly withdraw the extractor from the bit; you
should retrieve an intact core (Figure 1C). Remove your
borer from the tree as soon as possible to prevent it
from being “frozen” in the tree. Immediately place the
core into a drinking straw, seal the ends, and mark the
straw with the sample ID number (long cores may
require two straws taped together). Be sure not to loose
any pieces of the core. If the core breaks, maintain the
order of the pieces in the straw and store the straws in
a protective container while in the field (an inexpensive
mailing tube or map tube works well).
Laboratory Methods
Minimally, you will need the following lab supplies:
wooden increment core holders, scissors, razor blade,
fine lead mechanical pencil (for marking holders and
tree rings), cotton string (to apply clamping pressure
during drying), dowel rod (to remove cores from
straws), and glue (to mount cores to holders). The incre-
ment core holders need to be prepared by a local wood-
worker equipped with a router table and 0.25-inch core
box bit. We typically use pine molding strips available
in 12-foot lengths from a local building supply store. Be
careful not to route the groove too deep; 55 to 65% of
the core needs to remain above the surface for subse-
quent sanding (exact dimensions for the mounts can be
trees will often have scar tissue or rot pockets that make found at the authors’ web site).
coring and dating difficult.
Remove cores from their straws, and lay them into
Your primary field tool will be the increment borer an increment core holder that has been pre-glued, mak-
which includes a handle, bit, and extractor (core retriev- ing sure that the vessels are aligned vertically in the slot.
er) (Figure 1A). They are sold at forestry supply houses Tightly wind the glued cores with cotton string so as to
in various bit lengths; we recommend borers of 30 cm apply pressure during the drying process. Use the hori-
(12”) in length. When in use, the bit is inserted into the zontal surface of the wooden holder to record core iden-
handle and twisted into the tree (Figure 1B); the tip of tity, student’s name, etc. Once the glue has dried
the bit is threaded to pull the hollow bit into the tree (overnight or microwave on low power for 2 to 3 min-
(always keep the extractor out of the bit and in a safe utes) cut the string off, and begin sanding the cores.
place while coring). To begin coring, hold the bit just Begin with a coarse grit paper (lower grit number =
behind the threads and lean into the borer to provide as coarser paper) and use progressively finer grits to attain
much body pressure as possible. Slowly turn the bit until a smooth surface (e.g., 100-, 180-, 220-, 320-grit papers).
the threads have become fully engaged in the tree. After Wrapping the sandpaper around a wooden block or
TREE-RING STUDIES 691
A high confidence in the actual date assigned to
Figure 2. each ring will be needed for subsequent analyses and
A. Counting rings on a mounted tree core; this process is hypothesis testing. Unfortunately, a tree may sometimes
have a missing ring or a false ring present. Therefore,
performed under a dissecting scope.
every ring counted thereafter is off by one year.
B. Measuring individual tree rings with a ruler; measuring is However, you can look for certain years that offer good
also performed under a dissecting scope. “signals” (very thin or very wide rings) and identify
them in each core to confirm year assignments.
Crossdating may be performed as a group activity by
having each student write on the board each year in
which very small or very large rings (signature rings) are
found. A quick check of the individual cores by the
instructor can confirm the years in which the signatures
are found. Students can now make sure that their cores
follow the dated pattern of large and small rings. After
crossdating, students may find that they need to shift
the pencil marks in one direction or the other to accom-
modate the cross-validated information. Once this
process is finished, students may assume that exact
years are assigned to each ring, and the pencil marks
should now be made permanent by using a steel dis-
secting probe to put pricks in the wood (these will not
smudge out with further handling like pencil will).
Once all of the years have been crossdated and verified,
students can begin to measure the rings.
If your teaching goal was to simply introduce den-
drochronology and its principles to the students or to
determine the age of trees, this is a suitable stopping
place. However, if you are interested in more sophisti-
cated analysis and hypothesis testing, measurement of
the individual tree rings will be necessary. Rings can be
measured to the nearest 0.1 mm using the process of
interpolation. A small plastic ruler (marked in mm) can
be held up to the base of the core. Have the students
visually line up the beginning of the earlywood and
measure to the end of the latewood (Figure 2B). A steel
rubber pencil eraser will facilitate the process. Stress to machinists rule provides better accuracy (if available).
the students that the smoother and cleaner the surface, Once measurement has been completed, the dates and
the easier the wood features can be observed. When measurements can be entered into a spreadsheet for
purchasing sandpaper, we recommend buying a multi- graphing (Figure 3) and subsequent analyses and inter-
pack that contains sheet(s) of various grits. As sanding pretation. We provide two upper-level lab exercises
progresses, the annual rings and ring boundaries will where dated and measured tree rings can be used for
become more visible. hypothesis testing.
Tentative dates may now be assigned to each ring by
using a dissecting microscope and light source. Dating
Upper-Division Option 1:
begins by counting backwards from the first known year Dendroclimatology
behind the bark (Figure 2A). Using a fine mechanical The goal of this exercise is to investigate the rela-
pencil, place a single dot on each decadal ring (e.g., tionship between tree-ring widths and climate.
1990, 1980, etc.). At this stage, these marks are just tem- Numerous questions can be addressed with this exer-
porary year assignments. Once tentative dates have been cise: How does climate affect radial growth rate? Is there
assigned to each ring, students can begin the process of evidence of severe droughts in this stand? We provide
crossdating. This is probably the most fundamental and instructions for investigating the former question.
important aspect of tree-ring research. In essence, cross-
dating involves a form of pattern matching to assure The first step is to remove the growth trend of the
accurate date assignments have been made. tree; growth trends are simply growth patterns observed
692 THE AMERICAN BIOLOGY TEACHER, VOLUME 64, NO. 9, NOVEMBER/DECEMBER 2002
reasons beyond our scope of
Figure 3. discussion [refer to Stokes &
Smiley (1968) and Fritts
White oak tree-ring widths plotted over a 68-year period (1933 – 2000).Note the growth release
(1976) for additional infor-
in the early 1960s.
mation)]. To use regression to
model tree growth, use the
7
dates as the independent
variable (x) and the ring
6
widths for the dependent
variable (y). Run the regres-
5
sion, and use the residuals
Ring Width (mm)
(the difference between the
4
observed ring width and the
expected ring width from the
3 equation for the line) for fur-
ther analysis. If no statistical
2 software package is available
for this analysis, we recom-
1 mend using the data analysis
1930 1940 1950 1960 1970 1980 1990 2000 add-in within Microsoft®
Year
Excel.
Now that detrened val-
ues have been obtained, the
relationship between growth
and climate can be performed using correlation
Figure 4. analysis. We have had excellent success using
Scatter plot of the residuals (detrended ring widths) vs. total growing annual total growing season precipitation for
season precipitation. Note the positive relationship between residual analysis. Local climate data (if not available from
and amount of precipitation. local sources such as airports or campus weather
stations) can be easily obtained from the National
Climatic Data Center and National Oceanic and
60 Atmospheric Administration (2001) web site:
http://www.ncdc.noaa.gov/onlineprod/drought/
55
Growing Season Precipitation (cm)
r = 0.742 xmgrg3.html. Summarizing and re-organizing the
50 P < 0.001 data may be required so we suggest downloading
and formatting climatic data ahead of time to save
45
precious lab time. To test the relationship
40 between radial growth and precipitation, perform
correlation analysis using the residuals and pre-
35
cipitation data for each year. Again, the add-in for
30 data analysis in Microsoft® Excel works satisfacto-
rily if no statistical software is available. If you do
25 not want to statistically test the relationship,
20 graphical inspection will prove satisfactory for
-1.5 -1.0 -0.5 0.0 0.5 1.0
beginners (Figure 4). Graphing can be accom-
plished using most spreadsheet applications. See
Residual
Rubino and McCarthy (2000) for an example of a
dendroclimatic analysis of a forest.
in a tree over a given time period. By removing the
growth trend, we can more accurately relate tree growth
Upper-Division Option 2:
to climate. To remove the trend, we recommend simple Dendroecology Lab
ordinary least squares regression to statistically model Numerous ecological questions can be answered
tree growth. Actual tree-ring widths are not used for cli- using tree rings. For example, tree rings can be used to
matic analysis since they do not perform as well as answer questions such as what is the frequency of dis-
detrended values for numerous statistical and biological turbance in this forest? or How has logging affected tree
TREE-RING STUDIES 693
growth in a stand? Disturbances can be putatively identi-
fied by release events — sudden, sustained increases in
radial growth rate. It is hypothesized that releases are due
to changes in competitive relationships among trees. For
example, when a neighboring tree is removed (e.g., cut or
killed by a wind storm event) more resources will become
available for survivors, and increased growth rates in sur-
vivors may be observed. These increases may be identi-
fied by subjective or objective methods. To subjectively
identify releases, students can study cores or plots of tree-
ring widths and identify the year(s) in which radial
growth rapidly increases (Figure 3).
Ring measurements and specific release criteria can
be used to objectively identify release events (Rubino &
McCarthy, in review). We recommend a simple technique
such as the “running mean” approach which compares
mean growth between two adjacent time periods. For
example, to test for a release in 1970, mean growth from
1960 to 1969 would be compared to mean growth from
1970 to 1979. If growth during the later 10-year period
was 50% greater than growth in the previous 10-year
period, you could assume that the tree released. Using
spreadsheet software to automate calculations will permit
rapid analysis of many tree cores. The analysis could be
taken further by having students list release years on the
board. Years with a large proportion of trees releasing
may indicate large scale disturbances such as selective
logging while years without high release synchronicity
may indicate gap disturbance in the forest. Checking local
aerial photos, logging records, or historical records for
large wind or ice storms can add an exciting, informative
dimension to the interpretation.
Evaluation
Implementation of the various tree-ring laboratories
in introductory level college classes has proven quite suc-
cessful, and the labs have received very positive respons-
es from students. We have found that students are quite
surprised to learn that they cannot only determine the
age of a tree by looking at its rings but also interpret past
growth conditions. We have found tree-ring laboratories
to provide a range of field, laboratory, and computer-
based experiences while focusing on scientific methodol-
ogy and hypothesis formulation and testing. Also, the lab-
oratories are of considerable benefit in bringing together
multiple concepts (secondary plant growth, plant stress,
competition) from various disciplines (botany, ecology,
forestry, geography, etc.) in a very practical way. Tree-ring
laboratories can be tailored to suit a wide variety of objec-
tives, approaches (e.g., inquiry-based learning), or limita-
tions (locale or budget). Furthermore, a relatively large
amount of web-based resources are available, permitting
students to explore various approaches to the discipline
and perform additional research for lab reports.
Class discussion and lab reports can greatly enhance
694 THE AMERICAN BIOLOGY TEACHER, VOLUME 64, NO. 9, NOVEMBER/DECEMBER 2002
these exercises by having the students critically examine composed of dead tissues, that the only living portion of
their results. For example, discussions of why two trees the stem was the thin cambium layer just under the
of the same species growing in close proximity show dif- bark, and that trees have evolved various wound
ferences in growth during a particular drought year responses to heal minor injuries such as coring holes.
have initiated meaningful discussions. Students hypoth- The possibility to introduce and provide entry-sites for
esized that differences in geology (proximity to a seep), pathogens can be mitigated by coring trees which can
genetics, disease, age of the trees, competitive ability, or compartmentalize or seal wounds quickly (e.g., oaks
amount of shading from neighbors could all lead to dif- and conifers) or by coring before annual cambial activi-
ferences. These answers show that the students try to ty has ceased (the trees will have time to heal during the
incorporate knowledge from numerous disciplines and same year they were cored). Furthermore, you can tell
topics to explain irregularities in the data. the students that borer-induced wounding is not much
Students are often concerned when they perform different than the wounds caused by sugaring maple
correlation analysis and find that their residuals and the trees. Finally, do not try to plug the hole or apply paint
climatic parameter of interest do not show a neat, clean or tar; studies have shown that this can be worse than
relationship (Figure 4). Assuring them that real-world simply allowing the tree to heal naturally on its own.
data sets are often messy does help alleviate some of
their concern. Also, explaining to the students that Conclusions
important (or significant) trends which are not perfect
(i.e., low P, but low r or R2 value) are still relevant. This exercise can be easily modified and tailored to
Asking the students where possible error or unex- locally available species or presented at various educa-
plained variance could have crept into the analysis often tional levels (high school to college courses, depending
leads to critical thought and discussion. Furthermore, upon the level of data analysis). We strongly recom-
explaining the difference between biological or ecologi- mend visiting our dendrochronology web page
cal significance and statistical significance can be useful. (McCarthy & Rubino, 2001) and The Ultimate Tree-
Ring Web Pages (Grissino-Mayer, 2001) before perform-
While performing the laboratory, we have also ing these exercises. You can find suppliers for all of your
observed that students rapidly utilize and develop their tree-coring needs, in-depth explanations of den-
botany skills. They quickly begin to understand the drochronology basics, and many useful links to answer
arrangement and structure of the various tissues pres- many questions you or your students may have.
ent in secondary xylem. Also, they begin to better
understand how biotic and abiotic parameters affect
plant function by relating factors such as climate to NOTE: All photo credits belong to the authors.
plant growth.
We have found that tree-ring laboratories serve as a References
springboard towards more advanced discussions of
Fritts, H.C. (1976). Tree Rings and Climate. New York:
ecology and prompt students to critically examine the Academic Press.
interaction between organisms and their growth rates
with the environment. We have successfully imple- Grissino-Mayer, H.D. (2001). The Ultimate Tree-Ring Web Pages.
URL: http://web.utk.edu/∼grissino/.
mented several ecological variants of the dendropeda-
gogy laboratory. Questions that we have employed McCarthy, B.C. & Rubino, D.L. (2001). Dendrochronology.
include the following (many more are possible): Do URL: http://www.plantbio.ohiou.edu/epb/instruct/ecol
trees growing on north-facing slopes grow faster or ogy/dendro.htm.
slower than those on south-facing slopes? Do canopy National Climatic Data Center & National Oceanic and
trees show different growth rates than understory trees? Atmospheric Administration. (2001). CLIMVIS
How do different species of trees growing in the same Information: Selection criteria for displaying period of record.
environment differ in growth pattern? Do trees of the URL: http://www.ncdc.noaa.gov/onlineprod/drought/
same species grow at different rates on ridges or in val- xmgrg3.html.
leys? Additionally, relationships between radial growth Rubino, D.L. & McCarthy, B.C. (2000). Dendroclimatological
and other climatic factors (mean growing season tem- analysis of white oak (Quercus alba L.) from an old-growth
perature or Palmer Drought Severity Index (a measure forest of southeastern Ohio, USA. Journal of the Torrey
of available water) can also be investigated. Botanical Society, 127(3), 240-250.
Students inevitably ask if coring kills or severely Rubino, D.L. & McCarthy, B.C. (In press). Comparative analy-
injures trees. In one laboratory, a student refused to take sis of dendroecological methods used to assess distur-
part in the coring process because she felt that we were bance events. (Dendrochronologia).
going to kill the tree. However, her fears were alleviated Stokes, M.A. & Smiley, T.L. (1968). An Introduction to Tree-Ring
when we explained that nearly the entire tree trunk is Dating. Tucson: The University of Arizona Press.
TREE-RING STUDIES 695