Experiential learning using rapid-cycling Brassica campestris L.T. E. Michaels*

ABSTRACT

Experiential learning strategies let students discover conceptsrather than passively acquire and integrate knowledge. A semes-ter-long experiment was designed to promote experiential learn-ing in an undergraduate plant breeding course. The learningobjectives of the experiment were that students should be ableto (i) calculate and interpret heritability, (ii) create and use computer database for storing and manipulating their data, and(iii) feel confident in their ability to breed plants. By usingrapid-cycling Brassica campestris L. as the model organism,heritability of open flower number 4 wk after planting and othercharacteristics could be estimated by parent-offspring regres-sion. Pooling of student data resulted in significant heritabili-ty estimates for flower number (h2 = 0.49 P < 0.05 in 1989),whereas estimates obtained by students individually were gener-ally nonsignificant. Student responses to an end-of-cou ,rse evalu-ation indicated that the experiment was helpful in understandingprinciples of quantitative genetics, appreciating plant breeding,and maintaining interest in the laboratories during the semester.

T HE LEARNING PROCESS has been visualized by Lewin(1951), Kolb (1984), and Bawden (1985),

others, as a cycle where students alternate between con-crete experience and abstract conceptualization. Concreteexperience includes experimentation and observation ofexperimental results. Abstract conceptualization involvesreflection, hypothesis building, and designing new experi-ments to test hypotheses.

Experiential learning strategies differ from convention-al strategies in that problem-solving exercises give studentsthe chance to discover concepts rather than passivelyacquire and integrate knowledge. Lecture and tutorial sec-tions of a conventional biology or agriculture course typi-cally provide basic information. They may also exerciseinductive reasoning skills necessary for integration andabstract conceptualization. But experiments conductedby students in a laboratory section where the students areactive investigators may provide the concrete experiencethat is missing in lecture and tutorial sections.

I teach a plant breeding course to biological and agri-cultural science students who are in their third or fourthyear of study. The course format is traditional, with three1-h lectures and one 2-h laboratory each week. I designeda laboratory that I hoped would provide the experiencesthat, coupled with information from the lectures, wouldallow students to pursue the cyclical learning processvisualized earlier. The key component of the laboratorywas a semester-long quantitative genetic e.xperiment where

Department of Crop Science, Univ. of Guelph, Guelph,ON, CanadaNIG 2Wl. Received 30 May 1989. *Corresponding author.

Published in J. Agron. Educ. 19:41-44 (1990).

each student was responsible for building hypotheses;contributing to the experimental design; and gathering,analyzing, and interpreting data from his or her own setof plants. The learning objectives for the experiment werethat by the end of the semester the students should beable to (i) calculate and interpret heritability, (ii) createand use a computer database for storing and manipulat-ing their data, and (iii) feel confident in their ability breed plants.

Rapid-cycling Brassica campestris L. (Hawk andCrowder, 1978; Williams and Hill, 1984; Edwards andWilliams, 1987) was chosen as the model organism forthe experiment because (i) two generations of these plantscan be grown during a 13-wk semester, which allowsstudents to estimate heritability by regression of offspringon parent, (ii) the plants and seeds are of a reasonablesize for students to handle, and (iii) the plants resemblecanola (Brassica napus L.), an important oilseed crop inOntario, and may seem more relevant to agriculturalstudents.

In this article, I outline the structure of the semester-long experiment and evaluate its effectiveness in (i) help-ing students understand principles of quantitativegenetics, (ii) helping students understand and appreciatethe integration of disciplines involved in plant breeding,and (iii) maintaining student interest in the laboratory sec-tion throughout the semester. General performance ofthe rapid cycling stocks under our growth conditions isalso reported.

METHODS

Experiment Structure

Week 1. Each student filled two nine-cell multipackswith Metromix (Horticultural Products, Ajax, ON),sowed three seeds of rapid-cycling Brassica campestris L.stock CrGC-1 (Williams and Hill, 1984) in each cell, andcovered the seeds with vermiculite. The pack were placedin light facilities consisting of growth racks in 1987, agrowth room in 1988, and a growth chamber in 1989. Thegrowth racks used in 1987 were equipped with residential-type light fixtures with 2, 40-W fluorescent tubes and 2,25-W incandescent bulbs. The growth room and growthchamber used in subsequent years were equipped with acombination of incandescent and fluorescent lights thatsupplied a photon flux density of approximately 300#mol/m2 per s for a 16-h day (1988) or continuous light(1989), and were maintained at about 24°C.

Week 2. Students thinned seedlings to one plant per cell.Week 4. Students counted the number of open flow-

ers 4 wk after planting on each of their 18 plants and alsomeasured three additional characteristics of their own

J. Agron. Educ., Vol. 19, no. 1, 1990 41

choice. I suggested that they choose characteristics thatappeared to show plant-to-plant variation. Number ofopen flowers was measured by all students because it waseasy to determine unambiguously and was heritable inpreliminary experiments (data not shown). In addition,having one characteristic common to all students providedthe opportunity for students to analyze pooled data forthis characteristic; this eased marking of the final reportsby making some answers uniform from student and, byincreasing the sample size, increased the likelihood of sig-nificant heritability for that characteristic. Popularchoices for the additional characteristics included plantheight, stem diameter, leaf area, and branch number.Each student identified and tagged the superior plantwithin their 18 plants for each characteristic, and thepacks were returned to the lighting facility. The studentsentered their data into a spreadsheet-type computer data-base. The database consisted of a line for each plant, adesignation of the location of each plant within the nine-cell packs, individual plant data for each of the charac-teristics measured by the plant, and an indication as towhether the plan had been selected. The same databasesoftware was not used by all students because some ofthe students had their own computers and software. Therapid-cycling Brassica campestris L. used in this study-was self-incompatible. Therefore, feather dusters in 1987and 1989 and leaf cutter bees (Megachile rotundata) in1988 were used to promote random mating among theclass’ plants.

Week 8. Students harvested seeds from each of theirparent plants and sowed offspring seeds in new multi-packs. Three seeds from each parent plant were used tosow one offspring cell. The students also sowed an addi-tional three cells with the seeds from each of the selectedplants.

Week 9. Students thinned seedlings to one plant per cell.

Week 11. Students evaluated offspring plants for openflowers and the same additional characteristics that wereevaluated for the parent plants, and added these new datato their computer databases.

Student Analyses

The students pooled their data for number of openflowers in 1987 and 1989, but relied only on data fromtheir own 18 plants in 1988. They regressed offspringvalues on parent values to obtain an estimate of half ofnarrow sense heritability. The students estimated realizedheritability by calculating the selection response/selectiondifferential ratio based on the performance of theirselected plants. Although the exercise focused on herita-bility, students were also asked to compare a number ofmeans using t tests and to predict gain from selectiongiven a hypothetical selection intensity. The statisticaltechniques used in the laboratory were reviewed in theintervening laboratory sessions. The students were en-couraged, but not required, to do these analyses on thecomputer. Finally, the students prepared and submitteda written report of their experimental results, a discus-

42 J. Agron. Educ., Vol. 19, no. 1, 1990

Table 1. Responses to course evaluations completed by studentsat the end of the semester.

Year

Statement type 1987 1988 1989 Mean

Specific statements regarding semester-long experiment

The laboratories were very helpful in under- 3.5~ 4.4 4.1 4.0standing principles of quantitative genetics

The hands-on aspect of the laboratories was 3.9 4.7 4.1 4.2important to understanding and appreciat-ing plant breeding

Having a series of interrelated sessions helped 4.1 4.4me maintain a greater interest in laborato-ries throughout the semester

Mean 3.8

Mean~

4.8 4.3

4.6 4.2

GenerNstatementsregardingcourse and~st~ctor

4.4 4.7 4.3

Mean of responses on a 1 {complete disagreement) to 5 {completeagreement} scale, n = 55, 35, and 28 students in 1987, 1988, and 1989,respectively.Mean of responses to 17 positively worded statements.

sion of their findings, and suggestions for subsequentexperiments.

At the end of the course the students were asked toevaluate the semester-long experiment by responding tothree statements (Table 1). These statements were ap-pended to the standard end-of-course questionnaire. Thestudents responded using a 1 (complete disagreement) 5 (complete agreement) scale.

RESULTS AND DISCUSSION

Laboratory Results

Lighting facilities influenced growth of the rapid-cycling Brassica stock. The growth racks resulted in plantsthat grew about 20 mm horizontally along the surface ofthe multipacks before beginning to grow vertically. Thetotal stem length was about 350 mm. Seed set was poorand the plants lodged. Both the growth room (1988) andgrowth chamber (1989) resulted in sturdy stems of ap-proximately 3 mm diam. at the base and upright plantsstanding approximately 300 mm high. Seed set was muchhigher than in 1987. Regardless of lighting facilities, theplants began flowering in approximately 18 d.

Narrow sense heritability of flower number based onpooled 1987 data was 0.59, but was not significantlydifferent from zero at the 5°70 probability level. Realizedheritabilities were not calculated in 1987. In 1988, herita-bilities for flower number based on data from individualstudents ranged from -1.29 to 0.70 and were not sig-nificantly different from zero. Realized heritabilitiesranged from -1.57 to 1.05. In 1989, heritability forflower number based on pooled data was 0.49 and wassignificantly different from zero at the 5070 probabilitylevel. The realized heritability based on pooled data was0.09.

Pooling class data, as in 1987 and 1989, seemed to con-tribute to calculation of positive heritability values forflower number, whereas heritability estimates in 1988based on individual student’s samples of 18 or fewer data

points were quite variable from student to student, andwere often negative. These variable results in 1988 wereunsatisfactory from a teaching perspective because theconcept of heritability was not uniformly reinforced forall students. Although some students recognized that thevariable heritability and poor plant growth were primar-ily technical problems associated with growing conditionsor experimental design, others were concerned that thenegative or nonsignificant values in some way were theresult of their own mistake or lack of skill as a plantbreeder. This latter perception was obviously verydetrimental to morale and particularly threatened theachievement of the third learning objective for the labora-tory: confidence in their ability to breed plants. In an ef-fort to build student morale and confidence, more lectureand laboratory time was spent discussing experimentalproblems in 1988 than in the other 2 yr. Although I didnot want the results of the experiment to be completelypredictable, it became apparent that the experimentalprocedure should be refined so that the results meet atleast some of the students' expectations. The significantheritability estimate in 1989 may have been associatedwith the higher-quality lighting and climate control sup-plied in the growth chamber and pooling of the students'data. Only 15% of the students in 1989 reported signifi-cant regressions when they used their own data only.

All students entered their data in a computer databaseand included the database as part of their final report.All students also did at least some of their analyses usingthe functions included in the software.

Student Perceptions

Students on average agreed with the positively wordedevaluation statements specifically relating to the semester-long experiment. They also agreed with statements re-garding the course in general (Table 1). According to thefirst statement, students perceived that the laboratoryhelped them better understand quantitative genetics. Thismay be associated with the hands-on nature of the labora-tory as evaluated by the second statement. I observed thatthe students seemed to enjoy growing and evaluating theirown plants, and seemed more committed to interpretingthe data they obtained from their plants over the courseof the semester than to interpreting data presened to themin homework assignments. According to the second ques-tion, the students perceived that the semester-long experi-ment helped them understand and appreciate plantbreeding. Student agreement was highest in all years withthe third statement regarding maintenance of interestthroughout the semester. This agrees with my observa-tion that absenteeism from laboratories was lower thanin plant breeding courses I taught in 1985 and 1986.

The mean of responses to the general course evalua-tion statements in each year are a baseline index of stu-dent satisfaction. The mean of the experiment-specificstatements can be compared to the baseline. The differ-ence between responses to general and experiment-specificstatements was least in 1988, which could be interpretedto mean that student satisfaction with the experiment was

highest in that year. This was the year in which moraleinitially appeared to decline. Additional efforts to dis-cuss problems with students may have resulted in thegreater level of overall satisfaction with the experimentthan in other years. It was also the year, however, inwhich the class data were not pooled. The students mayhave been more satisfied with the experiment because theyrelied on their own data. The difference was greatest in1987, when, as a result of the lighting facilities, plantgrowth was the poorest. The lodging and intertwining ofthe plants resulted in shattering and loss of the few seedsthat had set as the dry, mature plants were separated priorto seed harvest. When students were asked to commenton the laboratory in the questionnaire, they often men-tioned that the rapid-cycling Brassica experiment pro-vided the opportunity and motivation to use and betterunderstand the statistical analysis skills that they learnedin intervening laboratory sessions and in previous courses.

Based on review of the laboratory reports that werereceived, the first two learning objectives were attained.The level of satisfaction with the specific and generalstatements suggested that the third objective may alsohave been achieved. The semester-long experiment at-tempted to provide concrete experiences that would allowstudents to discover concepts for themselves. Unfortun-ately, as the experimental procedure was refined over theyears, some of the opportunity for individual discoverymay have been lost as spontaneity, as well as the poten-tial for failure, declined. In retrospect, I realize that theexperiment should not be so predictable as to be only ademonstration of concepts, but should be sufficiently re-fined that the plants grow normally and the data arereasonable. Some frustration is good, but too muchreduces motivation.

Experiential learning by means of a student-run experi-ment that spans the semester might contribute to the over-all impact of other courses. The advantages are thatcomplex subjects can be investigated, the students havethe chance to discover concepts on their own, and stu-dent interest in the laboratory sessions is maintainedthroughout the semester. Also, the impact of interven-ing sessions is which other related skills are exercised canbe augmented if the students find that these new skillsare relevant and may help them interpret the results theyobtain in their main experiment. If the instructor designsa laboratory that includes concrete experiences, and alsoencourages abstract conceptualization in the course, thenstudents will be able to cycle through the process thatleads to effective learning.

J. Agron. Educ., Vol. 19, no. 1, 1990 43

Kolb, D.A. 1984. Experiential learning: Experience as the source of York.learning and development. Prentice-Hall, Englewood Cliffs, NJ Williams, P.H., and C.B. Hill. 1984. Rapid-cycling populations of Bras-

Lewin, K. 1951. Field theory in social sciences. Harper and Row, New sica. Science (Washington, DC) 232:1385-1389.