DATA: Baar et Ter Braak (1996) DIRECTORY: \CANOCO\SAMPLES\PERMUTIO\SPLITPLT PROBLEM: Ectomycorrhizal fungi occurrence as affected by the manipulation of litter and humus in Scots pine stands of different age ILLUSTRATION OF: + what a nested design or split-plot design is + what a whole plot variable is and what a split-plot variable + the requirements to be able to perform permutation tests in a split-plot design + how to test for whole-plot variables + two ways to test for split-plot variables + how to test the interaction between a split-plot variable and a whole-plot variable _______________________________________________________________________________ FILE TYPE | FILE NAME | DESCRIPTION -------------+---------------+------------------------------------------------- SPECIES | fungi.dta | numbers of sporocarps of 33 ectomychorrhizal | | fungi in 64 plots in 6 Scots pine stands | fungi2.dta | as fungi.dta with two extra variables (total | | numbers of sporocarps and of species in each | | sample). Added as supplementary species in | | fig1.con ENVIRONMENT | design.dta | 6 Stand and 3 treatment (S,A,C) indicators, Age | | and Soil type (Podzol, Arenosol) of the stands | nutrient.dta | nutrient concentrations and pH in the ectorganic | | layers (log-transformed). Used in fig1.con as | | supplementary environment file DERIVED | treatmnt.dta | as design.dta but without the stand indicators PROJECTS | stand1.con | project to show that age and soil are constant | | within stands(=whole plots) | stand2.con | idem by regression of Age and Soil on Stand | age_soil.con | test of the relation between the fungi and the | | two stand (=whole-plot) variables Age and Soil | | type in a balanced subset of 6*8 = 48 plots | | (all A-plots deleted) | age.con | idem, but now for Age adjusted for Soil type | lh.con | test of the effects of the treatment of the | | litter and humus layer (LH, a split-plot | | factor) on the fungi on all data, using block | | permutation within stands | lh_split.con | as above but on the balanced subset of the | | data, and using "within whole-plot" permutation | | of split-plots | lh_x_age.con | test of the interaction of LH treatment and age | | on all data | agetrial.con | attempt to test Age on all data (wrong P-value) | agsl_err.con | wrong test of Age and Soil (whole-plots not | | permuted) | fig1.con | project upon which fig.1 of the paper is based -------------+---------------+------------------------------------------------- Details of the data: In this example we study the effect of sod-cutting and sod-addition on sporocarps of ectomycorrhizal fungi in six Pine stands differing in age and soil type using data from an experiment by Baar (Ectomycorrhizal fungi of Scots Pine as affected by litter and humus, PhD dissertation, LUW, 1995) reported in Braar & ter Braak (1996). Experimental layout: 6 Pine stands: In 4 stands, three treatments control (C), sod-cutting (S), sod-addition (A), each treatment replicated 4 times, resulting in 12 plots (each of size 15x15 m2). in 2 stands (the oldest two) two treatments: Control (C) and sod-cutting (S), each treatment replicated 4 times, resulting in 8 plots. Species data: number of sporocarps of 40 ectomycorrhizal fungal species per 225 m*m plot (fungi.dta) Explanatory data: - experimental design factors: design.dta variable no. 1 St1_Aren: stand 1 of age 1.0 with Arenosol 2 St2_Podz: stand 2 of age 2.0 with Podzol 3 St3_Podz: stand 3 of age 3.0 with Podzol 4 St4_Aren: stand 4 of age 4.0 with Arenosol 5 St5_Aren: stand 5 of age 5.0 with Arenosol 6 St6_Aren: stand 6 of age 6.0 with Arenosol 7 S 8 A 9 C 10 Age (log-age) 11 11 Podzol 12 12 Arenosol - nutrients in ectorganic layers (humus in C and A, litter in S): nutrient.dta - stand indicator (dummy) variables, which might be used as covariables: stand.dta (stand 1-6) The sample names consist of the Stand number (Table 1 of the paper), the treatment (S, A, C) and the original plot number within stands (1-12). For example, the sample with name S2 A 3 is from stand S2 and was assigned the treatment A (sod-addition). Its original plot number was 3. Baar, J. & Ter Braak, C.J.F. (1996) Ectomycorrhizal sporocarp occurrence as affected by manipulation of litter and humus layers in Scots pine stands of different age. Applied Soil Ecology, 4: 61-73.