Data for the tobacco mosaic virus (TMV) experiment described in McIntyre (1955).

The data is available in the following files that differ in their formats:

An R script that produces the anatomy for the design used is available in the following file:

McIntyre (1955) reports an investigation of the effect of four light intensities on the synthesis of tobacco mosaic virus in leaves of tobacco Nicotiana tabacum var. Hickory Pryor. It is a two-phase experiment: the first phase is a treatment phase, in which the four light treatments are randomized to the tobacco leaves, and the second phase is an assay phase, in which the tobacco leaves are randomized to the half-leaves of assay plants.

In the first phase, four successive leaves at defined positions on the stem were taken from each of eight plants of comparable age and vigour that had been inoculated with the virus. Arbitrarily grouping the plants into two sets of four, the four treatments were applied to the leaves, which had been separated from the plants and were sustained by flotation on distilled water, in a Latin square design for each set with tobacco plants as columns and leaf positions as rows.

In the second phase, virus content of each tobacco leaf was assayed by expressing sap and inoculating half leaves of the assay plants, \emph{Datura stramonium}, on which countable lesions would appear. Lots of eight sap samples were formed from pairs of tobacco plants, the pairs being comprised of a plant from each set in the treatment phase. The eight samples from a lot were assigned to four assay plants using one of four 4 x 4 Graeco-Latin square designs, with the leaves from a single tobacco plant assigned using one of the alphabets and the second tobacco plant using the other. Actually, this design is a semi-Latin square.

The square root of the counts of lesions on the assay plants, and the factors indexing them, are given in the files. The factors are as follows:

  1. Set
  2. HalfLeaf
  3. NicPlant
  4. Posn
  5. Lot
  6. DatPlant
  7. AssPosn
  8. Treat
  9. followed by the response variable
  10. SqrtCount