Out of doors biology is a fine source of data but it is difficult to recognise which of the many possible studies will be worthwhile, rewarding and manageable, especially for those who are inexperienced. One such study is described here and its results are presented and analysed. It uses a lawn that is adjacent to some tall shrubbery and was devised for a class of ten fourth form pupils of mixed CSE and 0 level ability. For larger classes pupils could work in pairs doing the field work.

The class had reached the point in their Environmental Science course where the next question to explore concerned the distribution of plants in lawns as a particular case of the distribution of plants in general, and with a view later to a study of the effect of man upon plant distribution. It was proposed (by the teacher) that the distribution of weeds was an essentially chance affair and for that reason lawn weeds would be expected to be distributed at random throughout the lawn. While the pupils agreed this was likely to be true, as scientists they had to check it was so before proceeding In this way the class became involved in this study—to confirm the hypothesis that weeds are distributed at random in well established lawns.

Twelve quadrats, each enclosing a square of 25 x 25cm, were laid in a straight row at right angles to the shrubbery and spaced at equal intervals on a lawn near school such that at one end of the row the quadrats were on the open lawn and at the other they were close to, even under, shrubbery. Each quadrat was numbered and each pupil was given a chart (Figure 1) and the task of marking with a cross in the appropriate column the species present in each quadrat.
 
 
 

Plant	Quadrat
	1	2	3	4	5	6	7	8	9	10	11	12
Plantain1
Clover
Moss
Buttercup
Dandelion
Grass
Dock
Daisy
Plantain II
Tree seedling

It is important that the pupils all work from copies of the same list, even if this requires the quadrats to be in place before the lesson begins and the species list compiled randomised and duplicated after the quadrats are in place. Identifying the weeds is not a problem. Most are familiar and are easily recognised and named. Those that are not can be identified from drawings in such books as those by Sutton’s and by Hessayon. Coded reference specimens can be used if necessary, each in a clear plastic bag and either named or given a code number.

For this particular study attention is confined to a strip of the lawn, itself sampled at regular intervals so that the area under investigation is restricted to a manageable size, and concentrated upon sites that are well spread but fixed in number. By having all 12 quadrats in place at once, and numbered, pupils may work in any sequence and so there are no bottlenecks. In addition, with the species list serving as a check-list, all observations concerning the presence/absence of each named weed is structured. The overall effect of these arrangements is that the pupils results are identical. This can be verified and discrepancies checked before everyone returns to the classroom.

The results which are shown in Figure 2 were obtained by the class and were analysed in an essentially pictorial fashion.
 

Plant

Quadrat

1

2

3

4

5

6

7

8

9

10

11

12

Plantain1

x

x

Clover

x

x

x

x

x

Moss

x

x

x

x

x

x

x

x

Buttercup

x

x

x

x

Dandelion

x

x

x

x

x

x

x

Grass

x

x

x

x

x

x

x

x

Dock

x

Daisy

x

x

x

Plantain II

x

Tree seedling

x

x

x

 

The fourth formers (Grade 9) were reminded that the distribution of the weeds on the lawn was now represented by the distribution of the crosses on their record sheets. The original question had become Are the crosses on the record sheets distributed at random? The answer was found by making up a new record sheet with the same number of crosses but this time distributing them according to the dictates of a table of random numbers. The two sets of results were compared (Figure 3) and the class decided there was a difference. Since one set was of crosses that were distributed at random, it followed that in the other set they were not.
 

We searched for ways to summarise the differences, for instance, by counting the number of crosses in the rows and the columns and comparing totals for each table and between the tables, and found the most convincing way was to divide the table into quarters and count the crosses in each quarter (Figure 4). The original hypothesis that the weeds on the lawn were distributed at random was rejected on the evidence that was best summarised in Figure 4. Since it is a rule of science that any non-random distribution has some underlying cause, it follows that there are reasons why the weeds are where they are, and that is something which the pupils could pursue.
 

	Results from Class				Results from Random Number table
	Columns		No. of crosses		Columns		No. of crosses
	1-6	7-12			1-6	7-12
Observed (O)	29	13	42		21	21	42
Expected (E)	21	21	42		21	21	42
O-E	8	-8			0	0
(O-E)2/E	3.05	3.05

chi squared value 6.10

Meanwhile, sixth formers (grade 11) could examine why the list needs to be randomised before being duplicated and analyse the results more formally, reducing the raw data straight away to a table and testing the Null Hypothesis with the Chisquared test. If the weeds are distributed at random then half the crosses are to be expected on each side of a line that separates columns 1 to 6 from columns 7 to 12, and indeed this is the case for the results obtained from the random number table (Figure 4b). However, the results obtained from the weeds on the lawn give a Chisquared value of 61 (Figure 4a) and this is significant at the 0.05 level. The Null Hypothesis is rejected. The weeds on the lawn are not distributed at random.
 
 

Keele University

Know your turf weeds, Sutton’s Seeds, Torquay.

D. G. Hessayon, Be your own lawn expert, Pan Britannica Industries Ltd.

Back to top

Back to contents of The Best of Teaching Statistics
Back to main Teaching Statistics page