Spray calibration trial with a drifting boom
A student investigating spray efficiency on a 40 ha block of cotton sets three nozzle types (A, B, C) on a boom and measures deposition on water-sensitive cards at three positions across the boom: left, centre, right. Each treatment is run once per position. Results: Nozzle A gives mean deposition of 62, 60, 63 (arbitrary units) across the three positions; Nozzle B gives 41, 78, 44; Nozzle C gives 55, 54, 56. Wind speed during the trial is recorded as 12 km/h gusting to 24 km/h, and the boom height is 0.5 m. The student concludes Nozzle A is the most efficient because it has the highest mean (61.7) and lowest spread, and writes this as a firm recommendation for the grower.
1. Evaluate the validity of the student's conclusion that Nozzle A is the most efficient nozzle.
Read the worked answer
The conclusion is not yet valid for three reasons.
No replication. Each nozzle is tested once per position, so the 'mean' is an average of three different boom positions, not of repeated trials. Position and nozzle are confounded: Nozzle B's high centre value (78) could reflect a localised blockage or a gust, not the nozzle itself. Without replicate runs at each position, there is no estimate of random error and no basis for claiming a real difference.
Uncontrolled wind. Wind of 12 km/h gusting to 24 km/h exceeds typical fine-spray drift thresholds and varies during the trial. If the gusts occurred during particular runs, drift loss - not nozzle efficiency - could explain the low values for B at left and right. Deposition on cards measures what arrived, not what was applied, so drift must be measured (e.g. cards downwind) or the trial repeated under calmer, recorded conditions.
Efficiency is undefined. The student equates 'highest mean deposition' with 'most efficient'. Efficiency usually means target deposition per unit applied, or reduced off-target loss. A nozzle could deposit more on the target simply by producing coarser droplets with more total volume. The student needs to state the operational definition and measure both target and off-target deposition.
A firm recommendation to a grower is therefore premature. The data are a pilot and should be described as indicative only, with replication, wind monitoring and a clear efficiency metric before any recommendation.
2. Design a revised trial that would let the student make a defensible comparison between the three nozzles.
Read the worked answer
A defensible design would include:
- Replication: at least three independent runs per nozzle per boom position (9 runs per nozzle), with nozzle order randomised each run to avoid order effects such as wind shifting during the day.
- Blocking: treat each run as a block across the three nozzles so that changing wind and temperature affect all nozzles similarly, then compare within blocks.
- Standardised conditions: record wind speed and direction continuously; restrict runs to a pre-set wind window (e.g. below 10 km/h) and note any run outside it. Keep boom height, speed, pressure and spray volume identical across nozzles.
- Measurement: water-sensitive cards at multiple heights in the canopy plus downwind drift cards, so both target deposition and off-target loss are quantified. Report mean and standard deviation per nozzle.
- Operational definition: state efficiency as target deposition divided by volume applied, or as the ratio of target to off-target deposition.
- Analysis: compare means with an appropriate test (e.g. ANOVA with replication) rather than eyeballing averages, and report variability.
This design separates nozzle effects from position, wind and order effects, and gives the grower a recommendation backed by replicated, controlled evidence rather than a single unreplicated run.