Draw where rain travels on a sloping model

Build a low relief from waterproof modelling material. Set the base in a tray and place a ridge that runs across the model.

Draw a crest and two low outlets in a small model made in a waterproof tray. Write the independent change at the start of each row and keep the outcome beside it. Include anything about a failed run that would matter to somebody repeating the setup. If you later make a neat graph, keep the rough record as well so it remains possible to inspect the reasons for leaving out a point.

Two nearby drips can drain in different directions where a small ridge separates them. Make a simple model of what you believe happens between the starting and final positions. Then check one point at which that model gives a visible consequence. A single carefully observed intermediate position can be more revealing than comparing two widely separated endpoints with an unknown change between them.

Add water in small amounts at different high points. Note which drops pass into the same channel and where pooling interrupts an expected route.

A tiny trench left by a previous run may capture a later stream. Put the observations into order and keep the raw measurements or drawings. A result can be useful at more than one level. The direction of a change may be unmistakable while the exact size is uncertain. Describe the strongest pattern the records support and qualify measurements that depend on a faint boundary or a moving object.

Smooth accidental tracks before adding water at the same point. If the results vary, look for a second feature that moved along with the one you meant to change. That can be the distance to an observer, the angle of a support or the way the material moves between runs. Isolate it in a small follow-up trial when possible.

Sketch how the inferred divide shifts when you adjust the saddle. Write one prediction alongside the setup. It helps to describe what would be observed if that prediction were right, including a direction, a clearer pattern or a measurable difference. This makes room for the outcome to disagree with what you expected and gives you a better question for a second run.

Write a row for each run and keep apparently odd results in the record. Mark things that went wrong, such as a shifted surface, uneven starting point or missed time, before judging the pattern. Changing how you record the behaviour halfway through can make two otherwise similar trials difficult to compare. Explain the outcome in a sentence that a reader can check against the figures or sketch. Use a second sentence for a proposed cause. If you can imagine a different cause that fits the same observations, describe a small test that would help distinguish the two. Comparing those predictions can make a follow-up clearer than simply repeating the same demonstration. On a second trial, add water at one known point upstream of the divide. A channel that appears at the bottom may have a different source than you guessed. Sketch your apparatus before you start and use the sketch to identify where each observation will be made. Write down the starting state before changing it. If another person repeats the trial, knowing the relative positions can matter more than the brand or appearance of the materials.

Move the saddle slightly and redraw the boundary between the two drainage areas. Decide how to mark an observation when two readings fall too close together to separate. Even spacing on a sheet is useful only to the extent you can see where a feature crosses each line. A visible range can carry more meaning than a single estimated digit that your instrument cannot resolve.