Run a traverse, sum the latitudes and departures, and the total almost never comes back to exactly zero. That leftover — the closure error — has to go somewhere before you can call the traverse “closed” and start computing final coordinates. Three methods have dominated how surveyors have done that for the past two centuries: Bowditch's Rule (also called the Compass Rule), the Transit Rule, and the Crandall Method. They don't just differ in formula — they differ in what they assume about where your error actually came from, which is exactly why picking the wrong one for your equipment produces a technically-valid but poorly-justified adjustment.
Why a Traverse Needs Adjusting at All
Every course in a traverse has a latitude (its north-south component) and a departure (its east-west component), computed from the measured bearing and distance. In a perfectly closed traverse, the latitudes sum to zero and the departures sum to zero — you end up back where you started. In practice, small errors in every angle and distance measurement accumulate, and the sums come back slightly off. That leftover is the closure error, and adjustment is the process of nudging each course's latitude and departure just enough to force the sums back to zero — without throwing out any of the underlying field data.
The three methods below all solve that same problem. They just disagree about the right way to spread a fixed amount of error across a set of courses that vary in length and direction.
Bowditch's Rule (Compass Rule)
Bowditch's Rule is named for Nathaniel Bowditch, the American navigator and mathematician who developed it in the early 1800s for correcting ship's course plots — it was adopted into land surveying not long after. It assumes error accumulates in rough proportion to distance: a 500-foot course is assumed to carry roughly twice the error of a 250-foot course, regardless of direction. The correction to each course is therefore proportional to its length relative to the total perimeter.
This is the method most survey software defaults to, and the one most clients and reviewing agencies expect unless a project specifies otherwise. It's a reasonable assumption with modern total stations, where angular and linear precision are both high and roughly comparable across a job — which is most boundary and land surveys today.
The Transit Rule
The Transit Rule distributes error differently: instead of weighting by course length, it weights each course's correction by the size of its own latitude or departure relative to the sum of all latitudes (or departures) in the traverse. The assumption behind it is historical — in the era of transit-and-tape surveying, angles measured with a transit were considered substantially more reliable than distances measured by steel tape. The Transit Rule was built to protect the more-trusted angle data and load the correction onto the less-trusted linear measurements.
That assumption doesn't hold the same way with modern equipment. A total station's EDM typically measures distance to a precision that matches or exceeds its angular precision, so the original justification for the Transit Rule mostly no longer applies. It's still taught, and some agencies or historical retracement work still call for it specifically — but for a job run entirely on modern equipment, it's the method with the weakest technical case.
The Crandall Method
The Crandall Method takes a different approach entirely. Rather than distributing error proportionally by length or by component size, it holds the observed bearings fixed and solves a least-squares system to find the smallest possible adjustment to each course'slengththat closes the traverse exactly. Mathematically, it's a constrained optimization: minimize the total length correction while forcing the latitude and departure sums to zero.
Traditionally, Crandall's method starts one step earlier than Bowditch or Transit — with raw angle observations, balanced first across the traverse's angular misclosure, before the linear least-squares step runs. It's built on the premise that angles measured with a precise instrument are close to error-free, and that essentially all of the remaining misclosure is attributable to distance measurement. That makes it the right choice when you have real confidence in your angular observations and want a mathematically rigorous adjustment rather than a proportional estimate — control surveys and high-precision work are the typical use case.
| Method | Assumes | Distributes error by | Best fit today |
|---|---|---|---|
| Bowditch (Compass) | Angle and distance error are roughly comparable | Course length | General boundary and land surveys — the default |
| Transit | Angles far more reliable than distances | Size of each course's own lat/departure | Legacy transit-and-tape work, historical retracement |
| Crandall | Angles near error-free, distances less certain | Least-squares distance correction, bearings held fixed | Control surveys, high-precision work |
Which Should You Actually Use?
For nearly all everyday boundary, topographic, and construction survey work run on modern total stations or RTK equipment, Bowditch's Rule is the right default. It's what most software ships with, what most clients and reviewing agencies expect to see without explanation, and its underlying assumption — that error scales with distance — holds up reasonably well when your instrument measures angles and distances to comparable precision.
Reach for the Transit Rulemainly when a project specifically calls for it, or when you're retracing a historical survey and want your adjustment methodology to match the original. It's a legacy method more than an active recommendation for new work.
Reach for Crandallwhen the job is a control survey or otherwise demands a rigorous least-squares treatment, and when you have real confidence that your directional observations are significantly more reliable than your distance measurements — the exact condition it was designed to handle. It's also worth knowing that any tool implementing Crandall from bearings alone (rather than raw angle-and-distance field observations) is performing the linear least-squares step specifically, not the full traditional procedure — worth understanding before you cite it on a submittal.
None of these methods will save a traverse with a genuine blunder in it. If your closure ratio is outside acceptable tolerance — see our guide to verifying control before you trust it for the same principle applied to leveling — the fix is finding the bad observation, not picking a different adjustment method to paper over it.
Quick Reference
- Bowditch (Compass) Rule — correction proportional to course length; the standard default for modern equipment
- Transit Rule — correction proportional to each course's own latitude/departure; built for transit-and-tape era precision assumptions
- Crandall Method — least-squares distance correction with bearings held fixed; suited to control work with strong angular confidence
- Adjustment distributes existing error — it doesn't fix a blunder. A closure ratio outside tolerance means re-checking the observations, not choosing a different method