Brass wheel of the Great Tide-Gauge at Bristol harbor ground

The brass wheel of the Great Tide-Gauge at Bristol harbor ground to a halt on the afternoon of June 12, 1834, because a single piece of seaweed had choked the float-wire. This was a clean failure, a terminal cessation of data that left the copper-plate record as smooth and uninformative as a blank sheet of foolscap. Such a failure possesses a certain aesthetic integrity, as the instrument refuses to lie when it can no longer function. Conversely, the observations I gathered from the amateur coastal observers during the Great Tide Week of 1835 constitute what we must recognize as a messy success. These men, often untrained and working by flickering lanterns, produced a thicket of numbers that seemed, at first glance, a forest of contradictions and clerical blunders.

What the local pilots call a “high sea” was, before this systematic collation, merely a subjective impression of danger. By forcing these disparate, jagged reports into a single analytical framework, the fundamental law of the semi-diurnal inequality began to emerge from the noise. This is the consilience of inductions: the moment when the messy success of the tidal charts predicts the behavior of the moon’s declination in a way that no single, perfect gauge at Bristol ever could. The Bristol gauge was a closed system, whereas the coastal reports were an open network. One might prefer the elegance of the machine, but the machine only confirms what is already known. The messy data set, plagued by human error and atmospheric interference, contained the very anomalies required to prove that the Atlantic is a single, oscillating basin.

A clean failure like the Bristol stoppage is a dead end in the history of science. It provides no friction against which a theory may be tested. The messy success, however, is a fertile ground for the naming of new phenomena. What we now call the cotidal line was, before this analysis, merely a series of disconnected delays. I have decided to retain the outliers from the Irish stations in the final draft. They do not fit the curve, but they suggest a secondary wave-interference that my current theory did not anticipate. I will mark these points in red ink to ensure they are not mistaken for errors.