This volume opened with planes that did not come back and left no holes to count. It closes with storms that nobody saw and so left no record. The question has the same shape: is the thin early record a thin early climate, or a thin early watch? The file can sharpen the question. It cannot answer it, and this chapter says so rather than pick the answer it prefers.
The index fixed the observable before this chapter existed: storms per decade, 1851–2025, from the held file. The script counts systems that reached tropical-storm strength (status TS or HU at any point), which is 1,752 of the file's 1,988 COMPUTED.
| decade | storms | |
|---|---|---|
| 1850s | 53 | 1851–59, nine seasons |
| 1860s | 72 | |
| 1870s | 75 | |
| 1880s | 89 | |
| 1890s | 82 | |
| 1900s | 84 | |
| 1910s | 61 | |
| 1920s | 71 | |
| 1930s | 115 | |
| 1940s | 104 | |
| 1950s | 128 | |
| 1960s | 113 | |
| 1970s | 96 | |
| 1980s | 90 | |
| 1990s | 108 | |
| 2000s | 147 | |
| 2010s | 152 | |
| 2020s | 112 | 2020–25, six seasons |
Per season, 1851–1899 averages 7.57 storms and 1970–2019 averages 11.86, a ratio of 1.57 COMPUTED. The early record is thinner. That much is a fact about the file. The next step is the usual mistake: reading it as a fact about the Atlantic.
The climate story: there were fewer storms in the nineteenth century, and the rise is real. The counting story: there were as many storms, but a storm in 1860 was recorded only if a ship crossed its path or it came ashore where someone was watching. Storms that lived and died in empty ocean left no entry, and the rise is the watch getting better.
This is Chapter 1's problem exactly. The record is the returners. A storm that nobody met is a plane that did not come back: it existed and it left no mark. Wald could recover his missing planes because he knew how many went out. Nobody knows how many storms went out in 1860, so the equivalent of his N is the missing number.
While writing the script, one further split suggested itself. It was not registered in advance, and by Chapter 2's rule it carries less weight than §1. If the early deficit were mainly storms that stayed at sea, then storms that came ashore — which people on land would see — should show a smaller rise than storms that did not. The file codes landfall with an L record, so the script splits on it:
| storms | 1851–1899 per season | 1970–2019 per season | ratio |
|---|---|---|---|
| with a landfall record | 3.08 | 5.02 | 1.63 |
| without one | 4.49 | 6.84 | 1.52 |
The simple account predicts the second ratio is the larger. It is not COMPUTED. Landfalling storms rose at least as much as the rest. This is the honest place to stop, not to switch sides. Coastlines in 1860 were also thinly watched, so a storm that made landfall on an empty shore could be missed too. And how a landfall flag came to be attached to an 1860 storm is not documented anywhere in the held file. The split does not confirm the counting story. It also does not refute it. What it does show is that the easy version of that story is not written in the data.
Separating climate from counting needs an independent record of the watch: how many ships crossed which parts of the Atlantic in which years, when aircraft reconnaissance began, when satellites began to see every storm. With that, one could estimate, decade by decade, the chance that a storm of a given track would have been seen at all. That estimate is Wald's q for the ocean, and dividing by it would recover the missing storms.
No such source is held. This volume's rule is that no chapter claims more than its sources carry, so the verdict is OPEN, and the gap is recorded as a want rather than filled with a figure from memory. When a source on the history of Atlantic observation arrives, it will be hashed into docs/floor-texts.tsv and this section rewritten from it.