In summary, my hypothesis is that certain translators transliterated the Greek letters but not the accents, and thereby lost the [h], because initial [h] is only a small mark above the vowel. They may have done so because their Greek was not very good, or because they did not consider the [h] mark a real letter worth transliterating.
The Greek word εὕρηκα was pronounced [hěu̯rɛːka] in classical Greek; the [h] is indicated by the spiritus asper ("rough breathing"), the tiny "c" above the ὑ. (The other accent, the acutus, indicates increasing pitch; note that the spiritus asper came from a tiny left half of the Greek letter H, which was used in some older dialects to denote an [h] sound.)
Many, perhaps most Greek words came to English through Latin. However, the Latin transliteration of εὕρηκα would be heureca, or possibly eureca. My own, quick research in the Bibliotheca Teubneriana Latina suggests that the great majority of Roman writers transcribed the spiritus asper as h, but a significant minority did not. In any case, neither heureca nor eureca exists in the BTL; therefore I suspect that the word was not the famous expression in Rome that it is now.
The earliest mention of Archimedes' exclamation is by Vitruvius (De Architectura IX, preface 10), and later usage is most probably all based on him. Both the Perseus Digital Library, whose edition of Vitrivius is based on the 1912 Teubner, and the edition by C. Fensterbusch (Darmstadt 1964/76) give the word in Greek letters. Perseus/Teubner give the Greek with accents, Fensterbusch without.
The first use of the word recorded by the Oxford English Dictionary is from 1570:
[1570 Dee Math. Pref., For this, may I (with ioy) say EYRHKA.]
Notice that the word is given in Greek capitals, without accents.
(The square brackets may indicate either something added by editors in a quotation (unlikely, because that only applies to parts of quoted sentences), or a "spurious" entry in the dictionary, which is a word that erroneously found its way into other dictionaries while in fact it was never used (just as unlikely). I conclude that the brackets either indicate nothing of relevance or that this is a dubious quotation.)
1603 Holland Plutarch's Mor. 590 [Archimedes] crying out, Heureca.
1658 tr. Porta's Nat. Mag. xviii. viii. 384 We have gone beyond Archimedes his Eureka.
1742 Fielding J. Andrews ii. xiii. (ed. 2) 267 Adams...returned overjoyed...crying out ‘Eureka’ [ed. 1 (1742) Ευρηκα; ed. 3 (1743) Heureka.]
By the 19th century, eureka seems to have prevailed. Before that, as it can be seen, various versions were used. Several editions did not print the breathing mark: perhaps this caused readers who were not well versed in Greek to translate the Greek letters one by one into the most similar English letters, disregarding conventions; this would give eureka. A commenter on Languagehat gives the following for the editions of Fielding also quoted in the OED:
- 1st edition (1742), Ευρηκα
- 2nd edition (1742), Eureka
- 3rd edition (1743), Heureka
This might suggest that the spelling eureka did indeed originate in a letter-by-letter transliteration of the Greek into similar English letters by someone who didn't know the traditional convention of using c for Greek κ: Fielding feared that his version Eureka might seem illiterate and changed it into Heureka in the third edition.
In the time of Vitruvius (1st century BC), the spiritus was most probably not yet in use, so that plain unaccented Greek letters would have been written. It would have been capitals, because Greek minuscules were probably not yet used either. (Note that we cannot be sure that the Greek wasn't added later by monks and that Vitruvius himself didn't use Latin letters.) In the edition by Fensterbusch mentioned above, the Greek letters printed without spiritus too. This might support my theory. But we would really need to compare many more editions from various dates.
So nothing is certain. Before the 19th century, transliterations of Greek words that were not used in Latin varied wildly—if not of other Greek words, then certainly of this one. The reason why it eventually became eureka in modern English, Dutch, and French, as opposed to heureka in German, could be any arbitrary circumstance, though I have a feeling that the German version is relatively new as compared to ours. In the edition of De Architectura below, printed in 1567 in Venice, the word is written in Latin letters as eurica:
This shows how varied transliterations were in the 16th century. The i can probably be explained by the fact that it was printed in Italy. But here is the Latin edition of Giambattista della Porta's Magia Naturalis from 1597 (the English version is in the OED's quotation from 1658 above):
It was printed in Frankfurt, and eureka is given in Greek, with spiritus. Another Latin edition of the same work, printed in Naples in 1589, uses the aspirated Greek too (scanned image no. 302 / printed page 285, at the bottom).
Best Answer
A spell is the opposite of a fixed period of time— (MW)
If my elderly neighbor asks me to come sit for a spell, I don't know whether she intends to keep me for two minutes or two hours.
As you know from the OED, this meaning derives from an old verb sense of spell meaning
which is said to persist in Australian and U.S. usage. This, in turn, is traced to spele, a rare or dialectical word meaning to take or stand in the place of (another); to represent.
This in turn is from Old English of obscure origin, but believed to be related to gespelia and spala; a spale was a sparing, respite or rest.
The use of spell to refer to a period of weather is attested from 1728—
— and paired with adjectives from 1740 in the same source.
While dry spell comes late, Thomas Jefferson did write in 1797
The assignment of spell to a fixed period is almost certainly a modern invention for use within a specific domain, just as geologists began to use the term eon (æon) to refer specifically to a period of a billion years, while physicists adopted terms like shake and jiffy for very short times. In general usage, those terms remain inexact, and only represent fixed values when used as terms of art. It helps avoid prosaic terminology like heating degree days or survey townships, even as we abandon colorful measurements like the oxgang or the fistmele.