Closed Form for a Two-Term Recurrence
Summary: Euler’s Binet-type formula (§226–§229) for any recurrent series whose scale of the relation has length two. From the recurrence with first terms , the general term is where are the roots of . The pair satisfies the invariant , which lets each term be obtained from a single predecessor by an apparent — but illusory — square root.
Sources: chapter13, chapter17
Last updated: 2026-05-11
Setup (§226)
Let the scale have two members , so
(Euler now uses sign rather than ; both forms are equivalent up to the substitution .) The series arises from a rational function with denominator , factored as with and .
By partial fractions the closed-form general term is
Determining and (§226)
Setting and :
Solving,
Equivalently — the formula is symmetric in .
The invariant (§227)
From and , multiply:
Substitute and :
(source: chapter13, §227). This is the principal property of the recurrent series: it is a constant determined entirely by the scale and the first two terms , independent of which two consecutive terms one uses to compute it.
Term from a single predecessor (§227)
If is known, then satisfies and . Eliminating via the identity and solving:
Although this expression appears irrational, the right side is in fact always rational, since the series coefficients are themselves rational by construction (source: chapter13, §227). The square root must therefore evaluate to a rational every time. Euler does not prove this — he merely observes it.
Remote terms from two consecutive (§228–§229)
Given two successive terms and , Euler derives a closed form for the doubly indexed term :
(source: chapter13, §228). Eliminating the term in favor of an expression purely in and :
For :
(Section §229 sketches the same kind of formula for , by iterating the doubling.)
Worked Lucas example (§229 Example)
For the series — sum-of-two-previous, so scale , in the §226 sign convention — the first terms are , giving
(Adjusting signs back to Euler’s shows the identity gives — the standard Lucas-Fibonacci identity.)
Then
(positive sign for even , negative for odd). At , , so . ✓
Doubled-index formula: . At , : . ✓ (The 9th Lucas-like term in .)
Also and — Euler gives both.
Notable points
- The §227 identity is the determinant of a Casorati-like matrix. Modern phrasing: is the determinant of the matrix . Euler arrives at it by direct multiplication.
- The “irrationality is illusory” phenomenon. §227’s square-root formula is conceptually striking: a rational sequence is computable from one predecessor only via an apparently irrational operation, yet the irrational part always cancels. The same phenomenon recurs in modern treatments of Lucas sequences — e.g. Fibonacci’s is the same formula in different signs.
- The doubling formula generalizes. §229 derives both and from . Iterating gives “Lucas’s doubling” — modern fast Fibonacci computation in operations is exactly this scheme.
- §230 hints at the cubic generalization. For a three-member scale, the analogous formula is a cubic in the next term given the two predecessors — Euler writes the cubic but does not solve it. The pattern continues: a -member scale gives a degree- algebraic relation.
Why this matters
The two-term scale closed form is the Binet formula in full generality. It compresses an arbitrary linear two-term recurrence into a single algebraic expression, with a discriminant-like invariant () that controls when the sequence is rational vs. integral. Euler’s §227 identity is the source of the modern theory of Lucas sequences developed in the 19th century.
Dominant-root limit (Chapter 17)
Because with , the ratio of consecutive terms converges to :
(in the sign convention; equivalently, the larger root of ). This is the simplest instance of Daniel Bernoulli’s method: the Chapter 17 Example I uses with series and quotient , matching to six digits. The closed-form general term explains why the ratio converges, and how fast — the convergence is geometric with ratio .