The Elder Futhark runes are commonly believed to originate in the Old Italic scripts: either a North Italic variant (Etruscan or Raetic alphabets), or the Latin alphabet itself. Derivation from the Greek alphabet via Gothic contact to Byzantine Greek culture was a popular theory in the 19th century, but has been ruled out since the dating of the Vimose inscriptions to the 2nd century (whereas the Goths were in contact with Greek culture only from the early 3rd century). Conversely, the Greek-derived 4th-century Gothic alphabet does have two letters derived from runes,  (from Jer  j) and  (from Uruz  u).

The angular shapes of the runes, presumably an adaptation to the incision in wood or metal, are not a Germanic innovation, but a property that is shared with other early alphabets, including the Old Italic ones (compare, for example, the Duenos inscription). The 4th century BC Negau helmet inscription features a Germanic name, Harigastiz, in a North Etruscan alphabet, and may be a testimony of the earliest contact of Germanic speakers with alphabetic writing. Similarly, the Meldorf inscription of 50 may qualify as “proto-runic” use of the Latin alphabet by Germanic speakers. The Raetic “alphabet of Bolzano” in particular seems to fit the letter shapes well.^[4] The spearhead of Kovel, dated to 200 AD, sometimes advanced as evidence of a peculiar Gothic variant of the runic alphabet, bears an inscription tilarids that may in fact be in an Old Italic rather than a runic alphabet, running right to left with a T and a D closer to the Latin or Etruscan than to the Bolzano or runic alphabets. Perhaps an “eclectic” approach can yield the best results for the explanation of the origin of the runes: most shapes of the letters can be accounted for when deriving them from several distinct North Italic writing systems: The p rune has a parallel in the Camunic alphabet, while it has been argued that d derives from the shape of the letter san (= ś) in Lepontic where it seems to represent the sound /d/.^[5]

The g, a, f, i, t, m and l runes show no variation, and are generally accepted as identical to the Old Italic or Latin letters X, A, F, I, T, M and L, respectively. There is also wide agreement that the u, r, k, h, s, b and o runes respectively correspond directly to V, R, C, H, S, B and O.

The runes of uncertain derivation may either be original innovations, or adaptions of otherwise unneeded Latin letters. Odenstedt 1990, p. 163 suggests that all 22 Latin letters of the classical Latin alphabet (1st Century, ignoring marginalized K) were adapted (ᚦ from D, ᛉ from Y, ᛜ from Q, Ƿ from P, ᛃ from G, ᛇ from Z), with two runes (ᛈ and ᛞ) left over as original Germanic innovations. There are conflicting scholarly opinions regarding the ᛖ (from E ?), ᚾ (from N ?), ᚦ (D ? or Raetic Θ ?), Ƿ (Q or P ?), ᛇ and ᛉ (both from either Z or Latin Y ?), ᛜ (Q ?), and ᛞ runes.^[6]

Of the 24 runes in the classical futhark row attested from 400 (Kylver stone), ï, p^[a] and ŋ^[b] are unattested in the earliest inscriptions of c. 175 to 400, while e in this early period mostly takes a Π-shape, its M-shape () gaining prevalence only from the 5th century. Similarly, the s rune may have either three () or four () strokes (and more rarely five or more), and only from the 5th century does the variant with three strokes become prevalent.

The “mature” runes of the 6th to 8th centuries tend to have only three directions of strokes, the vertical and two diagonal directions. Early inscriptions also show horizontal strokes: these appear in the case of e (mentioned above), but also in t, l, ŋ and h.