As a rule, transverse (perpendicular to the ridge pole) roof beams (koya-bari 小屋梁) in minka are spaced at two to three metre centres. Transverse beams are always present in minka as they are necessary to prevent the spreading of the longitudinal walls under the thrust action of the rafters, but there may also be beams that run longitudinally (parallel to the ridge pole) where it makes structural or economic sense to do so, such as near the gable ends of the building, where the span from an internal post on the structure’s longitudinal centreline to the gable wall may be shorter than that of the transverse span from the post to the longitudinal walls, but not long enough to justify adding another transverse beam. Framing systems which incorporate longitudinal beams are called ‘cross framing’ (jūji-gumi 十字組 or igeta-gumi 井桁組). Both terms derive from the visual resemblance of the crossed beams to the respective kanji in their names: ‘十’ (ten) and ‘井’ (water well).
In the simplest structures, a single transverse beam spans the full distance between the longitudinal walls, but in all but the smallest houses the more common practice was to use a central longitudinal beam ji-mune (地棟, ‘ground ridgepole’), supported by an intermediate post or posts, and support the transverse beams (either one continuous-span beam as long as the width of the structure but supported in the middle by the longitudinal beam, or two single-span beams each half the width of the structure, spanning from the longitudinal beam out to their respective wall plates (keta 桁). When it runs over the raised floor zashiki part of the minka, this longitudinal beam is called the nakabiki-bari (中引梁, ‘central pull beam’), and when over the earthen-floored doma it is known as the ushi-bari (丑梁, ‘ox beam’).
There are several ways of resolving the structural relationship between the transverse beam, wall plate, and post:
In the orioki-gumi (折置組, ‘folded framing’) method, the mortised beam and wall plate sit directly on a long, stepped or ‘stacked’ tenon (kasane-hozo 重ね枘) cut into the head of the post, with the wall plate sitting on the beam and cross-lapped (ai-kaki 相欠き) into it.
In the kyо̄ro-gumi (京呂組, lit. ‘capitol backbone framing’) method, the wall plate is mortised into the beam and the beam sits on and is lapped and dovetailed (ari-tsugi 蟻継ぎ; ari 蟻means ‘ant’ and the joint is named for the supposed resemblance of the dovetail tenon to the pinched waist and abdomen of an ant) into the wall plate. There are two variations of the kyо̄ro-gumi: the kabuto-ari (兜蟻, ‘helmeted dovetail’) in which the end of the beam is flush with the outer face of the wall plate, and the watari-ago (渡り腮, lit. ‘crossing jaw’) in which the end of the beam extends out past the wall plate.
In the sashi-zuke (差付け, lit. ‘insert attach’) method, the beam is tenoned and pegged into a through-mortise in the post, and the wall plate sits on the head of the post.
These various methods each have their own advantages and disadvantages. In the orioki and sashi-zuke methods, there must be a post directly under each beam, whereas in the kyо̄ro method, beams can be placed at any location along the wall plate as long as it is deep enough to bear them, and intermediate posts can be offset from beams or omitted altogether. Orioki and sashi-zuke provide an uninterrupted upper surface of wall plate to support common rafters (taruki 垂木 or 棰), and a ready-formed lip between the beam and the top of the wall plate to set principal rafters (sasu 扠首) into. Orioki and sashi-zuke are older methods, while kyо̄ro is relatively newer.