I know … I am combining languages with this expression, but fusion is what we do at the Red Top Workshop, and “Tana Moderne” is how I describe the latest Project. Here’s the background. Collectively, tana is the Japanese word for “shelf” or “shelving”. When used as stand-alone furniture, these pieces are further identified by their intended use and location: simple tana are the shelves used for storage in the kitchen or shop; chanoya-dana are pieces used for ceremony in the tea room; and kazari-dana are ornamental shelves used for displaying objects in the drawing room or study. I am making a set of end tables for a customer who is fond of Asian decor, and after batting around a few ideas we settled on a multi-tiered, kazari-dana inspired design. Two defining examples of this form, made of mulberry wood, are shown below.

*reproduced from: Koizumi K. Traditional Japanese Furniture, A Definitive Guide, K. Koizumi, Kodansha International: Tokyo, 1986.

Design

The kazari-dana motif goes back some 500 years. These early examples are beautifully delicate and even the so-called “informal” versions are often flourished with intricate carvings. The style was right for this Project, but to match the customer’s modern living room setting we would need to come up with a sans serif version. Modern furniture design generally eschews frills and a moderne version of this recognizable classic was the goal: moderne (Fr.), adj. - [design] characterized by strong horizontal lines and minimal surface ornament. Still unclear on exactly how to accomplish this transformation I happened upon the answer one day while poking around Google. The “prototype” example shown below came from a web auction site with no further information as to age, dimensions or price. It is an unmistakably modern interpretation of a kazari-dana that would anonymously inform this Project on both proportions and modesty. But first the double doors and lacquer finish would need to go. We played around with the overall scale and door design until the customers’ desires were satisfied (seasoned with a healthy dollop of trust on their part). The final request was for two such end tables, one with a doored compartment and the other left open and doorless. Again with the fraternal twins!

Materials

These pieces would be made from black walnut which, like mulberry, is a dark and boldly figured wood. Depending on how bold things looked after mating the individual boards I could apply colorant (stain/dye) to modulate the grain “volume” so as not to drown-out the kazari-dana form. A single, light colored door made of birch plywood would operate on pivot hinges and open with a modern wooden knob. The twin tana would be constructed in the same manner and lack only the compartment feature (i.e., door and backboard). It looked to be a straight-forward build with the wood taking center stage. I purchased 42 board feet of 5/4 walnut and also added a leftover 5/4 plank from a previous Project. These were “rough” grade boards, meaning no additional surfacing had been conducted following their removal from the log, and while the presence of knots and other natural defects could be detected through the fuzzy exterior, the detailed grain pattern and color would only be revealed once my tools got involved.

Dimensioning

To begin, all of the planks were thickness planed to 1 1/8 inch in depth. This operation created a uniform dimension and also revealed the grain, allowing for segregation of board regions by “character”: Showy for the tops and base; Serene for the sides and pedestals; Straight laced for the posts. They were then jointed on one edge to give a square reference. Solving the next puzzle would set-up all subsequent operations on the Project, and that was: Where, exactly, would each of the 42 parts be extracted from? There was over 4,500 sq. inches of 1 1/8 in. deep material to choose from, but this area was parceled within the haphazard lengths, widths and “characters” of 8 separate boards. It wouldn’t do to begin chopping out “cut list Item 1.” from the near end of the nearest board and hope for things to work out in the end - they would not. A more strategic approach was called for, informed by the actual part dimensions and, more importantly, the pairings for the doublets and triplets of smaller boards that would need to be harmoniously joined together to create bigger boards. Indeed, all six of the horizontal surfaces [(base, lower top, upper top) x 2] would each be created from three component boards. The final width of these platforms was 16 in. and, while there are many ways to divide this into three, making them all from uniformly wide (5 3/8 in.) pieces would simplify things both in terms of construction and appearance. The vertical members [(2 sides, 4 posts) x 2] would also need to be assigned within the emerging plat map and similar strategies were employed to locate these. The following checklist helped to account for all parts on twin pieces A and B, and were sized 1 in. longer than required for final dimension.

SHOWY: upper top A (19,19,19), B (19,19,19); lower top A (21,21,21), B (21,21,21); base A (35,35,35), B (35,35,35) = 5 3/8 in. wide stock

SERENE: sides A (13,13) (13,13), B (13,13), (13,13) = 6 1/2 in. wide stock; pedestal A (32/32) (14/14), B (32/32) (14/14) = 3 in. wide stock

STRAIGHT: A (19/19) (6/6), B (19/19) (6/6) = 1 1/8 in. square stock

#= length in inches; (#,#,#) = glued together; (#/#) = paired on the same piece, but not joined

Of course, you don’t need to know any of this. I just thought I would give you a feel for the planning that happens before the cutting begins. Thanks for sticking it out!

To begin, the dark boards were marked appropriately with white colored pencil, cross cut to the desired lengths and then ripped to the requisite widths on the table saw. Since I was building both pieces at the same time, clearly labeling the parts as to what they were and to which tana they belonged was critical. Once the surface components (tops, base, sides) were glued together, thereby converting 26 boards into 10 (i.e., 5 per twin), colored tape served to keep things straight.

Each board would have to be refined further to create the mortice sockets, dados and tenons required for assembly. This involved careful measurement followed by the faithful reproduction of cuts within and between tana parts. Making two pieces at the same time gave it all a “production” feel. First, dados (technically “housings”) to hold the side pieces were fashioned in the base and lower top. Since these were so-called “stopped” dados (i.e., stopping short of running across the entire width of the piece) I first created two mortice holes at the start and stop points. A hand held router was then used to excavate the 3/4 in. wide canal in between. I employed a home-made jig to keep everything aligned and performed two successively deeper passes with the router to obtain the 1/2 in. desired depth. Worked well.

Additional grooves were required in tana A to house the backboard of the doored compartment. These were cut into the base, the lower top and the two sides using a 1/2 in. straight bit on the router table. A piece of 1/2 in. cabinet grade plywood would tuck into these slots during assembly.

Both twins would be elevated 3 inches above the floor by a rectangular structure that I call a “pedestal” to avoid the indecorous “toe kick”. Joining the corners together is the only woodworking involved here and since no information could be gleaned from the prototype picture I decided to use a simple box joint (sammai-gumi-tsugi) for this element. Box joints are found on the corners of many antique Japanese furniture pieces, often secured further with small bamboo pegs. It creates a distinctive feature that I would try to emulate here on a chunkier scale. The pedestal members were cut to final width/length dimensions using the table saw/miter saw and then the ends were notched appropriately using a bandsaw and chisel. Pilot holes were drilled into the frame parts where wood screws were then used to fasten the corners during this sub-assembly. In lieu of bamboo, light-colored maple plugs were tapped into place to cover the screw heads and then sawn flush. Prior to final assembly, pilot holes for the pocket screws that would affix the pedestal to the base were created along the frame members.

The upper top is supported by four posts that fit into morticed sockets created on the underside of this part, the topside of the lower top, and the base. It was a pretty routine operation to create the sockets, the only tricky part is to ensure that the smaller, 5 in. posts connecting the lower and upper tops are mounted perpendicular to these surfaces. Since this is a matter of socket alignment from two separate planes, ruler measurements could only get you close. To achieve a more accurate placement I delayed cutting the mortices in the lower top until the very end when a dry-fit of all the other parts would allow me to square things up, insert a “mock” post (possessing a single tenon on top) and then trace the exact position for drilling into the table top. Once the first socket was marked it was created at the mortiser. The “real” post was then trimmed to fit and inserted. The mock post was used to template the second socket, and … [repeat 3x]. The final challenge will be to achieve this same “dry-fit” pose during glue-up.

The posts themselves were fashioned by ripping straight-grained stock to 1 1/8 in. square, cutting the pieces to length, and then creating 3/4 in. square x 1/2 in. deep tenons on the ends using the dado blade on the table saw. The tongues along the top and bottoms of the side pieces, that slide into the housings described earlier, were also created at this time. After removing the surface mill marks, the sharp corner edges of all upright pieces were rounded-over slightly at the router table.

Getting close! A full day was spent card scraping all surfaces to even-out glue seams and remove mill marks acquired during the earlier operations. Then the edges of the tops and base were rounded over at the router table. Two successively deeper passes on both sides of all edges were required to create the desired, blunted bullnose profile in this dense material. Finally, all surfaces were sanded silky smooth down to 400 grit.

Assembly & Finish

After much middle-of-the-night pondering on how to install that door it was decided to mount the hinge hardware prior to glue-up. Unlike conventional butt hinges which attach to the side frame, the pivot hinges used for this piece “sit” within the top and bottom frames, which, since this is a frameless inset door, translates to the lower top and base. The mounting mechanism is a single hole drilled into these surfaces into which a plastic collar is inserted followed by the metal hinge pin. As the door opens these pins rotate within the collared holes. Drilling said holes 11/16 of an inch away from a fixed side board would not be possible with my hand drill, thus necessitating the pre-glue-up gambit. There is just a little front to back “play” within the housings of the side board tongues and so careful attention to hole alignment would be needed during assembly. To assist during glue-up I constructed a “mock” door of correct height but shortened width. This allowed easy manipulation inside and outside of the cavity, and its temporary installation facilitated vertical alignment during assembly. Following glue-up it could be swapped out for the actual door, prepared according to the finished dimensions. These hinges are easy to work with and provide a nice, hardware-free appearance for mimicking a sliding door design. The over-sized plate knob (in walnut!) would complete the look.

In the end, I decided not to stain the wood but, instead, allow the color variation and grain currents to flow free. More formal examples of kazari-dana found online reveal that, in the high-end homes of old, these pieces were often wildly decorated using gold figures on lacquer. Turns out this shelving was meant to let loose! and the walnut grain now had a real responsibility to put out. Doing what I could to assist, I went beyond the usual gel poly finish in favor of the oil-based products, Seal-A-Cell and Arm-R-Seal. This sealer and wipe-on varnish duo from the General Finishes company are reputed to be excellent for accentuating wood grain while also providing good water & wear protection. They did not disappoint! One coat of sealer was followed by three of the wipe-on varnish. The door and knob were finished in a similar sequence and then mounted into place. Final buffing gave a soft lustrous sheen on these walnut renditions of Tana Moderne.