Since the start of the pandemic ‘lockdown’ and the subsequent passing of my wife and partner in publishing, I have had next to nothing to do with miniature naval ship models except for the final editing of my book ‘Naval Ship Models of World War II’ for Seaforth. Recently, however, I saw a damaged Konishi Maya for sale on the web site 1250ships. While I had aircraft and a submarine from this Japanese firm, I had never owned any of their larger models. I was eager to examine one of them closely, as they were made from lost-wax cast brass, a technique not used by any other ship model producer.


Comparison of Maya 1944, as converted to an anti-aircraft cruiser, made by Neptun and Konishi. Neptun Maya gravity- or drop-cast in silicone mould with soft metal alloy, while hull of Konishi Maya is diecast zinc, like Triang models, and superstructure and all deck components are lost-wax cast brass. This is an ancient technique, now largely used for making jewelry. Brass is much stronger than Wismut, the German soft metal casting alloy, making parts like yards, masts and gun barrels much less liable to breakage, but does result in care being needed in straightening them. In addition, brass is harder to file, burr or drill, as is the zinc hull. Note differences in the hull bulge on the two models, which may be more accurate on Neptun Maya. Both are 16.3 cm long.

Close-up of Neptun Maya, showing casting residue not cleaned up, like on foremast below radar room, and gun barrels not fully separated on twin 12.7cm mounts. Note poorly painted and incorrectly placed running light, as well as no yards on foremast. By 1944, Maya had Type 13 air search radar on rear of foremast; this is missing in both Maya models.

Close-up of damaged Konishi Maya, with superstructures on the hull. Note that brass mast parts and gun barrels are slightly thicker than those on the Neptun Maya, and there is no radar hut on the foremast. Windows, searchlights and blast bags are painted, but not running lights.

In this process, wax models, usually cast in moulds from a metal or master from a hard enough material to withstand mould-making, are invested in a plaster-like substance, which is then put in a kiln to melt out the wax, leaving a cavity shaped exactly like the now melted-away model. Molten metal is then poured into the hardened mould; after the metal hardens, the mould is broken off, to reveal the metal casting, which is then cleaned up. Thus, in this process, the mould is sacrificed with each casting, requiring a new mould to be made each time. This is unlike silicone moulds, which can be re-used for between ten and two hundred times before the mould is too damaged for further use.

Close-up of Konishi model after stripping paint with acetone, which also dissolved most of the remaining glue, visible as white remnants. I have not yet stripped paint off part of the superstructure and main battery turrets. These images show superstructure components, with masts and most secondary gun mounts now separate. Note detailing on zinc hull, with raised lines meant to represent brass strips to hold down linoleum deck covering and anti-skid pattern on deck at stern. Close-up shows that masts still need to be straightened further, as well as twin 12.7cm gun barrels. Thick, cast cable for crane will be removed, and masts will be detailed and cleaned up.

Damaged models, like damaged jewelry, attract me like a magnet. Although my training and first career was as a biomedical scientist, in which I still maintain an active interest, for the past forty-five years I have been a researcher and writer about personal adornment. In this field, damaged objects almost always reveal more than intact examples. Thus I was eager to examine this Konishi Maya and compare it to a similar Neptun Maya, which would have been cast in a silicone mould with a soft alloy of tin, antimony and bismuth, called Wismut by German producers.

Close-up of superstructure and secondary AA mounts, revealing paint remnants. Most likely an enamel had been used, resulting in its tenacity even after soaking and brushing with acetone. Note that most of the castings are very clean, without casting residue. All castings have distinct lines and are complete, unlike silicone-cast models, where parts can be partially cast or distorted due to mould wear. Since the wax models for this type of casting are done from a mould not subject to high heat or wear, each cast is more like the original master. The economic cost is in the large amount of labour required for lost-wax casting and assembly of the model.

Before working on the damaged Japanese-made model, the superstructure of which had come unglued, I carefully took comparison photos using a stock Neptun Maya, also configured for 1944. Whenever I do a restoration or enhancement of a model, I always document the initial condition. The plan was to clean up and repair minor damage to the Neptun Maya, and slightly enhance the model by adding correct yards and a ‘Jake’ floatplane. The soft alloy 12.7cm gun barrels are very susceptible to damage, as they extend beyond the hull, so handling often bends or breaks them. A number had to be replaced with copper wire, a difficult gluing task. This was not a problem with the Konishi model, as the cast brass barrels are much stronger.

Mayas after enhancements added to Neptun Maya, limited to yards, replacement of broken off 12.7cm barrels, since they extend beyond hull and are easily broken when picking up model. Much more extensive work on Konishi Maya, with addition of copper and iron wire masts, copper RDF loop on bridge, Type 13 radar on rear of foremast, addition of seven 3D-printed PaperLab resin 25mm AA aft and on no.4 turret, as well as considerable cleaning up and adjustment of masts so that they fit properly. Note foremost starboard 12.7cm barrels bent through careless handling.

For the Konishi model, since it was now almost bare metal, I did much more detailing: adding correct mast features, yards, RDF loop, Type 13 radar, a radar room in the foremast, replacing a missing ship’s boat, a ‘Jake’ floatplane and adding individual single 25mm, using PaperLab’s new 3D-printed resin guns. Seven such weapons were shown in the Maya 1944 drawing on page 56 in Steve Wiper’s 2008 book on Takao class cruisers, although Japanese plans show up to twenty-one single 25mm guns. The actual Maya had twenty-eight single 25mm AA. The over-scale cable for the crane was also removed, and not replaced.

The most difficult parts of this project were the replacement of broken 12.7cm barrels, the fabrication and gluing of the Type 13 radar, and the installation of single 25mm 3D-printed resin guns. Resin is very fragile and breaks easily, and the guns’ small sizes and very light weight results in their easily slipping off tweezers and getting lost. I broke or damaged four out of the set of ten resin guns, so I had to adapt one single 25mm from a triple 25mm mount, also made by PaperLab. Thus this improvised 25mm gun consisted of three parts, made of resin, stretched plastic sprue and forged copper to form the round base of the mount. As shown in the close-up photo, this scratch-built 25mm did not quite match the other guns. In order for the yards and air search radar to be at right angles to the hull, considerable adjustment was necessary for the legs of the foremast.

Close-up of Konishi Maya, to show three single resin 25mm AA, with one furthest aft modified from PaperLab triple 25mm mount. Single guns come ten to a set but I broke the barrels of two, leaving only one spare. During the difficult clipping, cleaning and gluing of these, another two were broken or lost, so I had to adapt an additional one from a set of triple 25mm, with a separate stand of plastic and forged copper for the round base of the mount. As a result it stands slightly taller.

Close-up of Konishi bridge and foremast, with copper RDF loop on bridge roof and constructed mount and Type 13 radar on rear of foremast, as well as radar hut within structure of mast. Thicker copper wire was used for radar to match appearance of cast brass mast parts. Note that a lot of paint still remains.

This modelling project demonstrates my primary goals: to learn about how miniature naval ship models are made and to portray them as accurately as possible, although this is not achievable, as in this instance. Obtaining enough resin single 25mm guns and installing all twenty-eight of them would have probably driven me beyond my patience level.

Why I continue to be intensely interested in naval ship models of WWII, and why I have written a book on the same topic, has to do with how much my life was affected by this conflict. I was born in 1938, in Rome, Italy, when my father was the Chinese ambassador to Mussolini’s government. Prior to that he served as Chinese minister to Germany and Austria, when Hitler was in power. The Sino-Japanese war had already started, and when Japan joined the Axis, the friendly relations between China and Germany/Italy ended. We returned to China on the Italian liner Conte Verde, my father leaving for the wartime capital of Chungking, while we became stranded in Japanese-occupied Shanghai. While there, the city was bombed, most likely by B25s of the Fifth Airforce. In late 1944, my mother and five children left on a three-month journey to Chungking, traveling by train, wheelbarrow, mule cart, on foot at night to cross enemy lines, pulled on wooden sleds across the frozen Yellow River, and finally by open US International cargo trucks. After VJ day, most of the family flew to Nanking in a C47 troop carrier, landing at an airport full of surrendered Japanese warplanes. My late, oldest brother Johnny flew back in the tail gunner’s position of a Chinese B25. In 1946, we came to the United States on the USS Marine Lynx, newly disarmed, for the older children to attend school, but the Communist takeover prevented our return.

All photographs by Robert K. Liu, using macro lens and studio strobe lighting.

Preorder a copy here.