Tokyo and Osaka in prewar aerial photographs from around 1936-42

Published: September 11, 2026 / About 14 min read

What appears in the prewar photograph and disappears in the next era

If you step old aerial photographs back toward the older eras, at most locations the black-and-white 1961-69 photograph is the end point. Around the 23 wards of Tokyo and the City of Osaka, however, one or two more photographs wait beyond it. These are the photographs taken around 1936-42. If you open this photograph and then go back to 1961-69, you can see at once what the town lost in the 20 to 30 years in between and what it carried on as it was.

The things that appear only in the prewar photograph and cannot be found in the next era differ completely from location to location. They are pitch-black squares with straight outlines, bands of tracks clumped into masses by freight cars, gravel riverbeds shining white across the full width of the river, and strip-shaped paddies filling the area south of the levee from end to end. This article lines up locations in Tokyo and Osaka and checks those differences one by one. The guide to the eras of aerial photographs summarizes the character of the photographs of each era and which eras can be used in which regions, so if you first grasp how far the two prewar eras reach, you will not get lost.

Squares blacked out as if with ink

When you open the prewar photograph, this kind of surface is what puzzles you most. In the photograph from around 1936-42 of Yebisu Garden Place, square surfaces with straight outlines sink at the upper right and the lower right of the frame, at a darkness completely different in tone from their surroundings. In the town around them you can read both the roads shining white and the fine roofs, yet only the inside of those squares is pitch black. If you switch to 1945-50, there are no black lots at the same positions, and a large, thickly mounded wood appears at the lower right.

In the photograph of Around Amagasaki Q's Mall, the same thing happens at several places in the middle of the frame. They are squares filled with an even black as if ink had been dropped on them, and their outlines are straight. The surroundings are strip-shaped paddies and fields, in a tone where you can read even the difference in brightness of each paddy, so only the black squares seem to float. In the photograph of Morinomiya, the black surface is larger and spreads irregularly from the east to the north of Osaka Castle. Its northern end touches the river, and its southern edge is cut off almost straight along a road. Countless diagonal lines run over the surface, and white spots left unpainted are scattered inside it.

It is not that the photograph turned out badly. In the same photograph, the double moat of Osaka Castle and the white buildings lined up neatly inside the castle grounds appear clearly. From the traces of the lines running over the surface, you can read it as a surface that was worked on after the photograph was taken. This photograph cannot tell you what was left out of it.

Two kinds of black surface

The photograph from around 1936-42 of Kento (Former Suita Rail Yard) also has a pitch-black surface. This one, however, is not a painted surface. The band of tracks appears as a black spindle-shaped mass, and the individual lines can no longer be read. This is because it is filled with freight cars. The surroundings are farmland as orderly as ever, and a wide, bright white road runs diagonally from the upper left to the lower left. The contrast between the fine grain of the farmland and the blackness of the band is strongest at this location in this era.

There are two clues for telling them apart. One is the outline. The edge of a surface that was worked on is a straight line or a broken line along a road, and it cuts off without regard to the shape of the ground. A band blackened by freight cars is long and narrow along the row of tracks, and it tapers at both ends. The other is how the neighboring era looks. If you step to 1945-50, the Suita band changes to a bright surface, and the individual lines become readable. This means that when few freight cars are left standing, the same band appears bright. As for the painted squares, the lots themselves cannot be found in 1945-50.

The burned town and the yards that did not collapse

Between the prewar photograph and 1961-69, one more photograph is inserted, from 1945-50. If you go through it without skipping, you can see that the change in the towns splits into two steps. In the air raids of 1945, wide areas of the towns of Tokyo and Osaka burned. If you open this era at Shiodome Sio-Site, the burned town has a lower density of buildings, and only the way the blocks were divided remains as lines. The surface of the yard lined with tracks and with the transit sheds that handled freight, however, has hardly lost its shape. It is an era in which things that break easily and things that do not can be seen apart within the same frame.

The same contrast appears at Umekita (Grand Front Osaka). If you go back and forth with the photograph from around 1936-42, there are places where the division of the blocks remains but the density of buildings seems to have dropped out, and only the band of tracks keeps its shape. The 1945-50 photograph of Around Amagasaki Q's Mall is a coarse-grained frame, bright and whitish overall. Lots blanked out white are scattered here and there in the town, having lost the outlines of their buildings, with only the shapes of the lots remaining. The same frame also shows lots with long, narrow buildings lined up neatly, and burned surfaces, working factories, and fields that still remain coexist in one photograph.

At Morinomiya, the area that was covered in black before the war returns in this era to a clear gradation of white and black. A gray expanse that has lost its outlines remains on the surface where the paint came off. Below the center there are almost no building shapes, only a few winding lanes running through. Toward the top of the frame there is a large rectangle with a white striped pattern, which can be read as buildings whose roof ridges still remain. If you look in succession at the surface that was hidden and the surface where nothing is hidden anymore, the meaning of the prewar photograph seems to turn inside out.

Futako-Tamagawa, with a wide riverbed and only one bridge

In the photograph from around 1936-42 of Futako-Tamagawa, the Tama River cuts the frame diagonally from the upper left to the lower right. The streams of water are thin, and on both sides of them gravel riverbeds shining white take up a surprisingly wide area. The flow does not gather into one; it splits within the riverbed, goes around sandbars, and joins again. There is only one bridge crossing the river in the frame, a road extends northeast from its foot, and the rows of houses thicken only on both sides of that road. Outside, it is mostly lots of fields and paddies, and the foot of the bridge was the core of the town.

In 1945-50 the white of the gravel spreads even further, and long, narrow sandbars remain like islands inside the riverbed. There is still only one bridge, and the rows of houses grow thicker only on its northeast side. If you step to 1961-69, both banks are already filled with houses, the lots of fields are pushed to the corners of the frame, and buildings come close to the inside of the levee line. On the sand, which had been only white until then, a faint grid of lots is drawn. The white expanse that caught the eye most in the prewar photograph did not vanish, swallowed by the town; it changed its role as lines were drawn over it.

Roofs lying on the island and flat land still empty

In the photograph from around 1936-42 of Okawabata River City 21, many long, narrow roofs lie in parallel on the northern half of the island. The buildings face the same way, with narrow passages running between them. The west bank is a town divided into a grid, with the grains of the roofs packed without gaps. The part not to overlook is the lower part of the frame. Bright flat land spreads there with only the grid of roads drawn on it, and the lots are cut but have no buildings. Land whose reclamation was finished but which had not yet become a town fits untouched in the same photograph.

In 1945-50, too, the way the river divides does not change. Long, narrow bright bands line up on the island, and small boats are dotted over the surrounding water. On the east side, long, narrow compartments divide the water in many rows, a shape that can be read as a place handling timber. In 1961-69 the lines over the water increase. A bridge crossing straight from the west bank to the island is built, and the buildings on the island thicken one by one as their roof surfaces widen. On the water east of the island, countless small shapes gather, and the bustle of the boat basin turns directly into grains. On this island, the shape of the water stayed still while only the way the water was used was replaced.

Takashimadaira, where the prewar photograph and 1961-69 look almost the same

There are also locations where setting the two photographs side by side brings out hardly any difference. Takashimadaira Danchi (housing complex) is one. In the photograph from around 1936-42, the Arakawa River crosses the top edge of the frame, and just south of it a wide levee meanders east to west. Inside the levee it is paddies from end to end. Long, narrow strip-shaped lots line up facing the same way, and since each lot has a different brightness, the whole frame looks like a fine checkerboard. White paths on paddy ridges run lengthwise and crosswise between the paddies, and the houses keep to the southern edge.

The 1945-50 photograph is fine-grained, and you can read even the shape of each paddy. A narrow waterway meanders along the south of the levee, and the grid of paddies draws water from that waterway. No grid of streets is drawn anywhere. Even if you switch to 1961-69, the left half of the frame is still paddies. Movement appears in the north: factories and warehouses begin to line up along the waterway, and only two large white roofs stand isolated in the paddies. The grid of paddies disappears all at once and rows of slab-shaped buildings appear in the next era, 1974-78.

At this location the three photographs, prewar, just after the war, and the 1960s, are almost the same picture, and the replacement comes all together within one era. The value of the prewar photograph is not its age itself. It lies in being able to bracket when the change came by the intervals between eras. If comparing the three shows that nothing moves, that is one of the answers too.

Senri, where the hills were still villages

The photograph from around 1936-42 of Senri New Town (Senri-Chuo) shows the form of the hills just as it was. All over the frame, valleys branch like twigs, and terraced paddies continue along their bottoms. The outlines of the paddies curve like bows along the contour lines, and each paddy remains small in area. Dense woods and bamboo groves sit on the ridges, and fields that look white cling to their shoulders. The settlements cluster at the mouths of the valleys and at the feet of the slopes, and no wide road crossing the hills can be found anywhere. Every road is only wide enough to link paddies and settlements. The elevation of this location is 66.6 m, more than 60 m different from 0.7 m at Umekita (Grand Front Osaka) and 3.4 m at Takashimadaira Danchi.

In 1945-50, too, this composition hardly moves. The paddies in the valleys show somewhat darker traces of care, and in some places the lots of fields are divided more finely, but no sign of a town comes in. The change comes in 1961-69. Residential buildings facing the same way stand neatly in rows only on the right side of the frame, while the left half is still paddies and woods, and the boundary gives you an idea of which side the grading approached from. According to the City of Suita's explanation, Senri New Town opened in 1962, and residents began moving in from Satakedai. The prewar valley paddies and the rows of residential buildings stand side by side as neighboring eras in the photographs.

Land price records begin long after the photographs

Even at locations where prewar photographs remain, the records of official land prices (Chika Koji) begin much later. The three locations Morinomiya / Around Amagasaki Q's Mall / Takashimadaira Danchi begin in 1983, and Yebisu Garden Place in 2011; in every case the photographs reach nearly half a century or more further back. As a means of checking the look of the land in times without value records, the prewar photograph plays no small role. The observation points are in residential areas about 300 m to 900 m from each location.

Even when you look at the same 2026 values, the way of reading changes if the entry points of the series differ. The three locations continuing from 1983 recorded high values from 1988 to 1991 and then went into a decline, and their 2026 values are below that peak. The Yebisu series, which begins in 2011, never included that peak in the first place. As with the eras of the photographs, if you compare without aligning the entry points, you misread the differences.

Clues for going back and forth between two photographs

The prewar photograph is coarse-grained, scattered with white spots and thin scratches, and offers no clues of color. Even so, there is material you can read. A surface with straight outlines is a surface people have worked on, and lines curving like bows along the contour lines belong to surfaces decided by water and terrain. If you fix one thing that does not move within the same frame, the amount of change becomes easier to measure. The way the river divides, the position of the levee, the direction of the tracks, and clumps of woods serve that role. The steps for switching and overlaying are summarized in How to read old aerial photographs.

When you come across a black surface, first look at its outline, and then try stepping to the neighboring era. If it is a square that was worked on, that lot can no longer be found in 1945-50, and if it is a band filled with freight cars, a bundle of bright lines appears at the same position. If you open the prewar photograph with this approach, 90 years of the town rise up as the difference between two or three photographs. On each location's page you can choose eras and compare them, so start checking from the photograph that caught your interest.

The prices in this article are actual examples at each point in time and do not reflect market prices after the article was written.