The Darlington Building Collapse: Modern Engineering and Obsolete Systems
Donald Friedman, P.E., F. ASCE
Old Structures Engineering, PC
90 Broad Street
New York, NY 10004
The 1904 collapse of the Darlington Apartments during construction was a sudden and
complete failure: eleven stories fell into a pile of rubble less than 15 feet high in a mat-
ter of seconds, killing 25 people. The collapse and forensic analysis were prominently
reported in the newspapers and engineering press; thirty years later, the publicity was
cited as a deterrent to structural use of cast-iron columns. This failure became perma-
nently linked to the shortcomings of cast-iron structure.
If completed, the Darlington would have been typical of an obsolete structural type: the
high-rise cage-frame building. Cage frames, ﬁrst built in the 1870s, had an iron frame
supporting the ﬂoor gravity loads and surrounded by a self-supporting masonry wall
that provided lateral stability to the building. The use of cast-iron columns in commer-
cial buildings with cage frames had eﬀectively ended by the mid-1890s; the structural
engineers who were increasingly used as consultants in commercial high- rise design
preferred wrought-iron and steel columns. Wrought-iron and steel were known to have
lower allowable direct compression stresses than cast iron, but were ductile and could
safely withstand accidental tension and moment. The gradual replacement of cast-iron
with the ductile metals in the late nineteenth century was encouraged by fears of col-
lapse caused by the brittleness of cast iron.
Cage frames remained popular in high-rise apartment houses for nearly a decade after
they were no longer used in commercial construction. Unlike tall commercial buildings,
which were built in cities throughout the United States, tall residential buildings were
concentrated in a few cities, especially New York. These buildings were typically de-
signed by residential architects working without consulting engineers. Common prac-
tice was for the iron sub-contractor to provide “engineering services,” often consisting
of sizing steel and cast-iron columns from tables based on the span and ﬂoor load sched-
ule. In short, lateral load analysis was not part of the design, so the deﬁciencies of cage
framing were not made visible.
This paper will describe the design and construction background to the failure, the
forensic analysis performed at the time, a modern review of the failure, and discussion
of cage-frame failures within the engineering community.
1 of 15
Building and Collapse Description
The Darlington Apartments was designed to be one of dozens of similar apartment
houses erected in New York City before World War I. The rapid growth of the city’s
population, from 1,500,000 in 1870, to 3,400,000 in 1900 (following the 1898 consolida-
tion of neighboring counties into “Greater New York”), to 4,800,000 in 1910, combined
with the even more rapid growth of the downtown business district to create a climate
of nearly continuous construction. New York, as the city with the highest residential
and commercial population densities in the country, became a center of new construc-
The building, located in midtown Manhattan at 59 West 46th Street, was to have twelve
typical stories with a smaller, set-back penthouse. The general plan was 55 feet by 90
feet overall, in a broad “U” around a light court. (Figure 1) The full height was to be
148 feet, with 10'-10" typical ﬂoor-to-ﬂoor spacing. Neville & Bagge, the architects of
record, had designed many smaller apartment houses, tenements, and row-houses. The
general contracting ﬁrm, Pole & Schwandtler, was also the engineer of record.
The structure of the building was relatively simple. The “Roebling ﬂat-slab” ﬂoors were
cinder-concrete slabs reinforced with rectangular steel bars; the interior partitions
were non-structural terra-cotta tile carried on the ﬂoor framing; the ﬂoor beams were 6-
inch, 7-inch, and 9-inch steel channels and I-beams; and the columns were hollow cast-
iron sections ranging from 6 inches square with a 3⁄-inch wall up to 10 inches square
with a 1⁄-inch wall. Per standard practice, the cast-iron columns were spliced at each
ﬂoor with pin details (top and bottom ﬂanges cast integrally with the columns received
two bolts) and the beams were connected with stiﬀened seat connections also integrally
cast with the column shafts. All connections were loosely bolted. (Figure 2) The design
live load was 60 pounds per square foot (psf) on the ﬂoors and the design dead load was
also 60 psf. The exterior wall was self-supporting – the frame spandrel beams were lo-
cated entirely within the inboard face of the exterior wall – and increased in thickness
from 12 inches at the top to 20 inches at grade in accordance with the New York City
Building Code. It should be noted that, per common practice at the time, all masonry
other than some front-facade trim was solid, unreinforced brick. (“The Collapse of the
Even though the frame did not support the exterior wall, the spandrel columns were
completely enclosed within the wall. This was an intentional part of the design: as the
wall was intended to resist wind loads, it was necessary for the frame supporting the in-
terior gravity loads and the wall to be structurally linked. Enclosing the columns within
solid masonry provided a mechanism for force transfer between the frame and wall.
This design meant that all spandrel columns were eccentrically loaded, with oﬀsets be-
tween beam and column centerlines as great as 12 inches.
2 of 15
On the afternoon of March 2, 1904, the building collapsed, killing twenty-four laborers
and a woman in an adjacent building. At that time, the steel and iron frame had been
erected to up to the tenth ﬂoor, the concrete ﬂoor slabs had been placed to “two or
three” ﬂoors below the ﬂoor beams, and the walls were partially complete to the third
ﬂoor (rear wall) to sixth ﬂoor (side walls) levels. The front wall was not yet complete to
the second ﬂoor level.
3 of 15
Figure 1: “A” indicates a column eccentrically loaded in both axes,
“B” indicates a column eccentrically loaded around one axis, “C” in-
dicates an interior column possibly loaded eccentrically, “D” indi-
cates a concentrically loaded column, and “E” is a metal-clad bay
Figure 2: Column connections.
The collapse destroyed the building, leaving portions of the side walls, bent steel beams,
and cast-iron columns broken in pieces. (Figure 3) Witnesses observed the upper ﬂoors
of the frame swaying in various directions but then collapsing inward toward the center
of the site. (“The Collapse of the Darlington...,” 1904) In the following weeks, the col-
lapse was examined and discussed by various engineers, city oﬃcials, and the press.
Catastrophic failure of buildings, in the form of collapse and ﬁre, was part of urban life
in the late nineteenth and early twentieth centuries. The increasing size and complexity
of building codes in large cities was a response to both the increasing size and complex-
ity of buildings being constructed and to the perception that new construction technol-
ogy might alleviate known problems.
Construction of any structure larger than a hut has always been a technological process,
requiring tool use, artiﬁcially-shaped materials, and organization of groups of people.
(Bijker et al, 1987, p. 4) The use of new building technologies after 1850 contributed to
the gradual reduction in the rate of death and injury by ﬁre in urban buildings, al-
though the growth in population and increasing urbanization masked this reduction.
Public expectations that large-scale ﬁres could be avoided did not come until after the
Chicago and Boston ﬁres of 1871 and 1872, despite the existence of some forms of
structural ﬁre-prooﬁng before then. The rapid growth of New York City between 1870
and 1910 was coincidentally accompanied by rapid development of new construction
4 of 15
Figure 3: View from the south after collapse. A portion of the west side wall is visible at
left, adjacent to the row house.
technology; other cities have experienced their greatest growth with older technology
(e.g., Paris) or with newer technology (e.g., Houston). Late- nineteenth-century New
Yorkers could not always clearly distinguish between those changes caused by growth
and those caused by technology.
With few exceptions, the general form of building construction in 1870 New York City
was traditional wood and masonry, little changed from medieval practice. The indus-
trial revolution had introduced new methods of producing the individual elements used
(for example, by introducing powered gang-saws to speed the production of individual
lumber “sticks” from felled logs), but had not modiﬁed the types of elements used. The
ﬂoors of buildings were wood plank supported on wood joists; the joists were supported
on masonry bearing walls, usually brick or brick with stone veneer. Large open spaces
were created through the use of wood girders supporting the joists in lieu of bearing
walls. Girders were supported on wood posts, masonry piers or, after 1840, cast-iron
In a city of wood and masonry, little could be done to prevent ﬁres, although there were
eﬀorts to limit their spread. The creation of the “ﬁre line” divided the city between the
densely built-up portion and the sparser suburban areas. As the city grew north up
Manhattan island, the ﬁre line moved northward as well. Buildings south of the ﬁre line
were required to have masonry exterior walls, including party walls that marked the
boundary between separate buildings on separate lots but were physically enclosed
with the buildings; buildings north of the line could be entirely wood. In theory, the use
of masonry walls as ﬁre separation could have been taken further by requiring that
building interiors be compartmentalized with masonry walls, but this step had not been
taken by the time that the introduction of new ﬁreproof materials such as terra-cotta
ﬂoors made the question moot. (Freyer, 1898, pps. 288-293)
Various ﬁreproof buildings had been erected in New York and elsewhere by construct-
ing ﬂoors out of masonry vaults supported on bearing masonry walls, but this method
remained in limited use because of its cost. Most “ﬁreproof” buildings of the early and
mid-nineteenth century therefore consisted of wood and masonry with some form of
protection against external ﬁre spread.
The introduction of new structural materials and new systems in the middle and late
nineteenth century changed the nature of the dangers and public and professional un-
derstanding of these dangers. Cast iron was introduced into American practice in New
York and Boston in the late 1830s, following use in Britain and France. In the United
States, this was followed by large-scale wrought iron in the 1850s, steel in the 1870s,
and reinforced concrete in the 1890s. The typical pattern of use was introduction as a
one-for-one substitute for a well-established building element (such as the ﬁrst use of
cast-iron columns as substitutes for granite piers in storefronts) followed by the intro-
duction of large-scale systems dependent on the properties of the new elements (such as
the use of cast-iron columns and wrought-iron beams in interior frames that made possi-
ble large commercial buildings with no interior walls).
5 of 15
The substitution of iron for stone or brick provides no obvious improvement in ﬁre- pro-
tection, but in some cases the material superseded was wood. Cast iron served the mid-
nineteenth century as a symbol of modernity in construction, more speciﬁcally of cheap
industrial production of buildings. Cast-iron promoters emphasized safety and economy
when compared to traditional construction. (Bogardus, 1856, p. 5)
In that era, there was no required testing of the new materials and systems before
large-scale use in buildings. New materials were empirically tested by use and the
record of their performance under real-life conditions begins in the 1870s, when enough
new-technology buildings had been constructed for their performance to be recorded.
This period coincides with two of the largest nineteenth-century conﬂagrations, Chicago
in 1871 and Boston in 1872, and therefore began with a public sensitized to ﬁre danger.
Reports from those ﬁres emphasized the performance of new-materials buildings, in
large part because they had been advertised as “ﬁreproof.” (Bogardus, 1856, pps. 12-13)
Reports on individual building ﬁres in New York began to focus on whether or not the
building burning was “ﬁreproof” and, if so, how it performed compared to traditional
construction. Public commentary, including statements from governmental oﬃcials con-
cerned with building safety, professionals familiar with the issues, editorial writers
from newspapers, and professional journal writers began with the Chicago and Boston
ﬁres to examine the eﬀect of new materials in construction and new forms of building
control (such as the ﬁrst comprehensive New York City Building Code, enacted in 1882)
on safety. Simply, people were asking whether or not the new forms of building were
making New York more safe or less, a question that would continue through the begin-
ning of the twentieth century.
The ﬁrst new material introduced was cast iron, usually used for columns. Cast iron is a
product of the industrial revolution because its production requires intense heat that
depended on engine-powered forced air drafts and coal ﬁres. The metal is extremely
strong in compression – stronger than most modern steels – but weak and subject to
brittle failure in tension. These properties make the metal inappropriate for use in hori-
zontally-spanning beams, where tension and compression exist in equal amounts, but
feasible for use in columns, which are substantially in compression. Cast iron columns
were ﬁrst substituted for masonry piers in the late 1830s and became common in build-
ing interiors by the 1860s. (Condit, 1982, p. 81)
On a larger scale from material substitution was the use of entirely new structural sys-
tems, including the substitution of cage- and skeleton-frames for older masonry- sup-
ported structure. Traditional buildings are deﬁned as bearing-wall systems, where ma-
sonry walls support the gravity loads of the building itself (the wall weight and the inte-
rior ﬂoors) and its contents. The walls also serve to resist the lateral push of wind and
to provide exterior enclosure. Just as iron framing was substituted for wood ﬂoor joists,
iron was gradually substituted for the structural functions of the walls, resisting gravity
and wind loads.
6 of 15
The main advances in building systems technology were the introduction, circa 1875
and 1890, of cage and skeleton frames. A cage frame is one with a complete metal
frame (in which iron or steel ﬂoor beams are supported by iron or steel columns) sur-
rounded by a self-supporting masonry wall; a skeleton frame is one in which a wrought-
iron or steel frame supports all loads, including the exterior walls. (Freyer, 1898, p.
465) Cage frames are now seen as a transitional form between bearing-wall buildings
and skeleton-frame buildings, but were a legitimate method of constructing buildings in
the 1880s and 90s. Nearly all cage frames used cast-iron columns.
There are links between ﬁres and non-ﬁre collapses. Obviously, buildings badly dam-
aged by ﬁre often collapse, in part or in whole. Unlike traditional construction, which
had hundreds of years of empirical testing through use, new technology was often not
well understood when it was put into use. Cast-iron use, for example, went from being
considered an improvement in ﬁreprooﬁng to being considered a detriment, and the use
of unprotected iron was eliminated by New York’s 1882 building code. Iron was no
worse in ﬁre than wood, but its more serious ﬂaw was that it could fail without warning
if overloaded or improperly manufactured. The superiority of wrought iron and steel,
which fail gradually under similar conditions, only became generally recognized among
engineers in the 1880s and among other groups in the 1890s or 1900s. The worst col-
lapses were the result of over-loaded or poorly designed cast-iron columns and occurred
in cage-frame buildings.
Despite the similarities, we can distinguish between public reaction to ordinary ﬁres
and the reaction to failures in modern buildings. The dangers of urban ﬁres were well
known before the United States had cities of any signiﬁcant size and were associated
with the combination of wood construction and the domestic use of open ﬂame. In a so-
ciety where ﬂame was common – in ﬁreplaces, stoves, lamps, and boilers – people were
familiar with the dangers of ﬁre. The ordinary person of the late nineteenth century
had little familiarity with new building technology, and had to rely on expert opinion
about the safety of iron and steel framing, tall buildings, and new frame types. On the
other hand, building professionals view catastrophic failures not just as tragedies but as
means by which design and construction practice can be studied. As examples of ﬁres
in new-technology buildings were discussed in the press, the public understanding of
the meaning of ﬁre and ﬁreprooﬁng changed from that based on common knowledge to
one inﬂuenced by technology as described by professionals.
Building collapses, like ﬁres, have always been part of urban society, but bear a more
complicated relationship to the introduction of new technology. The substitution of non-
ﬂammable metals for wood posts and ﬁreproof ﬂoors for wood joists may not have cre-
ated the “ﬁreproof building” so often held as an ideal, but they did reduce the gross ﬁre
load (the amount of ﬂammable material in a given building) and, when coupled with ﬁre-
protection methods, gradually reduced the incidence of severe ﬁres. Collapses, on the
other hand, became more frequent in the 1880s and 90s as the average height of build-
7 of 15
ings in New York increased and as experimental structural systems were empirically
tested by use.
There were collapses of other cage-frame buildings before the Darlington, most signiﬁ-
cantly the August 8, 1895 collapse of the Ireland Building at 3rd Street and West
Broadway. The building was planned to be a commercial loft eight stories high and was
still in construction, although substantially complete, when its center collapsed, leaving
the perimeter walls and minor portions of the interior framing. The collapse killed 16
laborers and buried their bodies under a large pile of debris that impeded rescue of the
survivors and recovery of the remains. A coroner’s jury performed the formal review of
the collapse, including testimony by technical experts and visits to the site to look at the
debris. (“Imprisoned in a Wreck,” 1895; “Mr. Constable Inspects,” 1895; “Bodies All
Taken Out,” 1895; “Visited the Ruins,” 1895; “To Go Before the Grand Jury,” 1895)
At the Ireland Building, at least three contributing causes were identiﬁed, including the
possibility of inadequate bearing material below the foundations, inadequate design of
the interior column footings for the expected loads, and over-loading of at least one inte-
rior column with bags of gypsum by the plasterers working on the upper ﬂoors. How-
ever, the dominant cause was identiﬁed as the failure of the interior cast- iron columns
under high but not exceptional loads.
An editorial in the Engineering Record used the collapse of the Ireland Building to
highlight two issues that concerned engineers as a group but were largely unknown to
other professionals and the public. (“The Investigation of Building Failures,” 1895) The
ﬁrst was the general inadequacy of cast-iron columns. These elements, common twenty
years earlier, were not necessarily seen as inadequate by the public and were not
treated any diﬀerently by building codes than steel and wrought-iron columns, but were
described in the editorial as “entirely unﬁt.” For an engineering audience, cast iron was
speciﬁcally compared to wrought-iron and steel and found wanting, which is not surpris-
ing as the state of the art in structural metals had advanced signiﬁcantly during the
1870s and 80s. Cast iron is brittle and can therefore fail without warning and com-
pletely as it did at the Ireland Building, while the other metals are ductile and fail “only
after the most extreme or violent distortion, and even then enough will frequently hang
together to sustain the existing loading.” This engineering distinction between cast iron
and the two other metals was not widely understood.
The other issue identiﬁed was expertise. Coroner’s juries were then a normal mecha-
nism for investigation of untimely death, but jury members were selected in the same
random manner as criminal and civil juries. The Record editorial states that the Ireland
jury, by focusing only on the inadequate footings as the immediate cause of failure
missed an opportunity to “confer great beneﬁts upon the city of New York in connec-
tion with the construction of modern buildings” by addressing the general inadequacy
of cast-iron columns. The editorial bluntly states that coroner’s juries were inadequate
for investigation of disaster because of lack of expertise: “The questions which come be-
fore such juries are largely of a purely engineering character, which can eﬃciently be
8 of 15
treated only by men ﬁtly qualiﬁed through education and experience. Exact knowledge
of the nature of material, its behavior under loads, the competent design of main mem-
bers and details, and a multitude of other similar matters, must be intelligently consid-
ered.” Obviously, engineers were the only people that the editors of the Record consid-
ered qualiﬁed to judge building-related disasters. Further, the eﬀorts of government
oﬃcials to create a safe building environment were indirectly criticized: “The recom-
mendations of one such competent jury [composed of engineers] would furnish a ready
means of clearing the Building Ordinance of numerous absurdities and of introducing
eﬃcient provision for the attainment of creditable building construction, so that build-
ings, in the process of erection at least, could not fall down like a child’s block-house.”
(“Ireland Building Inquest,” 1895) This criticism was not entirely fair, as the jury heard
testimony by expert builders and engineers and arguably was therefore no less well-in-
formed than any jury examining a technical issue with the help of expert witnesses, but
it did highlight a gap between engineers and the public. Engineers saw vindication of
their work in the fact of a disaster that occurred from un-engineered design, in the
method of determining blame through professional investigation, and in their prefer-
ence for steel over cast iron; while the public saw a building of “modern” construction
collapsing catastrophically and without warning.
Investigation and Forensic Analysis
Like the Ireland Building, the Darlington had no engineer formally involved with struc-
tural design; like the Ireland, there were contributing causes that suggested negligence
on the part of the general contractor; but most importantly, like the Ireland, the root
cause of collapse was eventually identiﬁed as the use of a cage frame with cast-iron col-
umns. By this date, that combination of material and system was used in New York
only for medium-rise hotels and apartment houses, having been otherwise superseded
by steel-frame technology.
Increased sophistication of both the public and city oﬃcials can be seen in the ﬁrst arti-
cle in the New York Times describing the collapse. (“Death in Collapse...,” 1904) In ad-
dition to a narrative description of the collapse and rescue eﬀorts on site, the article
contained notice of arrest of the site foreman for negligence in overloading the ninth
ﬂoor with materials to be used in construction of the tenth ﬂoor; mention of the begin-
nings of investigation by the coroner’s oﬃce, the district attorney’s oﬃce, and the Build-
ing Department; and three distinct theories of the cause of collapse: overloading, faulty
construction, and the explosion of a hoist steam engine. The article compares the over-
load theory to the collapse of the Ireland Building and quotes the Fire Department’s
Chief Croker as stating that “complaint after complaint” had been ﬁled with the Build-
ing Department during the preceding months for faulty construction. There was a coor-
dinated eﬀort by various city agencies to assign blame by assessing the collapse for spe-
ciﬁc technical causes, and the Times assumed that the reading public had enough un-
9 of 15
derstanding of tall buildings to know the context of the Ireland collapse and the dangers
of speciﬁc poor practices.
One of the disturbing details found was the extreme ﬂexibility of the frame. Since the
exterior walls were meant to provide lateral bracing to the building, it is no surprise to
hear that the frame wobbled with iron erection having so far outpaced masonry, but re-
ports that the frame was 18 inches out of plumb when erected to the 9th ﬂoor and had
to be pulled back erect with a block and tackle, or that it swayed noticeably in the
breeze despite the lack of signiﬁcant sail area, raise the question of damage from sec-
ondary stresses. (“Concerning the Fall...,” 1904; “Further Notes...,” 1904)
Over the next few days, the public was given detailed explanations of the pre-collapse
Building Department violations and the methods used to cure them as well as descrip-
tions of the diﬀerence between inadequate construction (where the fault would lie pri-
marily with the contractor) and inadequate design (where the fault would lie primarily
with the architect and engineer). In both cases, there would be secondary blame on
those responsible for review: the building department inspectors and, for inadequate
construction, the architects and engineers. Interestingly, the ﬁrst technical opinion
stated by an oﬃcial – that the building code had been violated by the engineers through
the use of too-small ﬂoor beams and too few bolts between beams and columns – came
from an Assistant District Attorney, not one of the experts on construction. His logic
was simple: “But it is all foolishness for any one to think that there was nothing wrong
when a building like that falls in on itself.” (“Six More Bodies...”, 1904; “Darlington
Owner Held...,” 1904; “Darlington’s Builder Surrenders...,” 1904) With its emphasis on
the failure of a common type of building, that is a statement of betrayal of the lay-per-
son by technology.
The formal results of the coroner’s inquest were mixed. Unlike the Ireland jury nine
years earlier, experts were consulted by the city agencies to ensure that technical infor-
mation was not reviewed unaided by jurors and lawyers. In addition to technical discus-
sion, the coroner’s jury heard testimony concerning confusion over the meaning of ar-
chitectural and engineering consultation. Guy Waite, a consulting engineer formerly
employed by the Building Department, described his refusal to reduce the size of ﬂoor
beams smaller at the owner’s request. Julius Tomiek, an employee of the general con-
tracting ﬁrm Pole & Schwandtler and an immigrant who testiﬁed in German, stated
that he had no contact with the city oﬃcials or direct knowledge of the work performed
on site, and that he had been told by his employers that the plans (showing the lighter
ﬂoor beams Waite thought incorrect) had been approved by the building department.
Similar testimony followed, focusing to a large extent on the lack of continuity among
the designers. One juror was quoted saying that “If there is a system by which one ar-
chitect comes into a job and starts and stops and then another and another starts and
stops so that no responsibility can be ﬁxed, we want to know it, so that the laws may be
amended so as to put an end to it.” This statement is technically an attack on architects,
but the legal blame was found to be the owner who hired and ﬁred the architects. The
10 of 15
jury found the owner, Eugene Allison, and Pole & Schwandtler guilty of criminal negli-
gence and recommended two new laws: that only the architects responsible for a design
or experienced contractors be allowed to supervise construction, and that the building
department hire engineers as inspectors of “all buildings requiring engineering skill.”
Four jurors added as unoﬃcial recommendations a requirement for licensing all archi-
tects and engineers working in the city. (“Darlington’s Builder Surrenders...,” 1904;
“Warned Allison...,” 1904; “Darlington Verdict...,” 1904; “Architects Condemn Law,”
1904) The District Attorney indicted Allison but apparently never tried him, saying
that he had disappeared. In 1911, Allison (who had been in new York the entire time)
had the charges dropped on account of his poor health. (Engineering News editorial,
The separate investigation conducted by the district attorney’s oﬃce was quite simple
in concept: Harry de Berkeley Parsons, a leading consulting engineer of the time, was
hired to investigate the wreckage and drawings to see if the design and construction vi-
olated any laws. Parsons’s conclusion was that in a narrow sense the failure was the re-
sult of poor beam-to-column connections that allowed the columns to buckle sideways
without restraint, but that in a broader sense the cage frame type was inherently
ﬂawed. This was a signiﬁcant conclusion for engineers, but his statement that the col-
lapse was the result of “faulty design and carelessness and neglect in the erection of the
members” was vague in a legal sense, allowing the designers, the contractors, and the
building department inspectors to share the blame. (“Darlington Disaster Report,”
1904; Parsons, 1904) The existence of his report was itself a statement of the profes-
sional method of conducting forensic engineering investigations, regardless of whether
that was perceived; the report was quoted in the popular press and an abbreviated ver-
sion published in both the Proceedings of the American Society of Civil Engineers and
the Engineering News.
Parsons was quite speciﬁc in his report, identifying a column near the center of the
building as the ﬁrst to fail. He explained how this argued against various other causes
that had been suggested, including the explosion of the boiler for the on-site steam en-
gine used for hoisting, column overload caused by storing plasterer’s gypsum and
unerected iron and steel on the topmost ﬂoor, and foundation movement. (“Further
Notes...,” 1904; “Concerning the Fall...,” 1904) Interestingly, at least one more recent
engineer, Jacob Feld, is of the opinion that the trigger for the failure was that some of
the column footings rested directly on rock while others were on clayey soil. (Feld,
1968, p. 42) However, this analysis focuses on the trigger for the speciﬁc failure rather
than the general cause; had the same foundation conditions existed below a structural
steel frame, any resulting failure would have been limited to settlement and minor
member deformation. Similarly, pre-collapse sway may have cracked the column-splice
ﬂanges, but this does not constitute a separate overall cause of failure as much as rein-
force the case against the use of cast iron and cage frames.
11 of 15
Various problems that may have led to the collapse – overloading of the top ﬂoor, cracks
in the columns caused sway and truing of the frame, the possibility of an isolated foot-
ing failure – are unrelated in their origins, but share a common theme: none of them
could cause catastrophic failure of a steel skeleton building. The complete loss of the
Darlington, regardless of the importance attributed to the diﬀerent causes in forensic
investigation, was the result of an inadequate structural system – one that had no re-
silience to prevent progressive failure when subjected to temporary overloading, iso-
lated connection failure, or diﬀerential settlement.
The shortcomings of the cage building system were known and had been described in
detail in the engineering press before the collapse. Prominent engineers in the 1890s
and 1900s criticized the cage frame, the use of cast-iron columns, and the use of inte-
grally cast connections – for example, William Sooy Smith describing in 1902 the use of
cast iron “ﬂimsy lug and bracket connections” and eccentric loads on cast-iron columns
as “radical defects” – but had only managed to eliminate the use of cast iron and cage
frames from the commercial buildings where they were regularly employed. In other
buildings, particularly hotels and apartment houses, where consulting engineers were
rare, the older and less safe construction types persisted. (Sooy Smith, 1902; “Current
In that sense, this disaster was foreseeable and preventable and hardly seems worthy
of forensic analysis. However, the shock of the utter failure of a formerly popular build-
ing type, the shock of the great loss of life, and the sense that structural engineers, as a
group, had allowed an unsafe type to continue in use combined to exaggerate the impor-
tance of this one incident. It is diﬃcult overstate the importance of the Darlington col-
lapse when sources as knowledgeable as the builder William Starrett and the steel ex-
pert Robins Fleming later cited it as the cause for the ﬁnal abandonment of cast-iron
structure. (Starrett, 1928, p. 41; Fleming, 1934)
One of the articles in the engineering press following the collapse expressed the engi-
neers’ view clearly: “Steel skeleton...construction...gives us a rigid steel framework con-
nected and braced like a bridge in all directions, carrying walls, ﬂoors, and ﬁnish
merely as so much clothing, and able to safely bear all loads which come on either the
completed building or the framework alone. Structures of the ‘oﬃce- building’ type are
practically all of this kind. But very many other buildings, from ﬁve or six-story ware-
houses and stores to twelve (and in at least one case seventeen-) story apartment build-
ings, are radically diﬀerent in structural respects, though masquerading under the gen-
eral name of ‘skeleton’ or ‘steel-frame’ construction. They have ﬂoor beams and girders
of steel, columns of either steel or cast-iron designed for vertical, central loading only;
they are provided with no bracing, and the connections between the beams and columns
are unﬁtted to resist any calculable bending or twisting moments. In such buildings the
walls and ﬂoors give lateral strength and stiﬀness to the structure, while vertical loads
12 of 15
are carried by the frame. Evidently this is quite diﬀerent from the condition in an oﬃce-
building structure, both after and during erection. Of course, it is known that when the
walls are in place and ﬁrmly set, they supply suﬃcient lateral resistance, though the
amount of this resistance cannot be calculated, especially when the wall is pierced by
windows in every panel.” (“Concerning the Fall...,” 1904) In other words, even though
cage buildings might be generally safe when the walls were completed, they were inher-
ently inferior because they contained undesigned structure. Engineers were looking for
more responsibility, not less: the owner has the responsibility “to employ competent ar-
chitects and engineers to design and supervise his work, and honest contractors to exe-
cute it; and it is with those parties that the real responsibility for safe work must lie be-
cause they do the work.” [Emphasis in original.] (Engineering News editorial, 1904)
An informal result of the Darlington collapse seems to have been the end of cast-iron-
column cage construction. By 1904 this type had been technically obsolete for at least a
decade and it was absent from commercial buildings because the professional engineers
involved with them no longer used cast iron. As William Starrett wrote in his memoirs,
“the suspicion [of cast iron] was ruinous,” leaving steel frames as the only acceptable
new technology for most buildings. (Starrett, 1928, p. 41) In other words, the public
perception of new building technology had caught up with engineers in distinguishing
between cast iron and steel. Public concerns over safety and liability put an end to cast-
iron use in a way that engineers’ concerns over the technical issues of safety had not.
The cage frame, which was associated with but not dependent on iron columns, disap-
peared from use at roughly the same time; steel- frame buildings were generally free
from a reputation for catastrophic collapse. This collapse marked the end of acceptance
of unanalyzed structure.
The lessons from the Darlington were available after the Ireland collapse, but the con-
text was diﬀerent. In 1895 there were only a handful of steel-frame buildings in New
York and Chicago, often visually indistinguishable from their cast-iron cage- frame
neighbors. By 1904, steel-frame construction, identiﬁed as such and visible during con-
struction, was an accepted part of the cityscape and had spread across the country. The
diﬀerence between this technology and its predecessors was dramatized by such build-
ings as the Fuller (Flatiron) Building at the triangular intersection of Fifth Avenue,
Broadway, and 23rd Street; the combination of a 285-foot height and a triangular plan
that narrows to a 6-foot apex was not achievable using cast-iron technology and created
a dramatic appearance that heightened public awareness. Compared to the increas-
ingly-tall steel-framed buildings, the design of the Darlington was seen as backward,
even though that technology was itself only thirty years old. Because the transition to
modern technology was largely complete by the time of the Darlington collapse, the
press coverage focused on process – who was responsible? – and regulation of how to
encourage the use of the best technology available.
13 of 15
“Architects Condemn Law,” New York Times, March 4, 1904, p. 3.
Bijker, W., Hughes, T., and Pinch, T., editors (1987) The Social Construction of Techno-
logical Systems, The MIT Press, Cambridge.
“Bodies All Taken Out,” New York Times, August 15, 1895, p. 9.
Bogardus, J. (1856) Cast Iron Buildings; Their Construction and Advantages, J. W.
Harrison, New York.
“The Collapse of the Darlington Apartment House in New York City,” Engineering
News, March 10, 1904.
Condit, C. (1982) American Building, 2nd edition, University of Chicago Press,
“Concerning the Fall of the Darlington Building,” Engineering News, March 24, 1904.
“Current Practice in Apartment House Construction in New York City,” Engineering
News, April 14, 1904.
“Darlington’s Builder Surrenders to Coroner,” New York Times, March 17, 1904, p. 7.
“Darlington Disaster Report,” New York Times, March 20, 1904, p. 20.
“Darlington Owner Held by Coroner,” New York Times, March 5, 1904, p. 3.
“Darlington Verdict Holds Three Guilty,” New York Times, March 23, 1904, p. 1.
“Death in Collapse of Hotel Skeleton,” New York Times, March 3, 1904, p. 1.
Engineering News untitled editorial, March 17, 1904.
Engineering News untitled editorial, June 15, 1911.
Feld, J. (1968) Construction Failure, John Wiley and Sons, New York.
Fleming, R. “A Half-Century of the Skyscraper,” Civil Engineering, December 1934.
Freyer, W. (1967 ) History of Real Estate, Building, and Architecture in New
York City, Arno Press, New York.
“Further Notes on the Collapse of the Darlington Building,” Engineering News, March
“Imprisoned in a Wreck,” New York Times, August 9, 1895, p. 1.
“The Investigation of Building Failures,” Engineering Record, 9/14/1895.
“Ireland Building Inquest,” Brooklyn Eagle, August 21, 1895, p. 1.
“Mr. Constable Inspects,” New York Times, August 13, 1895, p. 6.
14 of 15
Parsons, H. De B., “Collapse of a Building During Construction,” Engineering News,
May 12, 1904.
“Six More Bodies From Hotel Ruins,” New York Times, March 4, 1904, p. 3.
Sooy Smith, W., “The Modern Tall Building – Corrosion and Fire Dangers,” Cassier’s
Magazine, May 1902.
Starrett, W. (1928) Skyscrapers and the Men Who Build Them, Charles Scribner's
Sons, New York.
“Steel Frames,” New York Times, March 4, 1904, p. 4.
“To Go Before the Grand Jury,” Brooklyn Eagle, August 30, 1895, p. 12.
“Visited the Ruins,” Brooklyn Eagle, August 23, 1895, p. 10.
“Warned Allison of Darlington’s Peril,” New York Times, March 18, 1904, p. 14.
15 of 15