Modern engineering techniques and physics equations unveiled the deadly phenomenon of the Great Molasses Flood, a molasses wave that smothered part of Boston nearly a century ago.

On January 15, 1919, 2.3 million gallons of molasses burst from a steel holding tank in the North End of Boston, Massachusetts. This sugary tsunami claimed the lives of 21 people and injured about 150 more. Some victims were killed when the molasses wave destroyed a city building, collapsing the structure on the workers eating their lunches within, according to a New York Times article from 1919.

“A dull muffled roar gave but an instant’s warning before the top of the tank was blown into the air …Two million gallons of molasses rushed over the streets and converted into a sticky mass the wreckage of several small buildings which had been smashed by the force of the explosion,” said the New York Times 1919 article.

Engineering students at Harvard University got their fingers sticky in a research project, initially designed as a final project for Dr. Shmuel M. Rubinstein’s Introduction to Fluid Dynamics course with the goal of exploring why the wave of molasses, a seemingly non-threatening material, proved to be so deadly when it erupted from the split tank.

The team discovered that the molasses had recently arrived from the Caribbean, and the molasses was likely slightly warmer than the Bostonian air. The molasses hardened quicker because of this temperature difference, trapping its victims. “Men and women, their feet trapped by the sticky mass, slipped and fell and were suffocated,” according to a retrospective article by The Boston Globe in 1968.

Some braved the molasses tide in attempt to save those who had been trapped, but a number of the rescuers paid for their heroism with their lives, according to the New York Times. The cause of the tank’s collapse remains a mystery for civil and structural engineers today. Ronald Mayville, a structural engineer who has studied this phenomenon for years, said that the rupture was probably due to a poorly built tank. With wartime pressures overseas, the tank’s builders were likely pressed to complete the tank quickly and never returned to inspect it before putting it to use in 1915, Mayville told The Boston Globe.

“The steel itself was prone to cracking,” said Mayville. The low manganese content of the steel caused the alloy to be brittle; this same kind of steel was used on the Titanic. While “the steel conformed to the standards of the time,” said Mayville, its manganese content falls far short of modern regulations. “Now it’s known you need to have a higher ratio,” Mayville said.

Using modern engineering software, Mayville says that the crack likely began at the 20-inch manhole. Today, a manhole is reinforced to decrease stress, he says, but in 1919, the manhole was not reinforced properly. Mayville’s hypothesis reflects the description from 1919: “The circular wall broke into two great segments of sheet iron which were pulled in opposite directions,” stated New York Times.