In a tiny garden in Boston petals fan out on tiny flower stems. Curved and delicate, row and rows of complex flower shapes form a feast for the eyes, spiraling like the tubes of a French horn and twisting round like finely blown glass.

Except the scene is not exactly in a garden; it's in a laboratory. And it's not exactly flowers; it's crystals being grown into flower-like shapes. And you can't exactly see it unless you have an electron microscope, because the garden is only microns in size, the nanoscopic crystal "garden" was grown at a Harvard lab.

The nanogarden shows great promise for scientists trying to better understand how complex shapes like those seen in flowers evolved in nature.

Researchers found they are able to adjust chemical gradients and manipulate environmental factors like temperature to get the crystals to form precisely-tailored structures like the flower shapes, rather than the jagged forms crystalline structures usually take.

"For at least 200 years, people have been intrigued by how complex shapes could have evolved in nature. This work helps to demonstrate what's possible just through environmental, chemical changes," said Wim. L. Noorduin, lead author of a paper published as the cover story for the May 17 issue of the journal Science.

Noorduin and his colleagues have grown the crystals on glass slides, metal blades and on a penny.

"When you look through the electron microscope, it really feels a bit like you're diving in the ocean, seeing huge fields of coral and sponges," said Noorduin in a release from Harvard. "Sometimes I forget to take images because it's so nice to explore."

The technique to grow the crystals is relatively simple. Chemicals common to most labs are all that's required. Mix the ingredients in a beaker that has a salt and silicon compound dissolved in it and place a drop on a metal surface that acts as a "soil" for the "plants" to grow. Carbon dioxide form the air naturally dissolves in the solution triggers a simple reaction that causes the crystals to form in intricate curves, rather than jagged edges. Noorduin found that increasing the concentration of carbon dioxide, for instance, will create "broad leafed" structures. Reversing the pH gradient in the solution at the right moment will create curved, ruffled structures.

"You can really collaborate with the self-assembly process," said Noorduin. "The precipitation happens spontaneously, but if you want to change something then you can just manipulate the conditions of the reaction and sculpt the forms while they're growing."

Noorduin's Harvard colleague, Joanna Aizenberg, a professor of chemistry and biology, said their approach is to study biological systems, focusing on what those systems can do that we can't, "and then to use these approaches to optimize existing technologies or create new ones."

"Our vision really is to build as organisms do."