The Pika Don Story: The Art of the Do-Over in Landscape Painting


Landscape painting of Hiroshima mushroom cloud by Mitchell AlbalaUnlike so many things in life, painting often gives us the chance for a do-over. We can scrape out what we’ve done or just start again. We learn something from that second (or third, or fourth) attempt that helps make our next effort better. As frustrating as that may be, it’s how we learn. In this post, I’d like to take you through the story of one of my own do-overs.

In 2014 I staged an exhibition called Acceptance. This series was a departure for me. As SeattleMet described it, “Mitchell Albala’s latest collection of works takes a dark turn. While still drawing heavily on soft, semi-abstract horizons, these paintings focus on historical human catastrophes … It’s beauty to ease the horror, and horror to make the beauty unsettling.”

In Japanese the atomic bomb detonation was described as a brilliant light, pika, followed by a thunderous blast, don.


At work on Pika Don, December 2013.

One of the paintings in the show, Pika Don, depicted the Hiroshima mushroom cloud. I wanted to make the mushroom cloud look like a regular cloud, but infuse something into it that would give it an unsettling quality — some clue as to what it actually was. My solution was to use saturated colors, to perhaps suggest the heat and radioactivity of the blast. That might be both beautiful and disturbing, which was the paradox I intended for each painting in the series. But I was never as satisfied with Pika Don as I was with the other pieces in the series.

Ordinarily, once I declare a painting finished, I move on. If it’s less than perfect, I let it stand as a statement in time, a record of my artistic history. Not so with Pika Don. After the exhibition the painting came back to my studio and sat in the corner for nearly three years. One day I pulled it out and was newly impressed by just how much I disliked it. The color seemed unnatural and the cloud itself — which can be one of nature’s most structural forms — looked flat. I was ready to sentence the painting to the dumpster (sometimes a necessary purging), when I thought, Why not rework it? What did I have to lose?

As I was starting the revisions, my friend and colleague Michael Stasinos stopped by the studio. I showed him the painting and the archival photograph I was using as reference. I expected him to comment on the color, but he noticed a compositional problem I had missed.

LEFT: The archival photo of the Hiroshima mushroom cloud, indicating the section that would form my composition. RIGHT: In the first version of Pika Don I made the mistake of following the photo too closely. At the time, I thought that some historical accuracy was important, so I tried to mimic the shape of the cloud in the photo. The column of the cloud tilted slightly, but along the left side, it left an empty rectangular space. By carrying that into the painting, I created an inactive shape that was not integrated into the rest of the composition.

Landscape painting of Hiroshima mushroom cloud by Mitchell Albala

Pika Don, 2016, oil on canvas, 26 x 20. In the revised Pika Don, the color is slightly less saturated. I also added more midtones (shadows) in the cloud to suggest volume and make it appear less flat. Compositionally, I exaggerated the twisting thrust of the cloud to create more movement. The sky shapes on the left and right undulate and move in and around the cloud, making the “background” more integrated with the composition.

In a sense, every painting is like a do-over. Each painting, each effort, each mistake can teach us something that we can apply to our next painting. When working from photos, I encourage my students to depart from the photo as needed. “A painting answers only to itself.” With Pika Don, that was advice I needed to take myself. By departing from the historical “truth” of the photo, I was able to create a stronger composition and a better painting.

Additional Resources

Using Photos Like an Artist

Reviews: Albala’s “Acceptance” Exhibit Takes a “Dark Turn” and “Transcends Painterly Realism”

“Acceptance” Series on Exhibit at Lisa Harris Gallery

Video: The “Acceptance” Series as Commentary on Human Consciousness


About Author

Mitchell Albala is a Seattle-based painter known for his semi-abstract and atmospheric landscapes. His book, "Landscape Painting: Essential Concepts and Techniques for Plein Air and Studio Practice," is a national bestseller with nearly 40,000 copies in print. Mitchell is also a popular workshop instructor at Gage Academy of Art in Seattle, Pacific Northwest Art School, and Daniel Smith Artist’s Materials. He led painting adventures in Italy in 2015 with Arte Umbria and in 2017 with Winslow Art Center. He has lectured on Impressionism and landscape painting at the Seattle Art Museum and written for International Artist and Artists & Illustrators magazines. His popular painting blog, which serves as a companion to his book, was awarded #12 on’s Top 75 Painting Blogs.


  1. Hi Mitch,
    I have found that re-working paintings has become a very good way for me to jump start painting. I now have on the easel a 10″ x 20″ painting. I do not know how many paintings are buried underneath. I usually paint in acrylics so I do not have to be overly concerned with the drying process as I would with oils. This particular canvas, a Fredrix Red Label is from years ago. Fredrix used to trim the excess of these canvases. I am curious if so many layers of paint will help pull the canvas away from the edges even more than what usually happens. When choosing a canvas to re-work I ask myself if I would ever decide to exhibit that painting. If not I consider it fair game. Some canvases such as the 10″ x 20′ I stay with the overall theme but extensively re-work. Others though receive a coat of gesso or acrylic paint. One of the problems with prepared canvases either from a roll or already stretched is that sometimes the texture of the previous painting cannot be eliminated. No matter how many times I may sand the canvas the texture cannot be eliminated so I try to incorporate it into the new painting. There is only so much sanding down that can be done even when using a fine sandpaper. Sometimes using gel medium or a very light modeling paste I can coat the canvas so that surface is more or less even.

  2. Mitchell Albala on

    With the painting I wrote about in this post, my revision was more or less on top of the same subject. The paint that was there from the first version would not interfere (texture-wise) with the newer version. What you describe is many, many layers, even re-gessoing the surface, wherein the texture of the previous layer cannot be eliminated. That is something I try to avoid because each painting generates it’s own texture. I don’t want that texture to be incompatible with the energy of the new painting. If Pika Don had a lot of texture that didn’t coincide with the revised painting, I would not have worked over it; I would started over and abandoned the first version altogether. Sometimes that’s necessary, too!

  3. Mitch, I truly enjoyed reading this article and hearing your interpretation and reasons for changes. Though I’m able to rework soft pastel paintings more easily and frequently do, there is often a knee jerk reaction to take it to task straight away. Lately I’m trying a slightly different approach. I do a notan study of the final piece to see how true I was to my initial sketch. This may lead to a correction or often leads to an exciting new interpretation for a second piece.

  4. Mitch,
    Thanks for another excellent presentation of your craftsmanship. You are also demonstrating that in making changes for one reason, in your case the offending rectangle and flatness of the clouds, one winds up making changes throughout the painting. I don’t think I would have ever thought of undulating the sky right to left to create a twisting motion as well as seeking a twisting motion to the cloud. Please elaborate on what you did to create the twisting motion as it seems highly complex, with shifts in chroma, temperature, contrast and more and I am interested in how you thought of the direction of twist, sky vs. cloud, and sideways vs. vertical. The change I find particularly significant is the ground plane which I see in the revision as a much lowered horizon creating greater sense of scale and height to the cloud. I also find the revision makes the cloud seem much taller, taking a look down plus a look up to view. That is a powerful device to create a sense of height, one I have recognized in mountain paintings, but never I never thought to apply it to clouds.