Solar Cat Steam Rocket Development: Part 1 (2020-2022)

by Bill inman, Reaction Research Society


During, and even before my first steam rocket, the “Scalded Cat”, first launched in 2000, I had contemplated the idea of heating the water with concentrated sunlight.  Also, in the years after the Scalded Cat, I became convinced that, as impressive as it was, it’s highest flight of 4660 feet still left room for improvement.  My late wife Carmela and I had discussed trying to combine an improved performance rocket and a solar heated rocket into one project and launching it on the 20 year anniversary of Scalded Cat’s first flight in December, 2000.  She was enthusiastic about the idea, but in 2019, died before it could happen, and I regretted her not living to see it.  As a new widower considering my options, I decided on this project for my next chapter in life.

Part 1 of the report on this continuing project covers the early development of a parabolic trough concentrating solar furnace to heat the water in a future steam rocket, and the results of the first “Solar Cat” rocket.


 Develop the knowledge and experience through hands-on design, fabricating and testing to eventually build and launch a 100% solar powered steam rocket with dramatically higher performance than the original propane heated “Scalded Cat” flown at the Reaction Research Society’s Mojave Test Area (MTA) in the years 2000, 2001 and 2002.

Some of my good friends have been a tremendous help along the way.  In addition to myself, “Team Steam” primarily includes (in alphabetical order):  Kime King-PatrawDave McKinnonKeith & Nadara SoulesDale Talcott, and Jon Wells

Launch of the Scalded Cat on it’s first flight, December 2, 2000.

Some of my friends and I have had an interest in solar energy since high school back in the 1970s, particularly in the heat potential of concentrating solar furnaces.  As such, we built several “science project” type models over the years – one reaching 600o F. – more than enough for a steam rocket.   Combining steam rockets with concentrated solar heat seemed like a good way to pursue both these interests simultaneously.

Receiver design with an upper window and two diagonal mirrors:

Before building the parabolic reflectors, I wanted to test an idea for the receiver since it will double as the launch pad and radiant heat collector. The receiver must be wide enough for the rocket fins to pass through for launch, it will cast a larger shadow than otherwise.  I hated losing that area of potential solar heat, and got the idea of trying to capture that otherwise lost sunlight from the top, with a window and  2 diagonally opposing mirrors inside, reflecting it onto the receiver tube.  The first device tested on October 3rd and 4th of 2020 was just a particle board box of this design.  (see cross sectional drawing below)

By itself, this receiver box was underwhelming, unable to even boil water in four hours of full sun.  Undoubtedly, there was significant heat loss on both ends of the receiver tube where metal fittings, valves & instruments protruded outside the box.  Wrapping those areas with makeshift insulation on the 2nd day’s test helped retain some of the otherwise lost heat, but it still fell short of expectations.

Testing the receiver box alone this way couldn’t tell us much though, since in this form, it was really just a small batch type solar water heater.  The real test would be comparing this design against one with the upper part just insulated instead,  and heated only by a parabolic reflector through the bottom window.

First of four pictures shows the 2” galvanized pipe receiver tube freshly painted with Thurmalox solar collector selective coating

Second photo of four shows the receiver box with the receiver tube and the two diagonal glass mirrors installed.

The third photo of four shows the reciever with an upper glass window, mounting yoke, instruments and a drain valve.  The receiver tube ends double as pivoting points for manually turning it to follow the sun

The fouth and final photo of the series shows the reciever assembly as it was set up at the RRS MTA on October 3, 2020 initially facing away from the sun before starting it’s first test.

The following graph shows the heating rate starting slower on the second day (blue line), but then picking up and passing the rate of the first day.  This was probably due to the start times.  The sun was lower in the east at the start on the second day and heating picked-up at mid-day.  

On the first day however, the sun was high at the start, but soon sinking into the west resulting in the heating rate dropping off after a period of two and a half hours.

The plot of temperature rise during the two tests on October 3rd and 4th of 2020.  The ambient air temperature was around 95 Fahrenheit.

Adding a four-foot span parabolic reflector focusing through the lower window:

This configuration was needed to actually test the difference in temperature achieved with and without using the top window and diagonal mirrors.  Not surprisingly, this version was able to produce higher temperatures than the receiver box alone, hitting the boiling point with ease.  However, when the use of the internal diagonal reflectors was compared with blocking the sunlight to them with fiberglass insulation, no discernible difference in heating was apparent.  And with larger parabolic reflectors planned, it seemed likely that the higher temperatures would result in larger radiant heat losses through the top window, exceeding any potential gain.  Therefore, this idea was abandoned.

​Adding the 4 feet of parabolic reflector involved enlarging the “U” shaped yoke to accommodate the corresponding increase in width and depth.  Trying to save the materials and work that went into the receiver box while continuing to use mostly surplus lumber for the project, still mostly particle board, yielded less than ideal results.

​However, we were now on to testing the outputs of the parabolic trough solar furnace, the actual objective.  The first test was attempted on November 7, 2020 at the RRS MTA, but clouds rolled in, blocking the sun until it was too low in the west to aim at.  So our first actual readings were taken with the 4 foot reflector back home in Carson City two days later.  

This version exceeded 300o F. twice, with a best of 371o F in 4 hours.  In both those tests, it reached 300o F within 2.5 hours with full sun.  In the other tests, clouds interfered, yet it still reached 267o F in one of them.  In all, 8 tests were conducted in this configuration, ending on November 23rd.

We also learned that heating was still possible in thin overcast conditions as long as the clouds were thin enough to allow the sun to create shadows with crisp edges.  But the heating rate was reduced, as was the stagnation temperature.

First of four photos shows the cutting out the parabolic curve in particle board.

The second of four photos shows the checking the focal spot on the ground with the reflector attached as seen by the intersection of the two bright lines.

Third photo of four shows the solar furnace set up but sitting idle at the RRS MTA on a cloudy day – November 7, 2020

Fourth of four photos shows another view of the assembly sitting idle at the RRS MTA.  Notice the absence of dark, crisp shadows due to clouds.

Same setup but with a larger reflector with two 3-foot spans:

​A primary goal of these steps of the project was to determine how hot the water in the receiver tube will get – and how fast it will get there vs. the size of the reflector, in both square footage and span (concentration ratio).  A number was derived from the tests with the receiver box alone, and now also from the 4 foot span reflector.  Of course, a foot wide swath right down the middle was shaded by the receiver box, so that area had to be subtracted from the total square footage.  By replacing the single 4 foot wide reflector with two reflectors, each 3 feet wide, a comparison was possible.  Also, the two reflectors were moved out from the center a few inches, reducing the amount shaded by the receiver box.

​Again, this particle board contraption was enlarged by adding about a foot to each end of the parabolic end boards to support the wider reflector.  Reinforcements were added to the yoke as well, because unlike the previous version which was hauled to the MTA disassembled in the back of my 2000 model year GMC Jimmy and re-assembled on site, the new 6 foot span and yoke would not quite fit.  Instead, it was rebuilt to sit on a platform on a utility trailer towed behind the Jimmy, but still transported disassembled.  Both the yoke and the reflector assemblies had also become so heavy and ungainly that it was getting difficult to assemble and disassemble without assistance.  Dimitri Timohovich offered to assist several times, and was a big help on those occasions.

New idea to add nozzle and fins to launch a demonstration rocket with the receiver:

Now, my late wife Carmela and I had had the idea to launch the new solar powered steam rocket on the 20th anniversary of the first successful Scalded Cat launch – the RRS December 2000 event.  It was looking like a lost cause until the best temperature and pressure achieved with the 4 foot reflector almost touched what we had considered the minimum needed for a steam rocket.  With the new, larger reflector, it should only do better.  By then, we were fast approaching the end of November, but what if?  … Could there be any possible way to do it that quickly?

Well, I had built and launched a couple of “Simple Cat” steam rockets starting in 2006, proving that a very simple rocket based on 2-inch nominal galvanized pipe with “hardware store” fittings and accessories, and released by a manual “pull cable” could actually fly.    One even reached approximately 1000 feet of altitude according to MaryAnne Butterfield’s calculations from it’s flight time.   And the receiver tube in this solar furnace experiment was already none other than 2” galvanized pipe!  Could the receiver box be made into a launch pad, and the receiver tube made into an actual flight vehicle?  – a solar heated “Simple Cat”?  That might just be the one chance we have to do this in under a month, making a December, 2020 launch possible!

​The first of two photos by David Allday shows the author by the “Simple Cat” in the tower on December 2, 2006, at the RRS MTA.

The second of two photos by David Allday shows the launch of the Simple Cat from the Large Vertical Test Stand at the RRS MTA on December 2, 2006.

The team’s decision was, “go for it!”  There was already a launch set for December 12th at the MTA, so that would be the target date – the official RRS December launch 20 years after our first steam rocket flight!  The fins and nozzle off the old Simple Cat were installed, and thanks to Dale Talcott’screativeness and fabrication skills, fin slots cut into the upper (north) end of the receiver box, and a launch tower added during a marathon session the weekend before the launch!  Provisions were also made to tilt the assembly far enough “south” to raise the tower & rocket vertical for launch.  Racing the clock, a lot was hurriedly done and not fully tested in the final form, if at all!   At the MTA on the 12th, we discovered that the newly reinforced reflector assembly was then too wide to fit into the newly reinforced yoke!  – OOPS!!!   With the quick thinking of Dave McKinnon and Dave Nordling, the yoke was hastily widened thanks to Dave McKinnon’s skills and cordless tools he had brought with him to the MTA.  But by the time we got it ready to start heating, the sun had moved behind clouds in the western sky.  – No 20th anniversary launch.

Or might there be…?   Back home in Carson City, Nevada, Kime King-Patraw pointed out that we still had nearly 3 more weeks left in December, and that launching it at ANY time in December of 2020 was still pretty close to the 20 year mark!  It wouldn’t be at an official RRS event at the MTA, but you can’t always get everything you want!  Some additional work was done, and heating tests were conducted on the 18th, 19th, and 20th, with the highest temperature reached being only 300o F.  All three were done in my backyard though, where shading from the bare trees and/or clouds still hampered the effort.

On December 21st, full of optimism, Kime and I drove to Tonopah, NV near where an undisclosed launch site had been chosen.  This area seemed fitting, as the 110-megawatt Crescent Dunes concentrating solar-thermal molten salt powerplant was also in the “rough vicinity of Tonopah”.   Jon Wells drove up from Las Vegas to complete our trio.  

The morning of December 22nd dawned completely cloudless in the high desert sky.  We drove out to the launch site early and set up.  Heating went quickly, but as pressure built, a drip started in the nozzle plug assembly.  We aborted heating, inspected, then cleaned & re-tightened the fittings involved, then re-filled and started the heating again.  Again, it started to drip once it started building pressure.

​By then, we didn’t have time to go back to town & look for a new fitting, but since the temperature and pressure continued to climb, even with the leak, we decided to just keep heating and see what happened.  As usual, the drip got worse as the pressure increased, and as it approached launch pressure, we started fearing it was almost out of water, too.  – So, knowing we’d need to release pressure and drain it anyway, and with nothing really to lose, we attempted a launch.  After a few tries, the stubborn plug/clamp finally released and it took off, becoming what just might be the world’s first ground launched, completely solar powered rocket!  – Nevermind that it was only at 175 psi and only went about 20 feet high on it’s half a pint of remaining water – it still cleared the tower!  – Hey, Dr. Robert Goddard’s first liquid rocket flight only went 41 feet high!

First of three photos shows the trailer mounted solar receiver at the RRS MTA on December 12, 2020. The only problem was no sunlight.

Second of three photos shows Kime King-Patraw and Jon Wells comparing notes during the heating experiment on December 22, 2020 conducted in the general area of Tonopah, Nevada.

Third of three photos shows the author conducting the first launch of a completely solar-powered, ground launched steam rocket just as it clears the launch rails.

Developing a motorized tracking system:

By manually adjusting the solar furnace to keep it targeted on the sun, it was necessary to move it and re-clamp it every 3 to 4 minutes.  This required someone physically being there to make repeated adjustments at that frequency.  The RRS pyrotechnic operators in charge took a dim view of this, and strongly advised us to come up with a remotely operated sun tracking method.  We had also been foiled in our tracking efforts at least a couple of times by other peoples’ projects needing us to take cover in the bunkers during their fueling, firing and/or troubleshooting.  This sometimes caused sun tracking to stop for an hour or more.

​I had once built a drive gear for a telescope mount with a clock-rated a/c gearmotor, so built one for this solar furnace using the same motor and concept.   Of course, this resulted in more pieces scabbed onto the monstrosity that was never originally designed for it, adding complexity, weight, and more potential problems.  In the end, this tracking gear worked … sort of…  We intentionally designed it to run “fast”, thinking it would be better for it to “outrun the sun”, allowing us to just momentarily stop it (with an “on-off” switch in the blockhouse) long enough for the sun to catch up, than risk it “running slow”, requiring approaching it to re-set it ahead by hand.  Again, that worked too … sort of…  It ran so fast that it had to be stopped for the sun to “catch up” about every 20 minutes.

​There were several other problems with it binding, hanging-up, and engaging/disengaging the gear teeth.  The rough road to the MTA broke the actual motor mounts one time, requiring a hasty field repair by Jon Wells and Dave McKinnon.  It had also become painfully obvious that the mount’s range of east-west motion was another limiting factor.  We never got set up too early to aim at the morning sun in the east, but there were several times we reached our western limit before the heat and pressure were sufficient, and the sun was still high enough that we could have kept going if only it was able to rotate farther west.  This wasn’t a fault in the tracking gear, but in the design of the mount itself.

​Along with a couple more unsuccessful launch attempts, mostly due to clouds, we still managed to make three more flights while also recording additional heating data.   It’s second flight was on March 20th at the MTA, again only reaching about 25 feet after more problems, including the broken tracking motor mount.  Dave McKinnon and Jon Wells both came again to assist on that attempt, Dave recording the heating rate notes, this time.  The 3rd flight was also it’s best, reaching an estimated 400 feet on April 1stat another undisclosed site, this time in “the rough vicinity of Lovelock, NV” where Dale Talcott and his brother Dave were able to attend.  Everything went right that time, other than a breeze and some high, thin clouds building, eventually slowing the heating rate.  My estimated potential of this fat slug of a rocket was it reaching 500 feet of altitude based on it being half the capacity of the earlier 6 foot long “Simple Cat-2” that may have reached 1000 feet of altitude.   An altitude of 400 feet wasn’t too far short of that estimate!!

Then it was back to the MTA for the fourth and what would be the final flight to another disappointing altitude of only 65 feet (again, derived  from timing it’s recorded flight duration).  – Clouds again…  Along with the ever faithful Jon Wells, Keith and Nadara Soules were present for this flight and kept a constant supply of energy bars and cold water bottles on hand.  They also helped clean the reflector, fill the rocket, and take care of trash.

First of four photos shows the Solar Cat reaching a height of 25-feet at the RRS MTA on March 25, 2021.

Second of four photos shows RRS member, Jon Wells, fielding questions at the RRS MTA. Note the new paint job on the receiver and tracking drive system.

Third of four photos shows the launch with the highest altitude on record of 400 feet on April 1, 2021 from just outside of Lovelock, Nevada. RRS member Dale Talcott can be seen to the left taking video of the launch.

The fourth of four photos shows the highest flight of the Solar Cat from the RRS MTA on April 10, 2021, reaching a height of 65 feet.

The device was being stored on my trailer in my back yard, oftentimes uncovered.  As the winter progressed, we had some rain and snow.  For protection – as well as aesthetics – we decided to paint it.  But what color?  Since the ultimate “improved performance rocket” would also be a tribute to my late wife Carmela, I asked her good friends Kime and Nadara what they thought Carmela’s favorite color was.  The answer was a certain shade of green. I had not known that, but they were both sure of it.  So Kime met me at Home Depot, and we picked out a color that was an ever so slightly lighter shade of what they both swore was Carmela’s favorite.  That’s how the color choice came about.

Developing data on heating rates versus reflector area:

Although we got a bit caught-up in trying to launch, certainly good experience to have before progressing to the ultimate “higher performance rocket”, the original purpose of collecting test data was never forgotten.  And while we collected heating data during these launches, most of it came from non flight heatings done strictly for the data.  We got in a total of 13 tests with the 6-foot span version between December 18, 2020 and April 18, 2021.

​We wanted to get the average of at least three “good” heating sessions for each of the three configurations to increase the confidence level of the averaged results.  With the constant problems being presented by tree branches, clouds coming out or the sun getting too low in the west to continue tracking, some of the results were settled on with less than 3 tests (marked by the dotted lines on the following graph).  Only two tests were conducted with the receiver box alone, because after the dismal performance, continuing seemed pointless.

​Also, while I’ve been referring to the 12-inch wide receiver box, the first 4 foot reflector, and then finally the two 3-foot reflectors (making a 6-foot total span), it’s important to account for that not being the actual “perpendicular intercept” of the sun’s rays.  For that, we need to subtract the width of the receiver box’s shadow, and the amount the reflectors are effectively shortened by their curvature.  The actual intercepted span of the receiver box is therefore, 8 inches, the “4 ft. mirror” is 34 inches, and the “6 ft. mirror” is 50 inches.  See the prior cross sectional drawing for illustration.

​And while we managed to get the “rocket” – actually a very heavy piece of 2 inch galvanized pipe with end fittings – off the pad and into the air 4 times, it’s clear that even with the 6 feet (actually 50 inches) of total reflector span, it was still pretty under-powered, requiring really good conditions to achieve launch.  The new “2nd generation” rocket and solar furnace with ten feet of total reflector span, should be a different story!

The results of the entire heating experiments over different spans of heater with projected curves.

Testing the commercial “Sun-Tura” 2-axis sun tracking system:

At some point, a novel idea occurred: “with all the solar activity going on these days, what if a commercial tracking system existed that we could just buy?”  A few Google searches eventually revealed the SunTura unit, and one was ordered.  

More headaches accompanied mounting it to the existing solar furnace, but it was eventually installed and operational … sort of …  The sun sensor when mounted as instructed, pointed the unit far to the east of the sun – about 25 DEGREES east of it!!!  The instructions said it could be adjusted by bending the stalks holding the little LED’s under the dome.  That idea made us pretty nervous though, as it seemed that getting it “right” on the first couple tries was unlikely, and too much bending could easily result in breaking something!

So several cardboard wedges to just sit the sun sensor on were made and tried until we found the angles that caused it to aim much better.  From this, the plan became making a more permanent version of this “mounting wedge” adding fine adjustment screws, in a manner borrowed from survey instruments Jon and I used a lot in the past.  Thus the last duty of the 1st generation solar furnace became to test the new sun sensor mount for the 2nd generation solar furnace.

This new mounting platform with adjusting screws was finally made, temporarily installed and tested, in August of 2022.  As hoped, it showed that it could be “dialed-in” to finally allow the SunTura sun tracking system to keep the assembly aimed at the sun accurately enough for our purposes.

The extreme angle of the cardboard wedge needed to correct the sun sensor’s output to have proper aim.

The new adjustable sun tracker mounting.

Now, with it’s last “assignment” finally completed, the battle weary Gen-1 solar furnace was mostly dismantled and removed from the trailer, clearing the way for construction of the new “Gen-2“ to begin.  Many of the screws and boards, many of them already “recycled” for this structure, were saved to possibly be used yet again in the future.

Jon Wells had come and stayed with me a few days, and was a huge help in this effort, including helping guide the decision that we’d leave a couple parts more-or-less intact to preserve some tangible evidence of this device: the two ends of the mount/yoke & the splice Dave McKinnon put in to adjust for a mistake of mine, and the launch tower that Dale had integrated into the receiver box so well that something would need to be cut to ever remove it.  Jon and I decided NOT to cut and remove it, but leave those last couple parts attached as testament to Dave’s and Dale’s creative solutions to make a solar powered launch on the 20 year anniversary of the original Scalded Cat’s first flight a reality.

Photo taken of the remaining pieces from the 1st generation Solar Cat receiver now sitting in storage.

This is a work in progress. Further updates will be reported as the Solar Cat project continues.

80th Anniversary of the Reaction Research Society

by Dave Nordling, President, RRS.ORG

Today, January 6, 2023 marks the 80th anniversary of the Reaction Research Society.

Founded in 1943 by George James, this small club of young boys gathered in Glendale to study and conduct experiments in reaction propulsion. It all began with a child’s chemistry set that included zinc and sulfur powder as a demonstration of exothermic reactions. This would lead to rockets and newer propellants and ever increasingly complicated experiments in the remote northern areas of Los Angeles from Mint Canyon to Rosamond to our present location east of Koehn Dry Lake. Many members of the Reaction Research Society have pursued professional aerospace careers making important contributions to the field.

In the war years of the 1940’s, rocketry was more the subject of science fiction and speculation than the robust science it is today in the early 21st century. Our society has seen eight decades of growth as a collective group of enthusiasts and colleagues supporting research and offering shared testing spaces while remaining dedicated to responsible projects and scientific progress. It is with pride that we share this milestone with the public on a bold new year of launch, facility improvement and looking forward to the future.

Photo credit: Mani Suri

Our month meetings are held at 7:30pm on the second Friday of each month by teleconference and in person at the front office of the Compton/Woodley Airport on Alondra Blvd. in Compton. For details, contact the RRS secretary.

November 2020 virtual meeting

by the Reaction Research Society

The RRS held its monthly meeting on November 13th. The teleconference brought several of us together again from different parts of the city and around the country.

The November 2020 monthly meeting by teleconference


The society discussed the prior launch event at the MTA held on November 7, 2020. Details of the event are in the firing report posted this month. More RRS members are buidling more rockets and this is a good trend, We were not able to work on site improvements at the event, but these tasks can be resumed at the next event.

Bent panel on the vertical thrust stand to be removed and replaced.

The 2020 Constitution Committee is almost ready to release the recommended draft with updated policy listings according to RRS vice-president Frank Miuccio who is the single executive council representative on the committee. Release of the draft will go out in December 2020. If ratified by the voting membership, this new Constitution will clarify RRS rules and policies and supersede all past revisions.


The society discussed the current progress of facility upgrades approved by the executive council. At the top of the list is a new restroom facility. A couple contractor bids are forthcoming with one being delivered just a few hours before the meeting. Cost is an important consideration and other designs besides a block and concrete design are being considered. Containerized, portable systems may offer an effective solution with fewer complications.

The council approved the construction of a similar pit toilet like the one being used now at the MTA. The second toilet is meant to replace the first as it will not be usable for much longer. A few RRS members are already working to put in this stop-gap measure until the better facility can be selected and funded.

Second pit toilet at the MTA approved as stop-gap measure.

Other site upgrades discussed at the meeting was replacement of the roof structure on the existing blockhouse which would be a temporary but safer solution until a new structure can br funded and built. Another was removal and replacement of the bent panel on the vertical thrust stand which may be done at the next launch event. Still another improvement option is purchasing some of the existing containers from our tenant for more storage space at the site and a water storage tank.

The blockhou at the RRS mTA. Replacement of the roof would enhance the safety of this shelter.

Other improvements include locked and secured areas for high-pressure gas bottles and cryogenic liquid cylindeers. Towable water-spraying trailers have also been discussed to augment fire-fighting at the MTA. A larger launch rail system for heavier liquid rockets at the RRS MTA is also being considered.


Amazon Smile is a program that allows Amazon to donate a small percentage of eligible purchases to a non-profit group, such as the Reaction Research Society.

go to Amazon Smile and see for yourself

You first enter:

then select the Reaction Research Society as your chosen recipient. There is no extra cost to you and you allow the RRS to recieve quarterly donations from Amazon when sufficient purchase amounts have been made in that quarter. With the holiday season approaching, this is a simple way to benefit the RRS and the many activities we support.

For more details and assistance, contact Chris Lujan or just go to Amazon Smile.


As per the Constitution, the society holds nominations at the November monthly meeting for the four elected executive council positions. Terms will start in the new calendar year, First, an election chairman is appointed by the council. This person can not run for office in that election cycle. Dave Nordling was selected as election chairman.

Only administrative members in active status can be nominated and hold office. The nominees will be listed on the ballots going out to the eligible voting membership. Ballots will be sent by email which underscores the importance of all members keeping their contact information up to date with the RRS treasurer. Also, one of the requirements of active membership is being current with dues payments also tracked by the RRS treasurer.


RRS policy is that all annual dues are to be paid on January 1 of each calendar year. New members joining the society throughout the year can pay a pro-rated dues for their first partial calendar year. Membership dues are $40 per year and student membership is $20 per year.

The executive council voted to keep dues at their current levels for the new calendar year, 2021. However, dues will increase to $50 per year starting in 2022 (January 1). All payments are made to the society to the RRS treasurer or the RRS president.


The next launch event at the Mojave Test Area will be held on December 12th. The nitrous oxide hybrid will be launched with the upgraded igniter and new airframe. Other rocket launches are planned.


The next monthly meeting will be held on December 11 by teleconference. Anyone with questions or suggestions should contact the RRS secretary.