Thursday, June 25, 2015

One Breeder’s Experience—doing his own thinking

From the September - October 2012 Newsletter

Most of you may not recognize the name of  Fred W. Owen (Homerville, Ohio) but his “Owenlea” herd of purebred Holsteins was well-known in Ohio as a productive and uniquely bred seedstock herd until his retirement in the 1990s.      His most famous cow was a World Dairy Expo winner,  Owen Marquis Wanda EX-96 , a product of linebreeding within the sire line of ABC Reflection Sovereign.

Mr Owen was not a “show type” breeder with an off-farm job milking just 24 cows in an old wooden stable— at its peak size in the 1980s, Owenlea had a 24,000 pound herd average on 300 cows in milk.  
By all measures, production, type and profitability, Mr Owen was a successful large-scale dairyman.

Following is Mr Owen’s approach to breeding, in his own words:

I would suggest that “linebreeding” is definitely not a high risk venture.    It is pretty darn hard to get a dangerously inbred animal using sire side linebreeding.

I used over 40 different sons of Rosafe Citation R over a long period of time.    [“Citation R” was a son of ABC Reflection Sovereign, and the last calf of Glenvue Nettie Jemima, who had 240,000m lifetime.]

In one particular case, we started with an own “Citation R” daughter and bred her to a “Citation R” son.   The resulting calf was 50% “Citation R”—or the SAME AS an actual “Citation R” daughter.

We followed this with seven consecutive generations sired by other “Citation R” sons.   So for eight generations, the sire was a “Citation R”son.

Why did I do it?    I did it because I had the resources, and the capability, and no one could stop me.   I did it because I could [and wanted to see first hand what would result].

To the unthinking, this may seem like extreme “inbreeding”—but it wasn’t!

All along these eight generations, the percentage of genetics from Rosafe Citation R remained the same as an own daughter of “Citation R”—exactly 50%.    If you doubt this, chart it out on a blank pedigree.

Most of those eight generations scored Very Good [above 85 points] or Excellent [above 90 points].   They definitely did not deteriorate toward the end.   They got better every round.   And guess what?   They looked exactly like own daughters of Rosafe Citation R.

We linebred our herd for decades to bulls as near as I could get to ABC Reflection Sovereign.   (He was the sire of Rosafe Citation R.)    I suppose there is a possibility my linebreeding wouldn’t work without “ABC”…    But he did exist and made many breeders look like a genius in his era.        But if he hadn’t, there would have been another.    Anyway, that linebreeding worked great.

I had a fabulous run, that produced over 100 homebred “Excellent” cows.    This great run included homebred cows scored up to EX-96, homebred winners at national shows (Waterloo, then Madison), many 30,000-pound and even a couple 40,000-pound producers, high priced consignments to national sales, and uncountable high-priced exports sold to many foreign countries.    I made money from cows.

During all this great run, I never paid more than a passing interest to progeny tests or PD [now PTA].   I did it MY way, ALL the way, and never once selected a sire based on milk production.   It was all done based on type.    I never thought about milk.   In fact I felt pity for the people struggling to wrench milk out of ugly cows.    When I was younger, I thought what they were doing was meaningless.

Yet we got great production, far better than any other big herd around Ohio at the time.   At one point it was over 24,000 rolling herd average on almost 300 cows.   We just kept our eyes on the ball, and tried to keep putting together pedigrees with ABC Reflection Sovereign in there as often as possible.

It would be very hard to convince me that “linebreeding” is a questionable practice.  I know it works.   I am not talking theory.   I lived my life immersed in it.   It worked great here and it worked that same era in Canada [ABC Reflection Sovereign was born in eastern Ontario in 1948].   I settled on “ABC” after seeing his Get of Sire from the small Rosafe herd dominate the Waterloo International in 1954 against all the moneyed show herds across the continent.    “ABC” had four All-American gets of sire in those years, which included many individual All-American winners.

Anyone who ever visited Rosafe in the 1950s or Romandale in the 1960s would never doubt the wisdom of linebreeding [to ABC Reflection Sovereign].


ABC Reflection Sovereign was the result in “aAa” terms of using a “sharp” (tall, dairy, style) sire on a “round” (smooth, strong, open) cow.     The dam of “ABC” was 20 years ahead of her time in her udder symmetry, teat shape and placement, housed within a wide and long rump.   She lived into early teenage with a moderate life production total.

His highest type average son, Rosafe Citation R, was the result of using a “round” (style, smooth, strong) sire—ABC—on a “sharp” (dairy, open, tall) cow who was a milk wagon and lived to be 16 years old, producing 240,000 pounds of milk in eleven lactations prior to 1960.

Unlike his sire (ABC) and more of his paternal brothers (Marquis, Roeland, Perseus, Caliban, et al), “Citation R” in his physique leaned to “sharp” (dairy, tall, open) qualities.    But his overall weight of qualitative possession was quite balanced.     Thus this balance made him more additive than subtractive across broad mating choices.   

This “balance” is essential for success in any linebreeding system.    To linebreed from more extreme physiques, the tendency is for physical balance to disappear—followed by depression of health, then fertility, and eventually production longevity.

Keep in mind that as Mr Owen was selecting from the many “Citation R” sons developed in both AI and by private breeders, they had different dams—the more successful sons having dams that complemented “Citation R” in qualitative [physical] mating balance.    In place of “PD” he wanted sons of high quality cows—traditional breeding attitude was “a bull rarely sires a cow better than the cow who birthed him.”
Thus in his linebreeding, he accumulated genes 50% from good bull dams, and 50% from the sire line anchor to a sire and his son who were ahead of their time for type while bred for longevity of milk.

In an era where genetic experts are so disillusioned by “index selection” results in commercial dairy herds that many will recommend “crossbreeding” as a solution, Mr Owen’s success is worth a thought.

Tuesday, June 23, 2015

Salvaging row crops for feed in a drought season

In my travels this year it appears that corn is the most stressed crop.   For the silage based dairy this is distressing, given its current stage of growth and nutrient demand.   In parts of the country (Indiana, Illinois, Minnesota) corn fields appear to be dying—farmers there are cutting it, wilting and baling it to salvage some feed value in an alternative hay form.

If you choose to do this, we still have enough growing season for summer annuals or grass options to yield some fall harvested forage.     (The easiest—plant some forage sorghum—is no longer possible as the national seed supply is already depleted.)  

One is to plant sorghum sudangrass and let it keep growing right up until frost threatens—then either chop it for ensilage or cut it to make baleage.     Sorghum Sudangrass is able to tolerate hot/dry conditions and grows fast as long as you have enough moisture to germinate (so time any planting right after a rainstorm).    It emerges quickly and will do a good job as long as you did not use a lot of atrazine.    Most varieties range from 75 to 90 day maturity, but the ones you can cut every 30-35 days offer great flexibility.    We have BMR-6 varieties.

Another is to wait into August, when we are more certain it will cool down at night, and put in Italian Ryegrass.     Barenbrug’s “Green Spirit” variety is pretty aggressive and still grows into those first days of frost.     It will yield a fall harvest of significant tonnage and then act as a winter cover crop, regrowing in the spring for early harvesting or green nitrogen plowdown.
Note “ryegrass” is never the same as “cereal rye”, it is a cool season high energy grass forage    producing dairy quality feed.   Italian ryegrass varieties act as annuals or biennials depending on whether seeded in the fall or the spring.    Cows love it as pasture, baleage or haylage.

If you do not want the crop to regrow in the spring, consider Forage Oats.   They will give a crop to harvest within sixty days (plant early August—harvest late September).   If you want a spring harvest, consider fall planted Triticale, also a dairy quality feed harvested on time.

A third “wild card” option, especially if you just harvested wheat or did not plant a Round-Up Ready crop in the intended field this spring, is to plant some Master Graze - which is a BMR, soft stalk, high energy and protein, 75-day corn from “Master’s Choice” capable of yielding five to seven tons dry matter as ensilage or baleage—or even direct-grazed.   Master Graze does fine on 1 ½ units of nitrogen per day and is feed quality in sixty days from planting.

As for our alfalfa fields, although harvests so far are only 50% of “normal” yields, they could still produce meaningful late summer and fall cuttings as soon as rain returns.    If feed needs find you taking a late season cutting it is always possible to overseed with high energy grass varieties next spring and maintain or gain yields beyond what you might expect from shorting the root energy storage by that late season cutting.

Drought seasons remind us of the risks of going “monocultural” (having our entire operation dependent on a single crop).    The forage-based cattle operation has flexibility not only from managing its stocking rates in low yield seasons, but from growing a mix of crops capable of capturing moisture when it falls, rather than a total dependence on summer season rainfall.

Friday, June 19, 2015

A Contrast between “new technology” and Time-Tested Selection

From the July August 2012 Dairy Newsletter

Prior to the advent of Genomics, selecting sires for AI sampling was a combination of pedigree review, physical inspection, and thorough analysis of the maternal cow line against the chosen mating sires.   In this, no AI system today is more thorough in screening maternal lines than International Protein Sires.

Among cow lines, one of the most enduring is the “Roxy” family (descendants of Glenridge Citation Roxy--  EX-97, Gold Medal Dam, All-American and All-Canadian show cow).     There are now over 135 “Excellent” maternal descendants of “Roxy” worldwide and up to 14 consecutive “EX” generations through one of her many “brood cow” daughters.   This is a “maternal traits” line, more about fertility and longevity than “fast-maturity” young cow production.     But lightning can strike at least once.

Qualifies top 20 TPI on production (82% Rel), needs a few more daughters scored for type (69% Rel)

6HO 1126  Brigeen ROCKWELL –ET         aAa: 2 3 4 1 5 6     “Tall + Open + Strong + Dairy”

+2068 TPI   +1920m  +07%  +90bf  +01%  +60pr    3.05 SCS   -0.2 DPR  -0.2 PL    $496 Net Merit
7% calving ease  (7% daughter calving ease)   +2.18 PTA Type  +1.21 Udders  +1.97 Feet and Legs

Note: the above is all progeny (and pedigree) based data, as “Rockwell” has not been Genomic tested!  (A quirk in the rules of the research—the contributing AI studs have a five year monopoly on G testing of bulls;  IPS was not yet in business in 1992 when the Genomic dairy research project began).

Sire stacks dictate the initial range of PTA values, but maternal cow lines can promise exceptions

“Rockwell” is the #1 TPI son of his sire “Toystory”, but his initial ratings for fitness traits suffer* a bit from this sire stack:   Toystory x Morty* x Convincer* x Emory x Integrity* x Prelude x Southwind ( x Apollo Ivanhoe x Citation R x Treasure Model x Roeland).    It takes really high Rel% to wean out the Parent Average assumptions built into “lower heritability” traits.

But when you consider the quality of “Roxy” cows in his pedigree:  EX 92 x EX 95 x EX 92 x EX 95 x EX 91 x VG 88 x EX 94 x EX 97 x VG Canada x EX Canada—and the first four added up to over one million pounds actual lifetime production (without rBST)—“Rockwell” has an exceptional inheritance.

Does new technology make proven technology obsolete over night?
The success of “Rockwell” would argue NO.     In fact, as a dairyman, your milk check is still based on actual progeny performance—not on imputed genetic value.     Breeding a profitable herd of cows still results from using reliable genetics for sires, mating cows carefully, and avoiding narrow (single trait, single index ranking) selection procedures.    Breed for an improved herd with “real” economic values.  

Who  has  the  emerging  “genetic  giants” ??

Taurus Service’s “EXPLODE” could have been #1 TPI in April.    International Protein Sires’ “ROCKWELL” could have been the highest milk solids yield of TPI ranked sires.
The world of Holstein breeding is going to pay attention to these two sires.

For decades, large AI systems have dominated TPI and $NM rankings by virtue of large sire sampling systems.    But with the advent of Genomics this playing field has leveled   a bit, while focusing more on what index selection overlooks that remains important for dairy profitability brings more uniquely usable sires to the forefront each generation.

Inside, we tell a story of two new Holstein bulls—but the contrasts in their path to value apply to any breed (as well as crossbreeding), as finding genetic value is defined by the approach you take to harvest the gene traits and qualities that can add to your income.

Hot summer weather and conception rates in Holsteins

From the July-August 2012 Dairy Newsletter

Everyone knows that it is harder to catch cows in heat in hot weather, and that even if you do, as many as 60% of conceptions begun in this kind of heat do not survive to term.    What options exist?

(1)   Sprinklers over feed alleys  allow cows a chance to cool down body temps and maintain normal
feed intake.      (2)   Increased access to drinking water  allows cows to rehydrate; if grazing, get the water tanks right out where cows are grazing.       (3)   Exhaust stale air away from stalls where cows are lying down, a bigger electric bill for fans may still be cheaper than lost pregnancies and delays.

Higher conception semen:  the #1 SCR progeny-proven Holstein sire (+4.7) is currently available here:
99HO 6145  Askew-James Barry   (aAa 4-2-3-6-1-5)   marketed through International Protein Sires.
He is from the “Bootmaker Mandy” cow line, sired by “Shottle”.     April data:   (86% Rel)   “balanced”
+393m  +08%  +36bf  +01%  +15pr   2.71 SCS   6% ce   +2.04 Type  +1.86  Udders  +1.46  Feet & legs

If  expectations of conception are really low, consider this option:   We have low cost Jersey semen:
179JE 026   Just A Few Jude Easton   (aAa 3-5-4-2-1-6)   which has proved extremely potent in cross - breeding on Holsteins as well as pure breeding on Jerseys.      Save your higher cost sexed semen when breeding heifers for calving ease until cooler weather—use this potent Jersey bull who maintains size now when the heat is going to influence conception rates in the best of conditions.

Wednesday, June 17, 2015

Is “somatic cell” as much a selection trait as PTA milk or type??

The USA is unique in that in every major dairy breed, you have two choices for ranking indexes—AIPL (USDA) “Net Merit” and each breed association’s TPI (Holstein), JPI (Jersey), etc.

Holstein USA’s “TPI” (and its Canadian corollary, LPI) is calculated from milk production and type classification traits, weighted roughly 60%/40%. Jersey USA’s “JPI” is somewhat similar, more heavily weighted to butterfat yield, then protein, then type. None of these indexes put as much emphasis on health, fertility and longevity as “Net Merit”—which is a curious switch from the traditional view of breed associations, founded on the belief in “longevity” as cow profitability.

“Net Merit” over time has been transformed from an index weighted 70% on “milk pounds” (1970) to 0% on “milk pounds” today—placing roughly equal emphasis on fat + protein, health, fertility and associated assumptions relative to cow size and feed efficiency. Net Merit became “lifetime” net merit in response to the loud complaints of commercial dairying—“cows turn over too fast”.

PTA- SCS (somatic cell score) is the key “health” component in “Net Merit”. Why is this?
Not just because mastitis is the scourge of dairying, the biggest source of lost income during any cow’s lactation. A somatic cell is not just a “mastitis” cell—it is any cell with a nucleus, thus it could be a bacterial or viral cell, it could be a spent white blood cell, it could be a shed skin cell.  
Thus the level of SCC in any individual cow’s milk indicates the total functioning of her immune system in fighting infections and protecting health.

Canadian Dairy Network (the research agency that calculates Canadian genetic evaluations) in a major study of the production patterns of dairy cows over their lifetimes concluded that SCS is an indicator of the potential productive longevity of dairy cows. They found (1) the sires with higher SCS ratings, had the earlier-maturing, faster aging daughters, less productive than their evaluations suggested in second and later lactations (if they even had mature lactations); (2) the sires with lower SCS ratings, produced daughters that matured better, lasted longer, and exceeded early age estimates of productivity in their mature lactations.

The summary conclusion:  if you are seeking longevity, pay attention to SCS ratings. In other words, SCS is not just about mastitis. Thus we cannot blame higher SCS scores on bad milking habits or maladjusted vacuum systems. Fragile health is a heritable condition.When we ignore SCS ratings in favor of bigger numbers, like PTA Milk, we can accumulate problems as well as add to generations of “single trait” selection in genetics.

An example of this is that the highest SCS cows in many herds are “lame” cows with clean udders on CMT tests. Foot rot, heel warts and laminitis will all elevate SCC levels, simply because the immune system is producing leucocytes in volume to reach these furthest extremities, where blood flow is small. Infection fighting cells, once “spent,” are expelled by the body through the udder. 
The cell counting machines at the dairy labs add them up in the SCC totals.

Even extraordinary high butterfat% and protein% levels in milk will elevate SCS a bit, because the thicker (more milk solids density) milk draws some skin cells along during final milking. Again, a basically benign cell gets counted the same as an infection cell. But this usually stays within 10% of the SCS levels on lower testing cows. (There is no evidence that higher SCS levels will affect cheese yields—possession of the “B” Kappa Casein gene and the level of protein % have the greatest positive impact on cheese production from any given volume of milk.)   
Are you prepared for the coming 400,000 SCS limits on “legal” milk? Many of us, under the existing system that uses 750,000 as the legal limit, see our herds range from 200,000 to 500,000 according to the season. Thus the new 400,000 limit exposes too many dairymen to penalties and diversion of milk to sub Grade A processing at a lower milk price. While a lower SCS limit forces us to pay more attention to milking system maintenance, cow stall bedding, hoof care, vaccination, post-calving cow health, milking prep procedures and teat dipping, without considering the genetic links to health, it can still feel like a losing battle.

Chronically high SCS levels are NOT “normal” and only environmental

I listened to a Holstein USA fieldman advise southern dairyman that “sire SCS data is meaningless, all high SCS scores are the fault of bad milking machines”—his further advice was that, unless sires are +2.50 for Type, they cannot help you breed good udders—meanwhile, the average herd SCS in the southern USA is dangerously close to the coming legal limit.   

Increasingly, astute dairymen are coming to recognize that too much of their advice on genetics is coming from opinion rather than research-based and observationally verified fact. Advice that is not leading you to solutions other than “cull the bad ones” is poor genetic theory and expensive in reality.

The facts are this—SCS has as high a level of heritability as most type traits other than udder traits.
We all pay a great deal of attention to linear type data—why then ignore SCS data??

The dollars and cents of the SCS scoring that adds up to PTA- SCS

AIPL standardizes each breed’s SCS average at 3.00. Thus an average bull will be SCS 3.00.   It is a trait where lower score is better—thus bull books that say “+3.30 SCS” are incorrectly stating this, given you might interpret the worst scores as better than the best scores.

Each log point of SCS represents approximately 70,000 cells for a Michigan dairyman (the average of all milk tested through MMPA in 2010 was slightly over 210,000 raw count).

Thus the difference between really healthy SCS and problematic SCS looks like this: (Dec 2011)

080H 1067  Atlantic            SCS 2.48    +0.2 PL                   200H 1758  Leheros                SCS 3.31    -0.8 PL
006H 1124  Tycoon Red     SCS 2.50    +6.1 PL                   007H 2236  Elton *BL            SCS 3.19    -1.2 PL
054H 0480  Legend             SCS 2.51    +2.0 PL                   029H 8343  Convincer *CV   SCS 3.18    -3.3 PL                             
076H 0500  Refresh *RC   SCS 2.52     +4.5 PL                   200H 7030  Talent *RC         SCS 3.18    +2.6 PL
076H 0569  Hi Tech            SCS 2.55    +4.3 PL                   011H 3073   Luke                   SCS 3.16     -3.8 PL
006H 1152  Zade                 SCS 2.57   +6.2 PL                   200H 0044   Morty                  SCS 3.11    -4.1 PL

Further correlations will often show up as differences in levels of UDC (Udder Composite) or FLC (Feet and Leg Composite) to indicate type-related sources of rising SCS levels, or differences in the DPR (Daughter Pregnancy Rate) to indicate slower fertility that can, with SCS, contribute to shorter herdlife. Now consider how multiple generations can add up negatively into big SCC differences.

In other words …

Health traits, primarily SCS, and fertility traits, primarily DPR, add up to major impacts on the cow Productive Life length. Individual exceptions aside, the pattern is clear from the above examples.

Genetics has changed since Hoover was President

Mostly, what has changed is the presentation of the data. Old assumptions as to the value of PD (now PTA) volumes for milk, fat and protein, and the value of type score information, need to catch up to what we have learned by more recent genetic studies of health and fertility patterns.

There are “pretty” cows that act frail, “milky” cows that lack fertility, and “functional but plain” cows that can live productively forever even if only herd average as heifers.
These variations are explainable by looking at all the data, and rather than just a blind look at your favorite trait number or ranking index, can help you to a more profitably precise selection of AI sires.

The most total value (herd equity plus milk check cashflow) cows come from genetic packaging of type, production and health in a positive balance, leading to longevity.
Taking shortcuts to save a couple bucks on a straw or gain a phantom 500 pounds on PTA Milk is where we miss out on accumulating gains in productive longevity.

Sunday, June 14, 2015

A Tale of Two Emerging “Exceptional” Sires

From the July August 2012 Diary Newsletter

The mainstream of dairy genetics has been enthralled with Genomics since the first release of G tested “super samplers” four years ago.    40% of  Holstein/ Jersey/ Brown Swiss semen sold is now from G tested sires that do not yet have progeny evaluations.    Taurus Service and its foresighted President Richard Witter were in the vanguard of the seven major AI studs that supported research into Genomic applications for dairy, alongside six purebred breed associations and the AIPL geneticists at USDA.

In Genomics, we look at fractions of the DNA for the “marker” genes associated with the genotypes of historical high performance sires.    PTA values are “imputed” from the gene possession of current sires compared to this historical database.    Rel% of Genomic data is almost double that of Parent Averages.

In this time there have been some spectacular failures (Shampoo, Mitey P, Spearmint) but each summary at least one sire will live up spectacularly to the Genomic expectations, keeping AI sire analysts excited by the potential of this tool to save them time locating sires who will “rank” in future summaries.

One more daughter and he would be the #1 bull on Holstein USA’s TPI ranking

76HO 581   Wabash-Way  EXPLODE –ET   (VG 88)        aAa:  2 3 4 1 6 5    “Tall + Open + Strong”

“Explode” generated enough excitement on his pedigree-exceeding Genomic values that he was used as a “sire of sons” as soon as semen was available.    He has sons in AI and daughters being ET flushed for AI, all based upon Genomic imputations.   Now with his first milking progeny, he looks like this:

+2265  GTPI    +1703m  -03%  +54bf  -01%  +49pr   2.75 SCS   -0.6 DPR   +3.5 PL   $548 Net Merit
8% calving ease   (6% daughter calving ease)   +3.74 PTA Type   +3.08 Udders   +3.02 Feet & Legs

What sort of pedigree generates this level of Genomic success?

“Explode” is a son of Sandy-Valley Bolton who was #1 TPI sire at time of his birth.  His dam Wabash-Way Emilyann (VG 88) (Dam of Merit) is a daughter of previous #1 TPI sire, Picston Shottle.   Emily Ann is the eleventh (11th) generation of VG or EX cows with calculated indexes, from the cow line of Whittier Farms Lead Mae  (EX-95), an early “Leadman” from a “Blackstar”.

The “sire stack” explains how you add up high indexes:  Bolton x Shottle x O-Man x Mtoto x Rudolph x Leadman x Blackstar:  all were #1 for Net Merit, TPI, LPI (Canada), BW (UK) or Inet (Italy).   Such are the sires on which the Genomic database is based and the imputation of value is calibrated.

Is crossbreeding truly a breeding “method” or just a breeding “option” ?

Crossbreeding has gained momentum in three major areas of dairy decision making:
(1)     use of Jersey sires on virgin Holstein heifers, to insure calving ease at earlier calving ages.
(2)     use in conversion of dairy herds to intensive rotation grazing to improve fertility and vigor.
(3)     use in large confinement herds to improve health, fertility, and feed conversion qualities.

On the last genetic base update by AIPL (USDA), it was noted that “crossbreed” has become the second largest breed on DHIA test, behind Holsteins, ahead of Jerseys and all other pure breeds. This has led to the inclusion of crossbred progeny in their sires’ individual genetic evaluations.

What is different between “pure” breeding and “cross” breeding?

A dumb question, right? A purebred is descended from a single breed sire line. A crossbred has a multiple of breeds in her recent sire stack. But on a molecular (DNA) basis, this is not as big a deal as it sounds—because 85+% of all bovine gene patterns are shared in common among all dairy breeds.

The entire theory of crossbreeding is based on the assumption that less than 15% of the gene total will be different enough to cause extraordinary “heterosis” [hybrid vigor] in such a mating. In a biological sense, the difference between “outcross” (different bloodlines within a breed) and “crossbreeding” (use of mates from another bred) is only a matter of degree.

The interest in crossbreeding stems from dairymen’s observation that over time AI sire lines become more closely interrelated. Fears of “inbreeding” combined with loss of calf and cow vigor and slower, more costly reproduction, adding up to shorter average herd life, led to crossbreeding experiments. 

What measurable effects are seen?

To me, three key research studies tell us how much to rely on crossbreeding:

(Netherlands)  “inbreeding depression” is caused by “single trait selection” (not pedigrees with ancestors in common)
(New Zealand)  “hybrid vigor” is worth 6% on a first cross, 3% on a second cross, 1.5% on a third cross—ie, the more crossbred generations, the level of heterosis declines to an undetectable level.
(Iowa State)  “crossbreed” vigor is good in a first cross, gains a bit on second cross, peaks by the third cross, and then declines—after eight crosses, you have lost all predictability of results.

To state it simply—“crossbreeding” is a short-term fix, but not a substitute for genetic selection.